Android Test-Driven Development by Tutorials

Android
Test-Driven
Development
by
Tutorials
By Lance Gleason, Victoria Gonda & Fernando
Sproviero
Licensing
Android Test-Driven Development by
Tutorials
By Lance Gleason, Victoria Gonda and Fernando
Sproviero
Copyright ©2019 Razeware LLC.
Notice of Rights
All rights reserved. No part of this book or
corresponding materials (such as text, images,
or source code) may be reproduced or
distributed by any means without prior written
permission of the copyright owner.
Notice of Liability
This book and all corresponding materials (such
as source code) are provided on an “as is” basis,
without warranty of any kind, express of
implied, including but not limited to the
warranties of merchantability, fitness for a
particular purpose, and noninfringement. In no
event shall the authors or copyright holders be
liable for any claim, damages or other liability,
whether in action of contract, tort or otherwise,
arising from, out of or in connection with the
software or the use of other dealing in the
software.
Trademarks
All trademarks and registered trademarks
appearing in this book are the property of their
own respective owners.
Dedications
"There are many people who helped to make
this book possible. My other half Marlene was
the one who initially suggested that I try out for
the Ray Wenderlich team. She gave me lots of
encouragement and, even before the editors saw
my work, edited every chapter to make sure my
sentences were coherent. My late mother passed
on her love of reading and many creative skills
for which I will always be grateful. The many
strong women and family members in my life
who taught me to live life with honesty and
conviction and were encouraging of my work. I
owe a debt of gratitude to the Ruby community
for teaching me about TDD and infecting me
with enthusiasm. I’d also like to thank Ray
Wenderlich and the team for giving me the
chance to share my love of TDD with the world.
Finally, I’d like to thank all of the editors and
co-authors of this book. It has been a very
rewarding experience working with everybody
on the team."
— Lance Gleason
"To my family, friends, and especially my
partner, who supported me while writing this
book. Tyler, thanks for all your encouragement
and patience you gave me as I spent evenings in
front of my laptop turning thoughts into words."
— Victoria Gonda
"To my girlfriend Romina, who will soon become

my wife. Thanks for your support and help,
examples and discussions about the topics of
this book. Yes, she’s a developer too, isn’t that
great? :] Also, to my family, who had to listen
repeatedly about what I was writing, and no,
they aren't developers!"
— Fernando Sproviero
About the Authors
Victoria Gonda is an author of this book.

Victoria is a software developer working mostly
on Android apps. When she's not traveling to
speak at conferences, she works remotely from
Chicago. Her interest in tech started while
studying computer science and dance
production in college. In her spare time, you can
find Victoria relaxing with a book, her partner,
and her pets. You can connect with her on
Twitter at @TTGonda.
Lance Gleason is an author of this book. Over
the years, he has worked in everything from
server side development, IoT, wearables, to
mobile development. He began his TDD journey
as a Ruby developer and has been an Android
fan ever since he preordered the G1 in 2008.
After a few years of being an Android fan he
finally took the plunge into Android
development when he became a Google Glass
Explorer in 2013. He’s had a chance to work on
project for organizations ranging from small
startups to large companies with millions of
users. When he is not writing code, you can find
him traveling to interesting places, wine tasting,
scuba diving or exploring a new hobby or
interest. You can find him on Twitter at
@lgleasain.
Fernando Sproviero is an author of this book.
He’s a mobile developer who works in Buenos
Aires, Argentina. He’s a graduated software
engineer who enjoys learning and discussing
about good practices and architecture. He’s
passionate about Android, iOS and API backend
development. You can reach out to him on
Twitter @fernandospr or on LinkedIn
@fernandosproviero.
About the Editors
Tammy Coron is an editor of this book. She is an

independent creative professional and the host
of Roundabout: Creative Chaos. She’s also a
Development Editor at The Pragmatic
Bookshelf, a Sr. Editor at Razeware, and a
content contributor at Creative Bloq,
Lynda.com, iMore, and raywenderlich.com.
Alex Curran is a technical editor of this book. He

is a lead developer in Berlin, with a keen
interest in development best practices and
encouraging collaboration between iOS and
Android developers. In his spare time, he cooks,
reads avidly, and makes natural soaps by hand.
You can find him anywhere at @amlcurran.
Matei Suica is a technical editor of this book. He

is a software developer that dreams about
changing the world with his work. From his
small office in Romania, Matei is trying to create
an App that will help millions. When the laptop
lid closes, he likes to go to the gym and read.
You can find him on Twitter or LinkedIn:
@mateisuica
Jonathan Wong is the final pass editor of this
book. Whether it's Android, iOS, or even cross-
platform frameworks like Flutter, as long as it's
mobile, he's interested. Jonathan runs the
mobile team at Tealium building data collection
libraries and is a member of the video team at
raywenderlich.com. He’s a software engineer in
sunny San Diego where he’s worked on
everything from mobile apps, to front-end
Javascript, to server-side development. When
he’s not taking care of his two kids, three cats,
or five chickens, he’s usually trying to steal away
time for a good run or a good cup of coffee. You
can follow him on Twitter at @fattywaffles.
About the Artist
Vicki Wenderlich is the designer and artist of

the cover of this book. She is Ray’s wife and
business partner. She is a digital artist who
creates illustrations, game art and a lot of other
art or design work for the tutorials and books on
raywenderlich.com. When she’s not making art,
she loves hiking, a good glass of wine and
attempting to create the perfect cheese plate.
What You Need
To follow along with this book, you'll need the
following:
Kotlin 1.3: This book uses Kotlin 1.3

throughout. The examples may work with
an earlier version of Kotlin, but they are
untested.
Android Studio 3.5 or later. Android Studio

is the main development tool for Android.
You'll need Android Studio 3.5 or later for
the tasks in this book. You can download
the latest version of Android Studio from
Android's developer site here:
https://developer.android.com/studio.
If you haven't installed the latest version of

Android Studio, be sure to do that before
continuing with the book. The code covered in
this book depends on Android 10, Kotlin 1.3 and
Android Studio 3.5 — you may get lost if you try
to work with an older version.
Book License
By purchasing Android Test-Driven
Development by Tutorials, you have the
following license:
You are allowed to use and/or modify the

source code in Android Test-Driven
Development by Tutorials in as many apps
as you want, with no attribution required.
You are allowed to use and/or modify all

art, images and designs that are included in
Android Test-Driven Development by
Tutorials in as many apps as you want, but
must include this attribution line
somewhere inside your app:
“Artwork/images/designs: from Android
Test-Driven Development by Tutorials,
available at www.raywenderlich.com.”
The source code included in Android Test-

Driven Development by Tutorials is for your
personal use only. You are NOT allowed to
distribute or sell the source code in Android
Test-Driven Development by Tutorials
without prior authorization.
This book is for your personal use only. You

are NOT allowed to sell this book without
prior authorization, or distribute it to
friends, coworkers or students; they would
need to purchase their own copies.
All materials provided with this book are

provided on an “as is” basis, without warranty of
any kind, express or implied, including but not
limited to the warranties of merchantability,
fitness for a particular purpose and
noninfringement.
In no event shall the authors or copyright

holders be liable for any claim, damages or
other liability, whether in an action of contract,
tort or otherwise, arising from, out of or in
connection with the software or the use or other
dealings in the software.
All trademarks and registered trademarks

appearing in this guide are the properties of
their respective owners.
Book Source Code &
Forums
If you bought the digital edition
The digital edition of this book comes with the
source code for the starter and completed
projects for each chapter. These resources are
included with the digital edition you
downloaded from store.raywenderlich.com.
If you bought the print version

You can get the source code for the print edition
of the book here:
https://store.raywenderlich.com/products/andro
id-test-driven-development-by-tutorials
Forums
We’ve also set up an official forum for the book
at forums.raywenderlich.com. This is a great
place to ask questions about the book or to
submit any errors you may find.
Digital book editions
We have a digital edition of this book available
in both ePUB and PDF, which can be handy if
you want a soft copy to take with you, or you
want to quickly search for a specific term within
the book.
Buying the digital edition version of the book

also has a few extra benefits: free updates each
time we update the book, access to older
versions of the book, and you can download the
digital editions from anywhere, at anytime.
Visit our Android Test-Driven Development

store page here:
https://store.raywenderlich.com/products/a
ndroid-test-driven-development-by-
tutorials.
And if you purchased the print version of this

book, you’re eligible to upgrade to the digital
editions at a significant discount! Simply email
[email protected] with your receipt for
the physical copy and we’ll get you set up with
the discounted digital edition version of the
book.
Book Updates
Since you’ve purchased the digital edition
version of this book, you get free access to any
updates we may make to the book!
The best way to get update notifications is to

sign up for our monthly newsletter. This
includes a list of the tutorials that came out on
raywenderlich.com that month, any important
news like book updates or new books, and a list
of our favorite iOS development links for that
month. You can sign up here:
www.raywenderlich.com/newsletter
About the Cover
Ravens are one of the smartest animals in the
world. They can solve puzzles, trick other
animals to gain an advantage, and are even
better communicators than apes! They're also
great planners: Ravens don’t leave anything to
chance and think ahead to ensure the best
chance of success.
Ravens are extremely adaptable and will test

new behaviors to stay ahead of other birds. In
fact, if ravens could code, they would likely be
big fans of test-driven development! Every
component of TDD contributes to greater
confidence and fewer mistakes, which ensures
survival of the fittest, whether you’re a bird or a
developer.
With Android Test-Driven Development, you’ll

be the top bird in your flock, with greater
confidence and fewer mistakes in your code!
Section I: Introduction
to Test-Driven
Development
This section introduces you to Android test-
driven development (TDD). If testing or TDD are
new concepts to you, we recommend starting
here. You’ll learn everything from what a test is,
why you should test, what you should test and
what you should not test.
You’ll then get your feet wet by writing your first

Kotlin test, independent of the Android
framework, with the principles of TDD. This
section lays the foundation for TDD.
Specifically, this section includes the following

chapters:
Chapter 1: Introduction: If you're new to testing

and want an overview of what this book covers,
start with chapter 1 for a gentle introduction
into testing.
Chapter 2: What Is a Test?: Do you know why
you should test or what to test? In this chapter,
"What is a Test?", you'll learn both of these and
also dive into code coverage.
Chapter 3: What Is TDD? Now that you know

what a test is, start writing your first tests! Not
only will you start writing tests, but you'll also
do so in a test-driven development way.
Chapter 1: Introduction
Welcome to the first chapter of this book. If
you’re curious about learning more about
testing on Android, you’re in the right place.
In this book, you’ll learn about test-driven

development (TDD) on Android. You’ll do this
by working step-by-step in real apps as you
increase your testing knowledge.
Whether you’re a testing beginner, or ready to

bring your testing skills to the next level, this
book is for you. By the end, you’ll be able to
fully apply the TDD process to both your
greenfield and legacy apps.
Who is this book for?
This book is intended for Android developers
who know at least the basics of the Android
framework. You should also know enough about
Kotlin that you’re comfortable working with it.
Although it’s not required, it’s helpful to have

an understanding of design patterns,
dependency injection and app architecture.
Whether you’re new to testing and TDD, or
eager to apply your existing knowledge to your
apps, you’ll find the information within this
book useful.
This book covers everything from the bare

basics of testing knowledge, to how to apply
that knowledge to existing code. If any of this
interests you, read on!
Note: If you already finished Android
Apprentice, you’re at a great place to start
this book. If your knowledge of Android is
limited, that book is the perfect place to
start. You can find it at
https://store.raywenderlich.com/products/a
ndroid-apprentice.
How to read this book

This book is divided into three sections, each
covering a subsection of material. In these
chapters, you’ll follow along writing tests while
you take on new concepts. Depending on what
prior knowledge you’re coming in with, you may
want to start in a different section or jump
around. The sections are as follows:
Section 1: This is where to start if you’re

new to testing or TDD.
Section 2: If you know the fundamentals of

TDD and are ready for some hands-on
practice, start here. You’ll learn how to
follow TDD to add tests to new apps and
features. If there’s a specific topic in this
section that interests you, feel free to jump
to it. You’ll be directed to a previous
chapter if there’s anything you need to
know that you might have missed.
Section 3: When everything in the previous

section looks familiar, you can jump to this
section to learn how you can apply your
learning to a legacy project.
With that, it’s time for a sneak peek of each

chapter, so you can choose where to start.
Section 1
Chapter 1: Introduction
You’re here now. You’re learning what this book

covers, and where you should start.
Chapter 2: What Is a Test?
Continue to this chapter to learn about

automated testing and why it’s important. This
is an excellent read if you’re new to writing tests
or need convincing for why it’s important. In
this chapter, you’ll also learn what you should
and should not test, and the importance of code
coverage.
Chapter 3: What Is TDD?
This chapter introduces a core concept of this

book: test-driven development. In it, you’ll
learn the basics of this software development
process and why it’s useful. You’ll do this by
using TDD to create a search URL generating
function. Through this hands-on experience,
you’ll learn some of the tricky things that can
come up when testing, and how you can
approach them.
Section 2
Chapter 4: The Testing Pyramid
As you might imagine, there are multiple ways

to test your app, each having a different focus.
In this chapter, you’ll learn about the testing
pyramid, which describes the types of tests you
can write covering different layers of your app,
and how many of them you should write. It’s a
great intro before diving into the specifics in the
next chapters.
Chapter 5: Unit Tests
This is the chapter where you’ll learn about unit

tests, what they are and where you should use
them. You’ll use JUnit while practicing TDD to
add some logic to a new Cocktail app. In this
app, you can search for drinks, “favorite” them
and play a trivia game. The building blocks
you’ll learn in this chapter will carry you
through the rest of the book.
Chapter 6: Architecting for Testing
The architecture of your project can make or

break your testing experience. So, in this
chapter, you’ll learn about some of the most
common app architectures. Along with this,
you’ll learn some of the pros and cons of each
architecture and how they affect your tests.
You’ll finish off with some design principles that
will help you when you practice the refactor step
of TDD.
Chapter 7: Introduction to Mockito
Revisiting the Cocktail app, you’ll take your unit

tests to the next level using Mockito. You’ll
learn the basics of how to use mocks and spies
to stub behavior to keep unit tests pure,
completely independent from other classes.
You’ll also come away understanding the
difference between black-box and white-box
testing.
Chapter 8: Integration
This chapter looks at testing how different

objects interact or integrate with each other.
You’ll practice writing integration tests by
working on a Wishlist app to keep track of the
wishlists of your family and friends. While doing
so, you’ll learn how to manage your interactions
with the Android framework while testing.
Chapter 9: Testing the Persistence Layer
Continuing with the Wishlist app, this chapter

teaches you the fundamentals of testing a
persistence layer while you build out the
database for this app. You’ll learn how to handle
statefulness in your tests and strategies for
creating randomized test data. While working
with the RoomDB library, you’ll also come
across the gray line of testing your code versus
testing the library you’re using.
Chapter 10: Testing the Network Layer
While HTTP requests can be unpredictable,

here, you’ll learn how to write predictable
network tests while working on an app named
Punchline that shows you random jokes to
share. You’ll learn three different libraries to
help you accomplish varying degrees of depth in
your testing. With all of these libraries, it opens
up a discussion about how to choose the right
libraries for your project.
Chapter 11: User Interface
Finishing out this section is all about testing

your user interface. While adding the visual
elements to the Punchline app, you’ll learn how
to match views, perform actions and make
assertions. By having these tests, you’ll be able
to guard against any future regressions.
Section 3
Chapter 12: Common Legacy App Problems
Launching into this section about testing a pre-

existing app, you start by learning some of the
common things you might encounter while
doing this. By recognizing some familiar
patterns, you’ll be prepared to tackle these
situations in future chapters. Along the way,
you’ll learn some reasons why these patterns
may have made their way into the code.
Chapter 13: High-Level Testing with Espresso
One of the first tests you might add to a legacy

application is an Espresso test. In this chapter,
you’re tasked with adding features to the Coding
Companion Finder app, an app to find a dog or
cat to pair program with. Along the way, you’ll
learn how to set boundaries and set up testing
in an existing app.
Chapter 14: Hands-On Focused Refactoring

g
Refactoring is often used to help make an app
more testable and easier to modify. This chapter
guides you through refactoring part of the
Coding Companion Finder app by writing tests
for safety. You’ll learn how taking small steps
and moving slowly can keep your test suite
green as you make changes — even if that
means you need to change your tests.
Chapter 15: Refactoring Your Tests
Your tests are code too, and sometimes the first

tests you write aren’t the quickest or most
maintainable. That’s why, in addition to
working on refactoring the app code, you’ll
refactor the tests of the Coding Companion
Finder app. You’ll be able to isolate your tests
better, making them less brittle and more
helpful.
Chapter 16: Strategies for Handling Test Data
There are many ways to manage test data, and

this chapter is perfect for helping you make that
decision. By learning the pros and cons of each
strategy, you’ll be better equipped to decide
what’s best for your app.
Chapter 17, Continuous Integration & Other

Related Tools
There are many more tools you can interact

with in the testing world. One type is
Continuous Integration (CI) frameworks. In this
chapter, you’ll explore what it means to have CI
for your tests, along with some tools you can
use to achieve it. You'll learn the pros and cons
of different strategies so you can make the best
decision for your team.
Chapter 18: Testing Around Other Components
One of the tricky parts of testing can be around

interacting with other libraries and parts of the
Android framework. One example you'll see is
with permissions. In this chapter, you’ll learn
how to set boundaries in your test and use
strategies for handling these situations,
including when it might be better to not spend
time on writing tests.
Appendix
Other Related Techniques
This chapter covers some techniques you can

use to compliment your TDD process. By using
these strategies, you can choose how to define
what you’ll test and build. You can also bridge
the gap between technical and non-technical
written specifications.
Key points
There are different places you can start
reading this book, depending on what you
already know.
You can jump to the chapters that interest

you most.
Previous chapters are referenced when

necessary.
Where to go from here?

You now know what this book is about, so it’s
time to dive in. Feel free to continue to the next
chapter to get started or skip ahead to another
chapter on a topic that interests you the most.
Chapter 2: What Is a Test?
A test is a manual or automatic procedure used to evaluate if the System
Under Test (SUT) behaves correctly.
The SUT may be a method, an entire class, a module or even a whole

application.
From now on, when mentioning anything related to writing a test this
book will be referring to the automatic procedure form of a test.
To write a test, you need to understand the feature, specification or

requirement of the component you are implementing. That component
may be an Activity, Fragment, View Model or several of these
components working together. These may come in different forms, such
as user stories, use cases or some other kind of documentation.
Testing is an important part of software development. By including tests

along with your code, you can ensure that your code works and that later
changes to the code won't break it. Tests can give you the peace of mind
you need to develop quickly and catch bugs before they're released.
Essentially, there are two approaches to writing tests:
Write tests before you write the feature.
Write tests after you write the feature.
This book primarily focuses on writing tests first versus writing them
after a feature has been implemented.
Why should you test?

Writing tests can take more time up front, and it is code you write that
the client won't "see", which is why tests are sometimes skipped by
developers. However, having tests can speed up development down the
road and it presents some advantages.
Change/refactor confidence
You have probably run into a scenario in which you have a section of your
application that works correctly before adding new functionality to the
application. After adding new functionality, either in Quality Assurance
(QA) or after it is released to customers you discover that this new
functionality broke the previously working section. That is called a
regression.
Having good, reliable, effective tests would have caught that at the
moment the bug was introduced saving time in QA and preventing
preventable bugs from making it to your users. Another related scenario
is where you have a section of your application that is working correctly,
but could use some refactoring to use a new library, break things up to
follow a more readable architectural pattern, etc. A good test suite will
provide you with the confidence to make those changes without having
to do time consuming manual QA regression test cycles to ensure
everything is still working correctly.
However, you should always bear in mind that this is not a 100%
"insurance". No matter how many tests you write, there could be edge
cases that the tests don't catch. Even so, it's absolutely safer to have tests
that catch most issues than not having them at all!
Usually, you will write tests for the most common scenarios your user
may encounter. Whenever someone finds a bug that your tests didn't
catch, you should immediately add a test for it.
Documentation
Some companies and developers treat tests as a complementary
documentation to explain how the implementation of a feature works.
When you have well-written tests, they provide an excellent description
of what your code should do. By writing a test, its corresponding
implementation and repeating this until a feature is completed, bearing
in mind that these tests can be treated as specifications, will help you
and your team when a refactor or a modification of the feature is
required.
When you're working on a piece of code, you can look at the tests to help
you understand what the code does. You can also see what the code
should not do. Because these are tests rather than a static document, as
long as the tests are passing you can be sure this form of documentation
is up-to-date!
How to write a test

There are many things to bear in mind when writing a test. You'll
understand them by reading this book and practicing writing tests.
However, the most important aspects of writing a test are as follows:
Naming: You should give a meaningful name to each test so that it is

clearly identifiable in code and in subsequent reports.
For example, consider a quiz game:
fun whenAnsweringCorrectly_shouldIncrementCurrentScore() {
...
}
This test's name represents the state of what you are testing and the
expected behavior. This is useful when looking at the report after
running a test suite.
Short and simple: You should aim to write tests that focus on a
narrow piece of functionality. As a rule of thumb, if your test
methods get long, and have multiple assertion statements to check
conditions of the system, it may be trying to test too many things. In
that scenario it may be a good idea to break up that test into
multiple, more narrowly focused tests. Take a look at this test:
fun whenIncrementingScore_shouldIncrementCurrentScore() {
val score = Score(0)
score.increment()
if (score.current == 1) {
print("Success")
} else {
throw AssertionError("Invalid score")
}
}
The test only has seven lines of code to bring the SUT in the desired state
and check the expected behavior.
Check one single thing: Check one thing at a time. If you need to
test multiple things, write an additional test similar to the one
you've just previously run, but change the check:
fun whenIncrementingScore_aboveHighScore_shouldAlsoIncrementHighScore()
score.increment()
if (score.highest == 1) {
print("Success")
} else {
throw AssertionError("Invalid high score")
}
}
As you can see, this test is very similar to the previous one; however, the
check is different. In the first test, you checked that the score of the quiz
game incremented correctly. Now, you check that the highest score also
increments along with the score.
Readable: Anyone in the team should be able to read and understand

what is going on in your test and/or what is the purpose of the test.
Consequently, you should pay attention to the naming of each
variable or method used and the logic sequence of the test. If you
don't, then the tests will become difficult to maintain and keep up-
to-date.
What should you test?

You should test code that is related to the logic of your app. This may
include code that you have to write to:
Show the UI and navigate between the screens of your app.
Make network requests to an API.
Persist data.
Interact with device sensors.
Model your domain.
Having tests for the logic of your application should be your main goal,
however, bear also in mind the following:
Code that breaks often

If you have a legacy project without tests, and it breaks often whenever
you modify its code, it's useful to have tests for them, so that the next
time you make a modification you will be sure that it won't keep
breaking.
Code that will change

If you know that some code will be refactored in the near future, tests
will be useful here, too, because if you wrote tests for this feature, you
can support on them to refactor the code and be sure you don't break
anything.
What should you not test?

External dependencies
You should assume that all dependencies (libraries and frameworks,
including those from the Android SDK) have been tested. Thus, you
shouldn't test functionality of external libraries because the goal is to
test things that you control, not a third party tool created by someone
else.
Note: In the real world, sometimes some of those dependencies are

not tested. So, as a rule of thumb, when you have to choose between
two or more libraries that have the same functionality, you should
go with the one that has tests. This assures you that the features of
the library work as expected and that the library developers won't
break the features when adding new features and releasing new
versions.
Autogenerated code
You shouldn't write tests for autogenerated code. Following the previous
principle, it's supposed to be that the library or tool that generates code
is tested properly.
When should you not test?

Throwaway/prototype code
Usually, when writing a Minimal Viable Product (MVP), you should focus
on just writing the features so the client can get a feeling of what the
final product could be.
However, all the stakeholders need to understand that all the code (or
almost everything) you wrote will be thrown away. In this case, it doesn't
make sense to write any kind of tests.
Code you don't have time to test

This is a controversial topic. Often, developers get stuck in a rut wherein
they are fighting fires instead of proactively writing quality code, and
they are not given the time to address code quality.
If you are working on a cash-strapped startup, where requirements are

changing rapidly, that extra time to test could cause this fledgling
company to miss key deadlines, not iterate fast enough, fail to raise it's
next round of funding and go out of business.
On a new greenfield project, writing tests can double the amount of time
to get features out in the short term. But, as the project gets larger, the
tests end up saving time. Writing tests has its benefits; however, it'll take
time to write and maintain tests. You and your team will need to make
sure that you understand the trade-offs when determining which path
you want to take.
A Note on Technical Debt
When you take out a financial loan, you get the benefit of an
immediate infusion of cash. But a lender charges you interest on the
loan in addition to the principal, all of which you will need to pay
back. If you take on too much debt, you can end up in a situation
where it is impossible to pay back the loan. In this case, you might
have to declare bankruptcy.
Technical debt has many parallels to financial debt. With technical
debt you make trade offs in your code, such as not writing unit tests,
not refactoring, having less stringently quality standards, etc. to get
features out quicker. This is analogous to getting a financial cash
infusion. But as the code base grows, the lack of tests increase the
number of regressions, bugs, time it takes to refactor and QA time.
This is analogous to interest on a financial loan. In order to pay off
that debt you start to add unit tests to your code. That is analogous
to paying down the principal on a loan. Finally, if too many
shortcuts are taken for too long, the project may reach a point where
it is more advantageous to scrap the entire project and start with a
clean slate. That is the same as declaring bankruptcy to get relief
from too much financial debt.
Code spikes
At some point, you may find yourself working with a new library or,
perhaps, you may realize that you aren't sure how to implement
something. This makes it very difficult to write a test first because you
don't know enough about how you are going to implement the
functionality to write a meaningful failing test. In these instances, a code
spike can help you figure things out.
A code spike is a throwaway piece of untested code that explores possible

solutions to a problem. This code should not be considered shippable.
Once you have a solution, you will want to delete your spike and then
build up your implementation using TDD.
What is test coverage?

You can measure how many lines of code of your app have been executed
when you run your tests. An app with a high test coverage percentage
"suggests" that it works as expected and has a lower chance of containing
bugs.
You may have asked yourself how many tests should you write. As
mentioned before, you should at least write those that cover the most
common scenarios.
In general, you can think of this metric as follows:
Criterion
To measure, there are several coverage criterion that you may choose.
The most common are:
Function/method coverage: How many functions have been called?
Statement coverage: How many statements of each function have

been executed?
Branch coverage: Has each branch in an if or a when statement been

executed?
Condition coverage: Has each subcondition in an if statement been

evaluated to true and also to false?
For example, suppose that the following code is part of a feature of your
app:
fun getFullname(firstName: String?, lastName: String?): String {

var fullname = "Unknown"
if (firstName != null && lastName != null) {
fullname = "$firstName $lastName"
}
return fullname
}
Having at least one test that calls this function would satisfy the
function/method coverage criteria.
If you have a test that calls getFullname("Michael", "Smith") you would

satisfy the statement coverage criteria, because every statement would
be executed.
If you also have a test calling getFullname(null, "Smith"), now it complies

with branch coverage criteria, because the line inside the if is not
executed and the previous test that called getFullname("Michael",
"Smith") executes the line inside the if statement.
To satisfy the condition coverage criteria, you need tests that call
getFullname(null, "Smith") and getFullname("Michael", null) so that each
subcondition, firstName != null and lastName != null would evaluate to
true and false.
Tools
There are tools that can assist you to measure the test coverage metric.
JaCoCo (Java Code Coverage Library) is one of them. Don't worry, it

handles Kotlin as well!
This library generates a report for you to check which lines were covered
by your tests (green) and which ones were not (red or yellow).
Android Studio also comes with a built-in feature to run tests with
Coverage.
100% coverage?
In real-world apps, reaching a test coverage of 100%, no matter which
criterion you use, is almost impossible to achieve. It often doesn't add
value to test all methods, of all the classes, all of the time.
For example, suppose you have the following class:
data class Pet(var name: String)
You shouldn't write the following test:
fun whenCreatingPetWithName_shouldTheNameSetFromTheConstructor() {
val aName = "Rocky"
val aPet = Pet(aName)
if (aPet.name == aName) {
print("Success\n")
} else {
throw AssertionError("Invalid pet name")
}
}
In this case, you are testing a feature (getting and setting a property) of a
Kotlin data class that is auto-generated for you!
Test coverage gives you an exact metric of how much of your code has
not been tested. If you have a low measure, then you can be confident
that the code isn't well tested. The inverse however is not true. Having a
high measure is not sufficient to conclude that your code has been
thoroughly tested.
If you try to reach 100% test coverage, you'll find yourself writing
meaningless, low-quality tests for the sake of satisfying this goal.
Neither you nor any team member should be obsessed with a test
coverage of 100%. Instead, make sure you test the most common
scenarios and use this metric to find untested code that should be tested.
If you feel that writing a particular test is taking too long, you might
want to take a step back and evaluate if that test is adding enough value
to justify the effort. Also, if a simple fix is causing a lot of changes to
your tests, you may need to look at refactoring your tests or
implementation to make them less brittle.
At the end of the day, your goal is to create software that provides value
to its users. If you are doing TDD well, as your project gets larger, the
total amount of effort spent on tests, implementation and QA should be
the same or less than if you were creating the same product, with the
same level of quality without doing TDD. That said, a project that is
doing a good job at TDD may still take more development effort than a
project that is not because the project with TDD will have a higher level
of quality. The key is finding the right balance for your project.
Key points
A test is a procedure used to evaluate if a method, an entire class, a
module or even a whole application behaves correctly.
This book focuses on writing tests before implementing the features.
You should write tests to have confidence when refactoring.
Tests also act as complementary documentation of the application

features.
The tests you write should be short, simple to read and easy to
follow.
You should only write tests related to the logic of your application.
You can use test coverage tools to find untested code that should be
tested.

Congratulations! Now you should understand what a test is, why it
matters and the coverage metric.
In the next chapter, you'll find out what Test Driven Development (TDD)
is and what the benefits are of writing tests before writing the feature. In
the following chapters, you'll also start writing apps with their
corresponding tests.
Chapter 3: What Is TDD?
Once you have decided to add tests to your project, you
need to start thinking about how and when to integrate
them into your work. You want to make sure that you add
tests for any new code and the tests you add provide value.
You also want to consider how to add tests for existing code.
Finally, you want to feel confident that the tests you add
will catch any regressions in the future.
There are processes for different ways to incorporate tests

into your codebase, one of which is Test-Driven
Development, or TDD. In this chapter, you'll learn:
1. The basics of the TDD process.
2. Why TDD is useful.
3. The Red-Green-Refactor steps to practice TDD.
4. Some of the difficulties when learning TDD.
TDD is a process in which you write the tests for the code
you are going to add or modify before you write the actual
code. Because it's a process and not a library, you can apply
it to any project, be it Android, iOS, web or anything else.
There are a number of benefits to this that you'll learn

throughout this chapter and this book. Through using the
TDD process, you can be confident that any new code has
tests and that the code you write adheres to the
specification of the tests.
Why is TDD important?
There are plenty of reasons for using TDD as your testing
strategy, building upon the benefits of having tests in
general:
Write intentionally: Well-written tests provide a

description of what your code should do. From the
start, you will focus on the end result. Writing these
specifications as tests can keep the result from
deviating from the initial idea.
Automatically document: When coming to a piece of

code, you can look at the tests to help you understand
what the code does. Because these are tests — and,
again, a process — rather than a static document, you
can be sure that this form of documentation is likely
up-to-date.
Keep maintainable code: When practicing TDD, it

encourages you to pay attention to the structure of
your code. You will want to architect your app in a
testable way, which is generally cleaner and easier to
maintain and read. For example, decoupled classes are
easier to set up test classes for, encouraging you to
structure your classes this way. Refactoring is also built
into this development process. By having this
refactoring step built in, your code can stay squeaky
clean!
Have confidence in your code: Tests help you to ensure

that your code works the way it should. Because of this,
you can have greater confidence that what you wrote is
“complete.” In addition, with any changes that you
make in the future, you can know that you didn’t break
that functionality as long as the tests you wrote with
the code pass.
Develop faster: Using a process that promotes more

readable and maintainable code and that acts as self
documentation means you can spend less time trying
to understand what the code does when revisiting it,
and use that time for solving your problem instead.
Also, the code you write using the TDD process is less
error-prone from the start, so you will need to spend
less time on fixing them down the road.
Higher test coverage: If you’re writing tests alongside

with your code, you’re going to have more test coverage
over the code. This is important to many organizations
and developers.
Getting started
You'll start from scratch using pure Kotlin independent of
any framework to learn the steps of TDD. You're not looking
at Android or any testing tools here, so you can focus purely
on TDD.
Imagine, with the Wishlist app — an app in which you can

keep track of people's wishlists — that there's a way to
select an item and open it in a Google shopping search. For
example, say you have a wishlist for your friend, Giuseppe.
On this list you have a gift idea for them, "Tea sampler."
When you click on that item it would open up your browser
with a Google shopping search for "Tea sampler."
You'll write the tests, and then the code, for a helper
function that returns the Google shopping search URL to
open in the browser when a list item is tapped.
Start by opening https://play.kotlinlang.org/. This is where

you'll write your tests and code. Remove any code inside the
main() function so that you have a place to write your tests:
fun main() {
OK! You're ready to begin! First up: Red-Green-Refactor.
Practicing Red-Green-Refactor
In this chapter, you will learn the basics of the TDD process
while walking through the Red-Green-Refactor steps. Red-
Green-Refactor describes the steps that you follow when
practicing the TDD process.
Red: Start any new task by writing a failing (red) test. This is
for any new feature, behavior change, or bug fix that doesn't
already have a test for what you will be doing. You only
write the tests and the bare minimum code to make it
compile. Make sure you run the test, and see it fail.
Green: Write the minimum code to make the test pass

(green). You're not worried about making it pretty, or
adding any additional functionality at this step. Write the
needs for that test, then run the test to see it pass. Stay here
at this step until your test is green.
Refactor: Now is when you can prettify your code, and make
any other changes you want to make sure your code is clean.
You know that your changes are safe and correct as long as
your tests stay green. Having these tests to ensure the code
meets the specifications helps you do this refactoring with
confidence.
Red: writing a failing test

Your first test will test that your helper function,
getSearchUrl(), returns null when given a null query. Add
this code to the main() function. There will be a compiler
error at getSearchUrl() at first until you create that function:
// Test getSearchUrl returns null if query is null
// 1
val nullResult = getSearchUrl(null)
if (nullResult == null) {
// 2
print("Success\n")
} else {
// 3
throw AssertionError("Result was not null")
}
Here, you:
1. Call the getSearchUrl() function with null passed in,

storing the result in a variable.
2. If the result correctly equals null, print the success.
3. Otherwise, throw an error because it failed.
You need this code to compile in order to run your test, so

add this empty function below your main() function:
fun getSearchUrl(query: String?) { }
This is as empty as you can write this function while still

allowing it to compile. You may be wondering why you
shouldn't add return null to the body now. Isn't that the
same as writing an empty function? The short answer is,
"No."
By including that in the function, the test would pass right

away. If there's one thing to never forget about TDD it's that
you always want to see your tests fail. By seeing it fail, you
have confidence that the test is correctly asserting what it
should. If you never see the test fail, the test might not be
checking for the right thing! If that's the case, it won't fail if
something is actually wrong. You'll learn more about this in
the False Positives section of this chapter.
Now that you have eliminated any compiler errors, click the
Run button to see your test fail!
Great! You have a failing test... now to make it pass.
Green: making your test pass

To make the test pass, the function needs to return null.
Make sure the return type is String?, and add return null to
the body of getSearchUrl(). Your function should now look
like this:
fun getSearchUrl(query: String?): String? {

return null
}
Run the test again and see the success message.

Congratulations on your first test with TDD!
Writing a second test

You may not need a refactoring step near the start, as the
code is still simple. In this case, before moving to the
refactor step, write one more test to practice the first two
Red and Green steps.This test is to make sure getSearchUrl()
returns a non-null value if a non-null String is passed in as
the query.
Add the following test to the bottom of the main() function:
// Test getSearchUrl returns not null if query is a string

// 1
val nonNullResult = getSearchUrl("toaster")
if (nonNullResult != null) {
// 2
print("Success\n")
} else {
// 3
throw AssertionError("Result was null")
}
This test is very similar to the one you wrote before. You:
1. Call the getSearchUrl() function with a String passed in,

2. If the result correctly does not equal null, print the

success.
Run the code again. You will see your first test pass, while
your second, new test fails.
Making it pass
Now, write the minimum amount of code to make your new
test pass. Change the body of getSearchUrl() to return the
query String.
return query
Run the tests, and see them both pass.
False positives
While testing saves you from many pitfalls, there are things
that can go wrong as well, especially when you don't write
the test first. One of these is the false positive. This
happens when you have a test that is passing, but really
shouldn't be.
To see this, next, test that the URL returned includes the
search query. Break the rules a little bit, and add this line to
the top of getSearchUrl():
val url = "https://www.google.com/search?q=$query"
Run your tests to make sure they still pass.
Note: Running your tests often is a habit you want to

acquire.
Now, add this test to make sure the result contains the
query:
// Test getSearchUrl result contains query
// 1
val result = getSearchUrl("toaster")
if (result?.contains("toaster") == true) {
// 2
print("Success\n")
} else {
// 3
throw AssertionError("Result did not contain query")
}
Again, you see this familiar pattern:
1. Call the getSearchUrl() function with a String passed in,

2. If the result correctly contains the query, print the

success.
Run the tests. They pass right away!
Correcting the mistake

Did you catch what is missing? At the start, the desire was
to test the returned URL contains the query. If you look
closely at getSearchUrl(), you'll notice that the URL is never
returned! It's the query that the function returns. The test is
not asserting what you want to test, because it isn't specific
enough.
As a challenge, you can write a test to check for the URL

portion of the result. For now, refactor the function to
return the URL instead of just the query. The getSearchUrl()
should now look like this:

return "https://www.google.com/search?q=$query"
}
Run the tests, again.
Oh, no! You can't see the result of the new test because the
first one (as you probably predicted) is failing now! If you
were using a testing framework such as JUnit, as you will
starting in Chapter 5, "Unit Tests," all the tests would run
even if some failed. When JUnit continues running the rest
of your tests like this, it helps you to have a comprehensive
understanding of what is broken. This is because your well
defined tests will tell you which parts of your code are
working, and which are not. What you know now looking at
these results is that the function is no longer returning null
if the input is null. This mean you need to update the
function so it returns null again when given null as a
parameter while ensuring your newest test is passing. That's
next!
Refactor: Updating your code

Now is your chance to refactor the code. You want to change
the code to make it better, and fix the failing test, while
making sure the rest of the tests pass.
Change the body of getSearchUrl() to the following:
return query?.let {
"https://www.google.com/search?q=$query"
}
This should now return null if the query is null, and the
URL containing the query otherwise. Run the tests and see
them pass to confirm this!
TDD takes practice

While this was a simple example, it shows that there can be
some tricky things that show up when testing. It takes
practice to learn what works and form good habits. TDD is
an art, and to do it well it takes a lot of time. It is hard, but
worth it.
When starting out with TDD, you will make many mistakes.
You might change a small piece of functionality that will
make half your tests break when the functionality is correct,
for example. Or you'll spot false positives. By making these
mistakes, you'll grow and learn how to write better tests and
better code.
Most projects have too few tests. As you're starting with

TDD a good idea is to always add tests, and when in doubt,
over test. As you learn and become more confident, you can
eliminate tests that are not valuable. Tests are another part
of the code you need to maintain, so you can develop
judgement about which ones are valuable to continue
maintaining.
Key points
TDD is a process of writing tests before you write your
actual code. You can apply TDD to any sort of
programming.
Practicing TDD helps you to code quickly and with

intent, document automatically, have confidence your
code is maintainable, and have better test coverage.
TDD follows the Red-Green-Refactor steps.
This process always starts with writing a failing test. No

matter what, you always want to see the test fail.
Only after you write your test and see it fail do you
write your new code or change your existing code.
You only ever write enough code to make your test

pass. If there's more code you need to write, you need
another test first.
You follow up with refactoring to make your code clean

and readable.
Learning to write good tests takes practice.

Congrats! You now know how to use TDD with the Red-
Green-Refactor steps, and how to apply them to building a
simple function. You can find the finished code for this
example in the materials for this chapter.
If you want to keep practicing, you can add a test to make

sure the URL portion of the search URL is returned. The
solution is also included in the materials for this chapter.
Revert the final two changes you made to the function first
so you can watch it fail. Before you start writing this test,
the getSearchUrl() function should look like this:

val url = "https://www.google.com/search?q=$query"
return query
}
You'll follow this pattern of Red-Green-Refactor for the rest

of the book. If you want more practice, you can check out
our helpful online tutorial, "Test-Driven Development
Tutorial for Android: Getting Started," found here:
https://www.raywenderlich.com/7109-test-driven-
development-tutorial-for-android-getting-started
Section II: Testing on a
New Project
This section dives deep into the art of TDD.
You’ll learn about the different types of tests
that make up the testing pyramid and how to
implement all of these different kinds of tests
into an app.
You’ll learn how to write unit tests, integration

tests, and using UI testing to verify your app
works as expected end to end.

chapters:
Chapter 4: The Testing Pyramid: If you're

wondering how UI tests, integration tests and
unit tests all fit into one application, learn
about the testing pyramid and how you can
structure various types of tests for your app.
Chapter 5: Unit Tests: Get your drink ready as

you explore a cocktail app and write your unit
tests in Android practicing test-driven
development.
Chapter 6: Architecting for Testing: The

architecture of your project can make or break
your testing experience. Learn the pros and cons
of each architecture and how they affect your
tests.
Chapter 7: Introduction to Mockito: Level up

your testing knowledge leveraging Mockito as
you learn the basics of how to use mocks and
spies in your tests.
Chapter 8: Integration: Often, you're going to

have to test the interaction between objects in
your application. In this chapter, you'll practice
writing integration tests by working on a
Wishlist app.
Chapter 9: Testing the Persistence Layer: Get

started learning how to test the persistence
layer in your app. In this chapter, you'll learn
how to handle statefulness in your tests and
strategies for creating randomized test data.
Chapter 10: Testing the Network Layer: While
HTPP requests can be unpredictable, your tests
don't have to be. In this chapter, learn how to
write predictable network tests working on the
Punchline app. Learn some random jokes on the
way as well!
Chapter 11: User Interface: UI tests allow you to

test your app end-to-end without having to
manually click-test your app. Learn the
fundamentals of UI tests using the Espresso
library.
Chapter 4: The Testing Pyramid
Traditionally, software testing was done manually. It consisted
of deploying the application to a test environment similar to
the real environment but with fake (test) data. The Quality
Assurance (QA) members would perform black-box testing on
it — without knowing the internals — and raise bug tickets.
Then, the developers went back and fixed the bugs.
Even nowadays, without any kind of automation, this is

happening on the Android ecosystem. Testing usually consists
of compiling the release candidate application, installing it on
a physical device or emulator, and passing it to the QA team.
The QA members would then follow a test plan, executing their
use cases manually to find bugs.
You'll find that automating these repetitive use cases is the

way to go. For each use case, you can write one or more
automated tests. However, first, you need to understand that
there are different kind of tests and how to classify them.
Tests are typically broken into three different kinds:

This is the testing pyramid, a concept originally explained by
Mike Cohn in his book Succeeding with Agile. The testing
pyramid gives you a way to group different types of tests and
gives an understanding of how many tests you should consider
on each layer.
You should have lots of small unit tests, some integration and
fewer UI tests.
You'll now go through each of the layers.
Unit tests
Unit tests are the quickest, easiest to write and cheapest to
run. They generally test one outcome of one method at a time.
They are independent of the Android framework.
The System Under Test (SUT) is one class and you focus only
on it. All dependencies are considered to be working correctly
— and ideally have their own unit tests — so they are mocked
or stubbed. This way, you have complete control of how the
dependencies behave during the test.
These tests are the fastest and least expensive tests you can
write because they don’t require a device or emulator to run.
They are also called small tests. To give an example of an unit
test, consider a game app.
The Game class is one of the main classes.

A common use case is to increment the score using a function
like incrementScore(). Whenever the score increments and
exceeds the highscore, it should also increment the highscore. A
simple and incomplete definition of the Game class can look like
this:
class Game() {
var score = 0
private set
var highScore = 0
private set
fun incrementScore() {
// Increment score and highscore when needed
}
}
Therefore, a test that checks this could be as follows:
fun shouldIncrementHighScore_whenIncrementingScore() {
val game = Game()
game.incrementScore()
if (game.highScore == 1) {
print("Success")
} else {
throw AssertionError("Score and HighScore don't match")
}
}
If you run this test, you'll see the test doesn't pass. We now
have our failing (red) test. You can then fix this to get our
passing (green) test by writing the actual method for the Game
class:
score++
if (score > highScore) {
highScore++
}
}
Some common libraries for unit testing are JUnit and Mockito.
You'll explore both of these in later chapters.
Google, in its testing fundamentals documentation, also

suggests Robolectric for local unit tests. Robolectric simulates
the Android runtime, it allows you to test code that depends on
the framework without using a device or emulator. This means
that these tests run fast because they run using just the regular
JVM of your computer, just like any other test that uses JUnit
and Mockito. However, some may consider Robolectric as an
integration testing tool, because it helps you test integrating
with the Android framework.
Integration tests
Integration tests move beyond isolated elements and begin
testing how things work together. You write these type of tests
when you need to check how your code interacts with other
parts of the Android framework or external libraries. Usually,
you'll need to integrate with a database, filesystem, network
calls, device sensors, etc. These tests may or may not require a
device or emulator to run; they are a bit slower than unit tests.
They are also called medium tests.
For a simple example of an integration test, think about a

Repository class that depends on a JSON parser class that reads
from a file. The repository asks the parser to retrieve the data.
Then the repository transforms the data to your domain
model. You could create a test that given a JSON file verifies
that the repository correctly returns the domain data. You
would be testing the integration between the repository and
the JSON parser.
Note: If you mock the JSON parser and verify only the
transformation to your domain model you would be
creating a unit test. You should create unit tests for both
the repository and also the JSON parser to ensure they
work as expected in isolation. Then, you can create
integration tests to verify they work together correctly.
Another example could be found in a retail app. You could

ensure that the LoginActivity is launched whenever the user
wants to add a favorite but hasn't signed into the app yet.
The test could look like this:

fun shouldLaunchLogin_whenAddingFavorite() {
// 1
val user: User = null
val detailActivity = createProductDetailActivity(user)
detailActivity.findViewById(R.id.addFavorite).performClick()
// 2
val expectedIntent = Intent(detailActivity,
LoginActivity::class.java);
// 3
val actualIntent = getNextStartedActivity()
if (expectedIntent == actualIntent) {
print("Success")
} else {
throw AssertionError("LoginActivity wasn't launched")
}
}
Here's what this test does:
1. Creates a new "details" activity, finds the favorites button

and clicks on it.
2. Creates the expected result: an intent to navigate to the

login screen.
3. Checks if the activity that launched is the same as the one

expected to be launched.
Although this test deals with activities, you'll see that it

doesn't require a screen for them to be rendered.
Another example of integration tests: In a social app, you

could check if the list of friends is retrieved correctly from a
REST API. Because this automated test will run frequently, you
shouldn't use the real production API. Usually, it is replaced
with a local or fake testing server. This is to avoid using server
quota and because the tests shouldn't alter any production
values. This will also ensure that the tests are repeatable.
You can still use JUnit and Mockito to create integration tests
to verify state and behavior of a class and its dependencies.
You'll focus on these kind of tests in later chapters. You can
also use Robolectric for tests involving the Android framework
but run locally without a device or emulator.
Google, in its testing fundamentals documentation, also

suggests Espresso for medium tests. For example, to perform
validation and stubbing of intents or to perform actions on
view objects. However, some may consider these kind of tests
as UI tests because you would be interacting with UI elements.
UI tests
Finally, every Android app has a User Interface (UI) with
related testing. The tests on this layer check if the UI of your
application works correctly. They usually test if the data is
shown correctly to the user, if the UI reacts correctly when the
user inputs something, etc. They are also called large tests.
These tests emulate the user behavior and assert UI results.

These are the slowest and most expensive tests you can write if
you run them on a device or emulator. While these are helpful
for testing your interactions on screen, you should limit your
UI tests. As in the pyramid diagram, you should perform your
tests with unit and integration tests as much as you can.
For a test example of this layer, think of an app with a login

screen. You'd like to check that, after logging in, the app shows
a TextView welcoming the user.
So, the UI test could look like this:

fun shouldWelcomeUser_whenLogin() {
onView(withId(R.id.username)).perform(typeText("Fernando"))
onView(withId(R.id.password)).perform(typeText("password"))
onView(withId(R.id.login_button)).perform(click())
onView(withText("Hello Fernando!"))
.check(matches(isDisplayed()))
}
On Android, a tool suggested by Google for UI testing is

Espresso. You'll write these kinds of test in later chapters. You
could also use Robolectric (since version 4.0) for UI tests to run
them without an emulator or a device. However, there are
times that you need to run them on an emulator or device.
These are called Android instrumentation tests.
There's another tool called UI Automator. Google recommends

it only when you have to do cross-app functional UI testing
across system and installed apps.
Distributing the tests

A typical rule of thumb is to have the following ratio among
the categories:
UI Tests: 10%
Integration Tests: 20%
Unit Tests: 70%
Google, in its testing fundamentals documentation, suggests

these percentages. This doesn't have to be absolutely exact,
but it's important to retain the pyramid shape. Remember that
tests in the lower layers are easier to maintain and run faster.
So you should avoid the following anti-patterns:
Ice cream cone or Inverted pyramid: The team is relying
on lots of UI tests, having less integration tests and yet
fewer unit tests. This type is commonly found in
organizations that don't have a testing culture, or are not
encouraging developers to create tests. Usually, a QA team
is responsible for testing and, in many cases, they don't
even have access to the code repository. In this case, if
developers don't create unit and integration tests, no one
will. As a result, the QA team will try to compensate by
creating tests of the upper layers, forming this anti-
pattern.
Hourglass: You start writing lots of unit tests, you don't

care that much about integration tests and you write many
UI tests.
Bear in mind that not following the correct pyramid shape
could affect productivity. This is because the test suite will run
slower, thus, taking a longer time to provide feedback to the
developers.
Key points
Testing is commonly organized into the testing pyramid.
There are three kinds of tests in this pyramid: unit,

integration and UI tests. These are also called small,
medium and large tests, respectively.
On Android, you can also distinguish between local tests,

which run on the JVM and instrumentation tests, which
require a device or emulator. Local tests run faster than
instrumented tests.
You'll write tests of different granularity for different

purposes.
The further down you get, the more focused and the more
tests you need to write, be mindful of how expensive the
test is to perform.

In the following chapters, you'll start doing TDD by writing
each kind of test with the appropriate tools and libraries.
If you want to go deeper on this subject, check the following:
Testing Fundamentals:
https://developer.android.com/training/testing/fundamen
tals#testing-pyramid
Unit Tests: https://www.youtube.com/watch?

v=pK7W5npkhho&start=111
Integration Tests: https://www.youtube.com/watch?

UI Tests: https://www.youtube.com/watch?
The Practical Test Pyramid:

https://martinfowler.com/articles/practical-test-
pyramid.html
Google Testing Blog - Just Say No to More End-to-End

Tests: https://testing.googleblog.com/2015/04/just-say-
no-to-more-end-to-end-tests.html
Succeeding with Agile book:

https://www.amazon.com/Succeeding-Agile-Software-
Development-Using/dp/0321579364
Chapter 5: Unit Tests
As mentioned in Chapter 4, "The Testing Pyramid," unit tests verify how
isolated parts of your application work. Before checking how things work
together, you need to make sure the units of your application behave as
expected.
In this chapter, you'll:
Learn what unit tests are and what are the best places to use them.
Write unit tests using the test-driven development (TDD) pattern to

learn these concepts in the context of TDD.
Throughout this chapter and Chapter 7, "Introduction to Mockito" you'll

work on an application named Cocktail Game. With this application,
you'll have fun with a trivia game about cocktails.
Find the starter project for this application in the materials for this
chapter and open it in Android Studio. Build and run the application.
You'll see a blank screen.
You'll start writing tests and classes for the application and, by the end of
Chapter 7, "Introduction to Mockito," the application will look like this:
When to use unit tests
Unit tests are the fastest and easiest tests to write. They also are the
quickest to run. When you want to ensure that a class or method is
working as intended in isolation — this means with no other dependent
classes — you write unit tests.
Before writing any feature code, you should first write a unit test for one
of the classes that will compose your feature. Afterwards, you write the
class that will pass the test. After repeating this procedure, you'll have a
completed, testable feature.
Setting up JUnit
You're going to write a unit test for the first class of the cocktail game,
which is a Game class. This first test will be a JUnit test, so, open
app/build.gradle and add the following dependency:
dependencies {
...
testImplementation 'junit:junit:4.12'
}
Notice that it's testImplementation instead of implementation because you'll

use this dependency only when testing. This means that it won't be
bundled into the application (APK) that your device or emulator will run.
Note: When creating a new project, you'll find that this dependency
is already there. You're adding it here manually for educational
purposes.
Creating unit tests

To start, switch to the Project View and open app ‣ src. Create a new
directory and enter:
test/java/com/raywenderlich/android/cocktails/game/model. Then,
create a file called GameUnitTests.kt.
Write the following code:
class GameUnitTests {
// 1
@Test
// 2
val game = Game()
// 3
// 4
Assert.assertEquals(1, game.currentScore)
}
}
Note: When importing Assert, you should choose org.junit.Assert.

1. Notice the @Test annotation. This will tell JUnit that this method is a
test.
2. Create an instance of the Game class — the one that will be tested.
3. Call the method that you want to test.
4. assertEquals verifies that the previous execution modified the

game.currentScore property to be equal to one. It's important to
understand that the first parameter is the expected value, and the
second parameter is the actual value.
There's also the possibility to write a message so that, when the test fails,
you'll see this message. For example:
Assert.assertEquals("Current score should have been 1",

1, game.currentScore)
Every test has the following steps:
Set Up: You first have a phase where you arrange, configure or set
up; in this case, you instantiate a class.
Assertion: You execute the method that you want to test and you
assert the result.
Teardown: Sometimes (not in this example) there are items in the

setup step that need to be reset after the tests are done running.
Here is where this would happen.
If you try to compile the test now, you'll get this:

Making the test compile
The test won't compile because the Game class doesn't exist. So, create the
Game class under the directory app ‣ src ‣ main ‣ java ‣ com ‣
raywenderlich ‣ android ‣ cocktails ‣ game ‣ model. You'll need to create
game and model packages first. In the Game class, write the minimum
amount of code to make the test compile:
class Game() {
var currentScore = 0
private set
// No implementation yet
}
}
Running the test

Now, go back to the test; you'll see that it compiles.
Run the test. There are several ways to run the tests.
You can click the Play button over a test:
You can also use the shortcut ^ + ⇧ + R.
Or, if you want to run all the tests (currently you have just one), you can
right-click over the app ‣ src ‣ test ‣ java ‣ com ‣ raywenderlich ‣ android ‣
cocktails ‣ game ‣ model package and select Run 'Tests' in 'model':
Either way, you should see that it doesn't pass:
This is because you didn't increment the current score yet. You'll fix that
soon.
You can also open the Terminal going to View ‣ Tool Windows ‣ Terminal
and run the tests from the command line executing:
$ ./gradlew test
Notice how the expected value is one and the actual value is zero. If we
had reversed the order of our expected and actual values in our assertion,
this would show up incorrectly. You'll also see that it generates a report
under /app/build/reports/tests/testDebugUnitTest/index.html; if you
open it in your preferred browser, you'll see the following:
Making the test pass

Modify the Game class to make it pass:
class Game() {
var currentScore = 0
private set
currentScore++
}
}
Now run the test again and see that it passes.
Or if you run the command in the Terminal:
$ ./gradlew test
It'll generate this report:
Creating more tests

The game will show a highest score. So, you should add a test that checks
that when the current score is above the highest score, it increments the
highest score:
@Test
val game = Game()
Assert.assertEquals(1, game.highestScore)
}
Again if you try to compile it'll fail because the highestScore property is
missing.
So, add the following property to the Game class:
var highestScore = 0
private set
Now the test will compile, so run it and watch it fail.

To make it pass, open the Game class and modify the incrementScore()
method as follows:
currentScore++
highestScore++
}
Run the test and you'll see that it passes.
However, you should also test that, when the highest score is greater
than the current score, incrementing the current score won't also
increment the highest score, so add the following test:
@Test
fun whenIncrementingScore_belowHighScore_shouldNotIncrementHighScore(
val game = Game(10)
}
Here, the intention is to create a Game with a highscore of 10. The test
won't compile because you need to modify the constructor to allow a
parameter. Because you need to start with a highest score greater than
the default, which is 0, you need to alter the constructor like this:
class Game(highest: Int = 0) {
And change the highestScore property to be set to highest:
var highestScore = highest

private set
Now, run all the tests and see that the last one doesn't pass. You can use
the green arrow button on the left-side of the class definition.
The last one doesn't pass because you're incrementing both the current
score and highest score regardless of their values. Fix that by replacing
the incrementScore() function with the following:
currentScore++
if (currentScore > highestScore) {
highestScore = currentScore
}
}
Build and run the last test to see the satisfying green checkmark.
JUnit annotations
For this project, you're creating a trivia game. Trivias have questions, so
you'll now create unit tests that model a question with two possible
answers. The question also has an "answered" option to model what the
user has answered to the question. Create a file called
QuestionUnitTests.kt in the app ‣ src ‣ test ‣ java ‣ com ‣ raywenderlich ‣
android ‣ cocktails ‣ game ‣ model directory.
Add the following code:

class QuestionUnitTests {
@Test
fun whenCreatingQuestion_shouldNotHaveAnsweredOption() {
val question = Question("CORRECT", "INCORRECT")
Assert.assertNull(question.answeredOption)
}
}
Here, you used assertNull to check if question.answeredOption is null.
If you try to run this test it won't compile because the Question class
doesn't exist. So, create the Question class under the directory app ‣ src ‣
main ‣ java ‣ com ‣ raywenderlich ‣ android ‣ cocktails ‣ game ‣ model
and add the following to make it compile:
class Question(val correctOption: String,

val incorrectOption: String) {
var answeredOption: String? = "MY ANSWER"
private set
}
Run the test again and watch it fail.
It failed because you hardcoded "MY ANSWER" which is not null.
So, modify the Question class to the following:

val incorrectOption: String) {
var answeredOption: String? = null
private set
}
Run the test again and watch that it now passes.

Now, you can add another test:
@Test
fun whenAnswering_shouldHaveAnsweredOption() {
question.answer("INCORRECT")
Assert.assertEquals("INCORRECT", question.answeredOption)
}
This test will check that, when you add the user's answer to a question,
the user's answer is saved in the answeredOption property.
You'll get a compilation error since you haven't written the answer()
method yet. Add the following to the Question class to make it compile:
fun answer(option: String) {

// No implementation yet
}
Now run the test and you'll see that it doesn't pass.
So add the following to the answer() method:
fun answer(option: String) {

answeredOption = option
}
Run it and watch that it passes.
Because you'll need to know if the question was answered correctly,

imagine that the answer() method now returns a Boolean. The result
would be true when the user answered correctly. Now, add this test:
@Test
fun whenAnswering_withCorrectOption_shouldReturnTrue() {
val result = question.answer("CORRECT")
Assert.assertTrue(result)
}
Notice, here, that you're using assertTrue. It checks for a Boolean result.
Running this test will get you a compilation error since the answer()
method doesn't return a Boolean. So, modify the Question class so that the
answer() method returns a Boolean. For now, always return false:
fun answer(option: String): Boolean {

return false
}
Run it and watch it fail.
Fix it temporarily by always returning true:

return true
}
Run it and watch it pass.
Add the following test:

@Test
fun whenAnswering_withIncorrectOption_shouldReturnFalse() {
val result = question.answer("INCORRECT")
Assert.assertFalse(result)
}
Run it and see that it fails.
Now that we have tests for when the answer is correct and when the
answer is not correct, we can fix the code:

return correctOption == answeredOption

}
Run all the Question tests and verify they all pass correctly.
Finally, you should ensure that the answer() method only allows valid
options. Add this test:
@Test(expected = IllegalArgumentException::class)
fun whenAnswering_withInvalidOption_shouldThrowException() {
question.answer("INVALID")
}
Notice, here, that the @Test annotation allows to expect an exception. If
that exception occurs, the test will pass. This will save you from writing
try/catch. If you run the test now, it will fail because the answer() method
doesn't throw the exception:
To fix this, modify the Question class as follows:

if (option != correctOption && option != incorrectOption)
throw IllegalArgumentException("Not a valid option")
return correctOption == answeredOption

}
Run the test and watch that it now passes.
Because later you'll need a property isAnsweredCorrectly, open the

Question class and refactor to the following:
val isAnsweredCorrectly: Boolean

get() = correctOption == answeredOption

if (option != correctOption && option != incorrectOption)
throw IllegalArgumentException("Not a valid option")
return isAnsweredCorrectly
}
Run all the tests again to see that everything is still working after the
refactor.
Refactoring the unit tests
Notice that each test repeats this line of code:
This makes the tests bloated with boilerplate code that makes them hard
to read. To improve this, JUnit tests can have a method annotated with
@Before. This method will be executed before each test and it's a good
place to set up objects.
Modify the QuestionUnitTests test class, adding the following to the top:
private lateinit var question: Question
@Before
fun setup() {
question = Question("CORRECT", "INCORRECT")
}
And remove the repeated line of each test:

@Test
fun whenCreatingQuestion_shouldNotHaveAnsweredOption() {
Assert.assertNull(question.answeredOption)
}
@Test
fun whenAnswering_shouldHaveAnsweredOption() {
question.answer("INCORRECT")
Assert.assertEquals("INCORRECT", question.answeredOption)
}
@Test
fun whenAnswering_withCorrectOption_shouldReturnTrue() {
val result = question.answer("CORRECT")
Assert.assertTrue(result)
}
@Test
fun whenAnswering_withIncorrectOption_shouldReturnFalse() {
val result = question.answer("INCORRECT")
Assert.assertFalse(result)
}
@Test(expected = IllegalArgumentException::class)
fun whenAnswering_withInvalidOption_shouldThrowException() {
question.answer("INVALID")
}
Now, run all the tests again to make sure you didn't break them while
refactoring. All tests still pass — great!
JUnit also has other similar annotations:
@After: The method will be executed after each test. You can use it
to tear down anything that you set up in @Before.
@BeforeClass: If you annotate a method with this, it'll be executed

only once before all the tests are executed. For example, opening a
file, a connection or a database that is shared in all the tests.
@AfterClass: To execute a method only once after all the tests are
executed, use this one. For example, closing a file, a connection or a
database that is shared in all the tests.
Challenge
Challenge: Testing questions
You have the Game and Question classes. The Game class should contain a
list of questions. For now, these are the requirements:
The game should have a list of questions; so, when getting the next
question, the game should return the first.
When getting the next question, without more questions, the game
should return null.
The question should have a getOptions method that returns the

correct and incorrect options as a shuffled list, so later you can show
them as Buttons. Hint: The method should receive a lambda
parameter to sort the list, by default it should be { it.shuffled() }
but having a parameter will let you use another one in your test.
Write a test for each one and add the corresponding functionality to the
Game class progressively to make each test pass.
Remember the TDD procedure: write a test, see it fail, write the
minimum amount of code to make it pass and refactor if needed.
Key points
Unit tests verify how isolated parts of your application work.
Using JUnit, you can write unit tests asserting results, meaning, you
can compare an expected result with the actual one.
Every test has three phases: set up, assertion and teardown.
In TDD, you start by writing a test. You then write the code to make
the test compile. Next you see that the test fails. Finally, you add the
implementation to the method under test to make it pass.

Great! You've just learned the basics of unit testing with JUnit. You can
check the project materials for the final version of the code for this
chapter.
In the next chapter, "Architecting for Testing," you'll learn about good
practices and design patterns, that will ensure a good architecture and
encourage testability. Afterwards, you'll continue working on this
project, creating unit tests using a complementary library called
Mockito.
For additional resources, there's a Google library you can use called
Truth, similar to JUnit. It has a couple of notable benefits:
More readable test assertions
Default failure messages
You can check it out, here: https://google.github.io/truth/comparison

Chapter 6: Architecting for Testing
Software architecture is a template or blueprint that you can use when
building new apps. It defines software elements and their relations.
When you create a new app, you need to make some fundamental
structural decisions — and those decisions may be difficult to modify
once they’re implemented.
In this chapter, you’ll focus on what it takes to architect an app for

testability; specifically, you’ll:
Learn the characteristics of a testable architecture.
Discover good practices to create a testable architecture.
How does architecture matter?

To understand why architecture matters, it’s essential first to understand
what qualifies as a poorly architected app.
A poorly architected app may have all of its logic contained within a
single method that consists of many lines; or it may have one large class
with too many responsibilities. Both of these scenarios make it
impossible to test groups or units of logic independently.
Apps that are architected for testing separate their code into groups of
logic using multiple methods and classes to collaborate. With this type of
architecture, developers can test each public method and class in
isolation.
You also need to consider the effort it takes when adding or modifying an
app’s features. In TDD, this process starts with creating new tests or
modifying existing ones. While it may take some additional time to do
this, adding and updating tests shouldn’t be a painful process. If it is,
you’ll eventually stop writing tests and avoid TDD all together. To
encourage TDD, it’s better to think of a software architecture that
encourages and facilitates the creation of tests.
But it’s not only testing that matters:
Communication: Software architecture establishes a common

language between the developers of an app and other members of
the team, like managers, QA testers, analysts and designers.
Reusable abstraction: Reusability saves time. Later in the chapter,

you’ll see that you can reuse patterns within different parts of an
app, across different apps as well as on other platforms. You’ll also
see that you can use architecture patterns to kick-off new projects.
Early design decisions: When you create a new app, one of the first
decisions is to decide on the architecture you’re going to use. These
early design decisions are important because they’ll set constraints
on your implementation, such as the way your classes will interact
and their responsibilities. Early decisions will also organize your
codebase a specific way and may even organize the members of your
team. For example, on a given architecture, you may divide your
team between people who only write domain classes and others who
only write visually-related code.
Better testing: By using good architecture from the start or

refactoring an existing one, you’ll enable the creation of tests that
would otherwise be impossible or difficult to write. Also, migrating
from an existing architecture to a better one — which is a difficult
task, but not impossible — will enable you to migrate slower tests,
such as UI or integration tests, to unit tests, which are faster.
To achieve a robust architecture, it’s important to know and understand

design patterns and the SOLID principles.
Design patterns
It's not uncommon for developers to encounter the same problems in
different projects and platforms, and to solve these problems using
similar solutions. Over time, certain developers started formalizing these
patterns into templates or solutions that other developers could reuse if
they found themselves in a similar context or situation.
Most of the time, these solutions are not specific blocks of code for a
specific platform. Instead, they’re diagrams, ideas and descriptions of
how to proceed when faced with similar circumstances. They tend to
show relationships and collaboration between classes. When reviewed
carefully, you’re able to then take this information and implement
solutions in your own language and platform.
These battle-tested and proven patterns can help speed up your

development process. Using design patterns, which is what this is known
as, you avoid searching for solutions that other developers have already
solved.
Design patterns are also useful to communicate about software ideas

between developers and to help with code readability.
According to the Gang of Four (GoF: Erich Gamma, Richard Helm, Ralph
Johnson and John Vlissides) you can classify design patterns into the
following categories: creational, structural and behavioral.
Creational
The patterns in the Creational category describe solutions related to
object creation.
Singleton
The Singleton design pattern specifies that only one instance of a certain
class may exist, known as a singleton. Usually, it’s possible to access the
singleton globally.
Kotlin has the object keyword to declare a singleton:
object MySingleton {
private var status = false
private var myString = "Hello"
fun validate(): Boolean {

...
}
...
}
You can use MySingleton by invoking:
MySingleton.validate()
This line creates the MySingleton object. If the object already exists, it
uses the existing one, so there’s no worry about creating more than one.
Although this might be the easiest pattern to understand and

implement, it’s important to use caution when using it. For instance, if
you have an object that collaborates with MySingleton, like this:
class MyClass {
fun methodA() {
...
if (MySingleton.validate()) {
...
}
...
}
}
You won’t be able to test methodA() properly because you’re using the
actual MySingleton object, which means you can’t force validate() to
return true or false.
Later, you’ll learn about a pattern known as dependency injection to deal

with this problem. In terms of testing, you should be able to identify a
singleton and flag it as code that might need refactoring to test.
Builder
The Builder design pattern abstracts the construction of a complex
object, joining several parts. For example, think of a restaurant that
serves different menus depending on the day.
You might have the following abstract class:
abstract class MenuBuilder {

var menu = Menu()
abstract fun buildMainDish()
abstract fun buildDessert()
}
You would then implement the builders depending on the day:
class DayOneMenuBuilder: MenuBuilder() {

override fun buildMainDish() {
// Add day one main dish to the menu
}
override fun buildDessert() {
// Add day one desert to the menu
}
}
class DayTwoMenuBuilder: MenuBuilder() {

...
}
You might also have the Chef class:
class Chef {
fun createMenu(builder: MenuBuilder): Menu {
builder.buildMainDish()
builder.buildDessert()
return builder.menu
}
}
Notice how Chef calls the corresponding methods to build the menu.
This lets you create the menu as follows:
val chef = Chef()

val menuBuilder = getDayMenuBuilder()
val menu = chef.createMenu(menuBuilder)
In this example, getDayMenuBuilder() returns the corresponding

MenuBuilder depending on the day.
With this implementation, it’s easy to test the separated parts. You can
test Chef to verify that it calls the right methods, and you can also test
each class that inherits from MenuBuilder by asserting the state of the
resulting Menu. You’ll see how to perform this kind of test in the next
chapter.
Effective Java by Joshua Bloch, introduced another Builder design

pattern, focused on readability. The AlertDialog.Builder in Android is an
example of this type of pattern:
AlertDialog.Builder(this)
.setTitle("Error!")
.setMessage("There was an error, would you like to retry?")
.setNegativeButton("Cancel", { dialogInterface, i ->
...
})
.setPositiveButton("Retry", { dialogInterface, i ->
...
})
.show()
The Builder design pattern is useful to avoid an anti-pattern known as a

Telescoping Constructor. A Telescoping Constructor consists of a
constructor with many parameters where some of them are optional.
This is not an issue with Kotlin where you can have default and named
parameters.
Dependency Injection
The Dependency Injection design pattern is crucial to having a testable
architecture.
The following is an example of a class that does not use dependency

injection. Usually, objects collaborate with other objects. For example:
class Vehicle() {
private val engine = CombustionEngine()
fun start(): Boolean {

...
return engine.start()
}
...
}
When you create a Vehicle, it creates, internally, a CombustionEngine. You

can then use it later to perform some operation. This makes it difficult to
properly test Vehicle because, for example, you won’t be able to test what
happens when the engine brakes or when the engine is used for too many
hours.
And what happens when you want to use an ElectricEngine instead of a
CombustionEngine?
You can solve these types of problems using the dependency injection
design pattern, which describes that collaborators are provided to an
object that requires them, instead of this object directly instantiating
them internally.
There are two ways to inject dependencies.
The first way is using Constructor Injection, like so:
class Vehicle(val engine: Engine) {

...
}
...
}
In this example, to create a Vehicle you need an Engine. Here, Engine is an

interface or an abstract class which lets you inject any implementation
(CombustionEngine or ElectricEngine) provided it complies with the
interface or abstract class. So, the creator of the Vehicle provides the
proper Engine.
If you combine the Builder design pattern with the dependency injection
design pattern, you end up with something like this:
class CombustionVehicleBuilder {
fun build(): Vehicle {
val engine = CombustionVehicleEngine()
...
return Vehicle(engine)
}
}
In this example, you aren’t injecting the engine here, so you may also
want to inject the engine to the builder. You could do that. However, at
some point someone or something needs to instantiate the class.
Usually, it's the entity that creates objects and provides their
dependencies. This entity is known as the injector, assembler, provider,
container or factory.
The second way to inject dependencies is by using Property or Method

injection:
class Vehicle {
var engine: Engine? = null
...
engine?.let {
}
return false
}
...
}
In this case, you create a Vehicle without an Engine. You can then set the
Engine type later.
Dependency injection favors testability because you can inject fake

Engine objects to test different situations. For example, you can inject a
broken Engine where its start() method returns false and asserts that the
Vehicle won’t move.
In Android, Dagger2 and Koin are libraries that help you inject objects.
In Android, Activity objects know about the Application object, so

another way to inject dependencies is to ask them from the Application
object, for example:
class MyActivity : AppCompatActivity() {

private lateinit var repository: Repository
override fun onCreate(savedInstanceState: Bundle?) {

...
repository = (application as MyApplication).getRepository()
...
}
}
Note: This is sometimes called a Service Locator because you ask a
locator object, the app in this example, for other objects, the
repository.
The Application object might be:
class MyApplication : Application() {

fun getRepository(): Repository {
val apiService = getApiService()
val inMemoryService = getInMemoryService()
return MyRepository(apiService, inMemoryService)
}
...
}
Where MyRepository implements an interface named Repository.
Structural
Structural design patterns ease the design to establish relationships
between objects.
Adapter (or Wrapper)

The Adapter (or Wrapper) design pattern describes how to let two
incompatible classes work together.
For example, in Android, when you have a list of contacts that you want
to show in a RecyclerView, the RecyclerView doesn’t know how to show
objects of the class Contact. That’s why you need to use a ContactsAdapter
class:
class ContactsAdapter(private val contacts: List<Contact>):
RecyclerView.Adapter<ContactViewHolder>() {
override fun onCreateViewHolder(viewGroup: ViewGroup,

i: Int): ContactViewHolder {
val inflater = LayoutInflater.from(viewGroup.context)
val view = inflater.inflate(
R.layout.row_contact, viewGroup, false
)
return ContactViewHolder(view)
}
override fun onBindViewHolder(viewHolder: ContactViewHolder,

i: Int) {
viewHolder.bind(contacts[i])
}
override fun getItemCount() = contacts.size
inner class ContactViewHolder(itemView: View):

RecyclerView.ViewHolder(itemView) {
fun bind(contact: Contact) {

...
}
}
}
Here, bind() sets the views (TextView, ImageView, and so on) using the
Contact model.
Facade
The Facade design pattern defines a high-level interface object which
hides the complexity of underlying objects. Client objects prefer using
the facade instead of the internal objects because the facade provides a
cleaner, easier-to-use interface.
For example, you might have a ProductsRepository class that provides

objects of the class Product, like so:
class ProductsRepository {
...
fun getProducts(): List<Product> {
if(isRemoteAvailable) {
return api.getProducts()
} else {
val localProducts = room.getProducts()
if(localProducts.isEmpty()) {
return sharedPrefsManager.getLastProduct()
} else {
return localProducts
}
}
}
}
In this example, getProducts() grabs the data from a remote server,

memory, filesystem, or maybe SharedPreferences or Room. It’s easier to use
ProductsRepository, which abstracts getting the products from the
corresponding source.
Composite
The intent of the Composite design pattern is to construct complex
objects composed of individual parts, and to treat the individual parts
and the composition uniformly.
In Android, View, ViewGroup and the rest of classes that inherit from View
— like TextView and ImageView — create a composite pattern because
ViewGroup inherits from View, and contains a list of child View objects.
Note: This is not the actual Android implementation; it’s simplified
for illustration purposes.
When you ask a ViewGroup to draw(), it iterates through all of its children
asking them to draw(). A child can be anything that inherits from View —
even other ViewGroup objects.
Behavioral
Behavioral design patterns explain how objects interact and how a task
can be divided into sub-tasks among different objects. While creational
patterns explain a specific moment of time (the creation of an instance),
and structural patterns describe a static structure, behavioral patterns
describe a dynamic flow.
Observer
The Observer design pattern gives you a way to communicate between
objects where one object informs others about changes or actions.
There’s an observable object which you can observe, and there’s one or
more observer objects that you use to subscribe to the observable.
Whenever there’s a change in the state of the observable object, it

notifies all of its observers.
A simple example in Android is to use the OnClickListener interface of a
button:
button.setOnClickListener(object: View.OnClickListener {
override fun onClick(v: View?) {
// Perform some operation
}
})
In this case, the observable is the button, and you subscribe to it with an
observer, which is the anonymous class that performs some operation
when you click the button.
When you use BroadcastReceiver objects — or any kind of reactive

programming like LiveData from Architecture Components or the
RxKotlin/Android libraries — you’re using this pattern. In each case, you
subscribe to notifications of some kind of change or event and react to
that. When you don’t need to observe anymore, you unsubscribe.
Command
The Command design pattern describes the encapsulation of an
operation without knowing the real content of the operation or the
receiver.
A concrete command object knows about the receiver and invokes a

method of the receiver. When an invoker calls the execute method of the
command object, the command is delegated to the receiver. The invoker
only knows about the command interface.
For example, if you have a drawing app where all user interactions are
implemented as commands, and you put them in a stack, you can easily
implement the undo by popping the commands and executing an undo()
operation on the command.
Architectural design patterns

There are some other design patterns that may be considered as a fourth
category. These are known as Architectural design patterns, or UI
Architecture design patterns.
Architectural design patterns are mostly used in rich clients. In other

words, they have some kind of UI, such as Android. Each pattern
describes ways to structure and organize your code. You can use them to
achieve a robust, stable, testable, modular and easy to extend codebase.
MVC
Model-View-Controller (MVC) states that each class you write should be
part of one of the following layers:
Model: The data classes that model your business belong to this
layer. They usually contain data and business logic. This layer also
contains the classes that fetch and create objects of those business
classes, networking, caching and handling databases.
View: This layer displays data from the Model. It doesn’t contain any
business logic.
Controller: The goal of this layer is to connect the objects of the

Model and View layers. It gets notified of user interaction and
updates the Model. It retrieves data from the Model. It may also
update the View when there’s a change in the Model.
Ideally, you have separate layers to allow testing them separately.
The View knows about the Model, and the Controller has a reference to
both the View and the Model.
Here’s how it works:
Whenever there’s user input, it’s received by the Controller. The

Controller updates the Model, and it also notifies the View that it should
update. Finally, the View requests the data from the Model.
In 2011, to apply MVC to Android apps, it was accepted that the

Activities, Fragments and custom views were both the View and also the
Controller layers because they show the UI and handle user input. The
problem with this approach is that you had to unify two layers into one,
which goes against the original purpose: To split things into layers,
making them independent and testable.
Also, combining View and Controller logic in the Activities, Fragments

and custom views makes them larger. This lead to MVC ironically being
referred to as Massive-View-Controller.
To make it better, this thought was changed. The Activities, Fragments
and custom views are just part of the View layer, and any user interaction
is delegated to a separate Controller class.
Here’s how that works:
1. The View (Activity/Fragment/Custom view) receives the user input.
2. It informs the Controller (a separate class) about it.
3. The Controller updates the Model and notifies the View that should
update.
4. Finally, the View requests data from the Model.
However, there’s still a problem: The Controller has a reference to the

View, which is the Activity, Fragment or custom view, which means it
won’t be unit testable because you need to create that Activity, Fragment
or custom view in the test.
The solution to this problem is to create an interface that the Activity,

Fragment or custom view can implement. The Controller will have a
reference to the interface instead of the actual Activity, Fragment or
custom view.
Using this MVC implementation in Android, the Model classes are unit
testable, and the Controller classes also are unit testable because they
don’t extend or use any Android UI classes. For the View classes, you can
create UI tests.
This pattern doesn’t explicitly state which layer should handle UI logic.
For example, if you have the following class, part of the Model layer:
data class Address(val street: String,

val number: String,
val zipCode: String)
Suppose in the View layer you need to display it with the following
format: "$number, $street, $zipCode". You could do the following in your
Activity:
addressView.text =
"${address.street}, ${address.number}, ${address.zipCode}"
However, the only way to test this formatting is to create a UI test.
But, if you want to create a unit test, you may instead add a property to
the Model, like this:
data class Address(val street: String,

val number: String,
val zipCode: String) {
val description = "$number, $street, $zipCode"
}
Then, in the Activity, you can do this:
addressView.text = address.description
Now, you could create a unit test for the Model. However, you’d be
making the Model dependent on the View layer.
The View simply knows too much: It knows about the Controller and the
Model. The Activity, Fragment or custom view knows what to display and
how to display it. Even more, if you have a direct reference to the Model,
you might be tempted to go directly to the Model to obtain data from an
API or database — without going through the correct flow, using the
Controller.
The solution is to avoid having a direct reference to the Model. Instead,

everything should go through the Controller, and the Controller should
handle the UI logic, knowing how to present the Model in the View. This
is exactly what the next pattern solves.
MVP
Model-View-Presenter (MVP) has these layers:
Model: This is the same as the Model layer from MVC.
View: Displays data presented by the Presenter but doesn’t have a

reference to the Model. It does, however, have a reference to the
Presenter to notify it about user actions.
Presenter: Similar to the Controller from the previous pattern, the

Presenter retrieves data from the Model and updates it accordingly.
It has UI presentation logic that decides what to display. It notifies
the View when a Model has changed. Therefore, it has a reference to
the View and the Model.
The View layer is composed by the Activities, Fragments or custom

views. The View entity notifies the Presenter about user interactions.
The Presenter decides if has to fetch something from the Model, updates
it, applies UI logic and finally tells the View what to display.
Usually, there’s an interface for the View and an interface for the
Presenter, and these are written in a single file or package as a sort of
contract.
For example, think about a Login flow where you might have the
following contract interfaces:
interface LoginPresenter {
fun login(username: String, password: String)
}
interface LoginView {
fun showLoginSuccess()
fun showLoginError()
fun showLoading()
}
And the corresponding implementations:
class LoginPresenterImpl(
private val repository: LoginRepository,
private val view: LoginView): LoginPresenter {
override fun login(username: String, password: String) {

view.showLoading()
repository.login(username, password, object : Callback {
override fun onSuccess() {
view.showLoginSuccess()
}
override fun onError() {
view.showLoginError()
}
})
}
}
class LoginActivity: AppCompatActivity(), LoginView {
private lateinit var presenter: LoginPresenter

...
loginButton.setOnClickListener {
presenter.login(usernameEditText.text.toString(),
passwordEditText.text.toString())
}
}
override fun showLoginSuccess() { ... }

override fun showLoginError() { ... }
override fun showLoading() { ... }
}
Once the View is loaded, you need to show that data has to be fetched
from an API. It’s common to have an onLoad() method in the Presenter
called by the View when it’s ready, e.g. in Activity’s or Fragment’s
onResume(), so that the Presenter starts fetching data from the Model. It’s
important to also have another method, onStopLoading(), in the Presenter
to cancel any asynchronous tasks whenever the View is not being shown,
e.g. in in Activity’s or Fragment’s onPause(). When using custom views,
the analogous methods are onAttachedToWindow() and
onDetachedFromWindow().
Differences between the MVC and MVP patterns are that in MVP, the
View doesn’t have a direct reference to the Model, meaning it is loosely-
coupled. Instead, the Presenter brings the transformed or simplified
Model, ready to be displayed to the View. MVP also proposes that the
Presenter should handle everything related to the presentation of the
View. In MVC, it's not clear where the UI logic should be, compared to
MVP where it’s common to put it in the Presenter. Because of this, the
Presenter could grow a lot, converting it into another anti-pattern called
God-class (you’ll learn more about this concept later). It’s a good practice
to create small classes (e.g. an AddressFormatter) with single
responsibilities to have better maintainability and unit testing. These
collaborator objects could be injected in the constructor, as explained
before.
Because the Presenter talks to a View interface, in production code the

Activities, Fragments and custom views implement that interface. But, in
your tests, instead of using those Android classes, you could create
custom classes that implement the View interfaces and assert that the
corresponding methods were called.
For example, using JUnit, if you want to test login() of LoginPresenterImpl

you can create this test:
@Test
fun login_shouldShowLoading() {
var didShowLoading = false
val testRepository = object: LoginRepository { ... }
val testView = object: LoginView {

override fun showLoginSuccess() {}
override fun showLoginError() {}
override fun showLoading() { didShowLoading = true }
}
val presenter = LoginPresenterImpl(testRepository, testView)
presenter.login("Foo", "1234")
Assert.assertTrue(didShowLoading)
}
Here, you create a testRepository and a testView, both implementing their

corresponding interfaces. You then instantiate the LoginPresenterImpl,
passing those test objects. Afterward, you call login(). If you
implemented everything correctly, it’ll call showLoading() of your testView,
and didShowLoading gets set to true and your test will pass.
Note: In the context of TDD, you should first create this test, and
afterward, the corresponding implementation.
MVVM
Model-View-ViewModel (MVVM) contains the following layers:
Model: Same as the Model layer from MVC/MVP.
View: Notifies the ViewModel about user actions. Subscribes to

streams of data exposed by the ViewModel.
ViewModel: Retrieves data from the Model and updates it

accordingly. Exposes streams of data ready to be displayed, but it
doesn’t know and doesn’t care about who is subscribed to the
streams.
Here, again, the Activities, Fragments and custom views conform the
View layer. The interaction between the View and the ViewModel is very
similar than the one that had between the View and the Presenter in
MVP. The View will notify the ViewModel about user actions, just like it
notified the Presenter in MVP, however this time, the ViewModel doesn’t
have a reference to the View, not even an interface. It just exposes
streams of data (Observables), it could be the Model or a transformed-
displayable Model.
As you can see, this is an event based approach, where the ViewModel
produces data and the View consumes it. The ViewModel doesn’t know
about the consumer, it just exposes streams of data. The View subscribes
and unsubscribes to that data as needed.
The ViewModel layer is composed of classes that don’t extend or use any
class related to the Android UI framework. Usually, the mechanism of
exposing data, observing and updating it, is done using reactive libraries.
The most common are:
RxKotlin/Android libraries: The ViewModel exposes one or more

Observable objects. The View subscribes to them. When the
ViewModel updates the Observable data (the actual Model), the
View will react and render the corresponding update. If the views are
Activities or Fragments it’s important to subscribe in onResume() and
unsubscribe in onPause(). If using a custom view the analogous
methods are onAttachedToWindow() and onDetachedFromWindow().
LiveData+ViewModel from Android Architecture Components: If

your ViewModel classes extend a ViewModel class from the Android
framework, bear in mind that this class has nothing to do with UI, so
it’s still unit testable. The ViewModel exposes LiveData objects and it
updates the values of them. The Activities have to start observing
(subscribe) these LiveData objects in the onCreate() method and
doesn’t need to stop observing (unsubscribe) because the base
LiveData class from Android Architecture Components is aware of
the Activity lifecycle.
If using Fragments, the suggested approach is to start observing in

the onActivityCreated() method. For custom views, unfortunately,
it’s not possible to use this ViewModel class from Architecture
Components, because it was thought to only work with Activities
and Fragments.
For the Login flow example, using ViewModel+LiveData from Android

Architecture Components, a possible approach would be:
class LoginViewModel(
private val repository: LoginRepository): ViewModel() {
private val loginStatus: MutableLiveData<LoginStatus>()
fun getLoginStatus(): LiveData = loginStatus
fun login(username: String, password: String) {

loginStatus.value = LoginStatus.Loading()
repository.login(username, password, object : Callback {
override fun onSuccess() {
loginStatus.value = LoginStatus.Success()
}
loginStatus.value = LoginStatus.Error()
}
})
}
}
sealed class LoginStatus {

class Error(): LoginStatus()
class Success(): LoginStatus()
class Loading(): LoginStatus()
}
And the Activity would do the following:

class LoginActivity: AppCompatActivity(), LoginView {
private lateinit var viewModel: LoginViewModel

...
viewModel = ...
viewModel.getLoginStatus().observe(this, Observer {
when(it) {
is LoginStatus.Loading -> ...
is LoginStatus.Success -> ...
is LoginStatus.Error -> ...
}
})
loginButton.setOnClickListener {
viewModel.login(usernameEditText.text.toString(),
passwordEditText.text.toString())
}
}
}
Suppose that login() from the repository returns a data User object, and
you need to show that in the UI, you could instead use another approach.
Having various LiveData objects exposed, for example, one
LiveData<Boolean> for the loading state, another one for the error state
and another one LiveData<User>, in the View you would need to observe
them and react accordingly.
Testing login() of LoginViewModel is similar to the one you saw previously

in MVP. In the next chapter, you’ll learn more about it.
S.O.L.I.D principles
TDD is closely related to good programming practices. Writing tests
before the actual feature makes you think on how the interface of a class
will be. Therefore, you’ll be exposing only those methods really needed.
On the contrary, without using TDD, sometimes you’ll find yourself
creating a class and exposing methods and properties that you don’t
need them to be public. Using TDD will also make you think on how the
classes will collaborate. After writing several features, while your app
grows, there will be times when you realize that things should be
refactored. Thanks to having tests you can refactor with confidence.
To complement TDD, the S.O.L.I.D principles are a list of principles to

build software following good practices, introduced by Robert C. Martin
(Uncle Bob), in his Design Principles and Design Patterns paper. Though
these items are independent from TDD, they both complement each
other: while writing a test for a class you’ll want to comply with these
principles. Also, before writing any test you’ll have these principles in
mind, so you’ll design and write tests and classes accordingly.
Single responsibility (SRP)

Each class should have a unique objective or should be useful for a
specific case. Any logic that is not part of the objective of the class
should be the responsibility of some other class. A class that has lots of
responsibilities is sometimes called a god class and should be avoided.
You’ll often find classes that clearly violate this principle, however, you’ll
also find that sometimes is not that clear. Whenever you’re refactoring or
evolving your code you may realize that the class you’re modifying starts
to have multiple responsibilities.
Thanks to TDD, you may realize a class is becoming a god class when you
spot some of the following signs:
Before adding new functionality for an existing class, you need to

add new tests to an existing test suite. If the new tests are not
related or don’t follow the essence of the existing ones, the
functionality violates SRP. For example, think of a Car class that has
an associated test class that check the methods startEngine(),
accelerate(), stopEngine(). Suppose you need a new feature that
drives the car to a place. You may be tempted to open the existing
test suite, and write a new test like testDriveToPlace() that would
check a new method driveToPlace() of the Car class. You’ll realize
that this new feature doesn’t follow the essence of the class, that it
actually should be the responsibility of a new Driver class.
When you have a class A that depends on a class B, and the tests of A
start to require you to stub many methods of B, B is turning into a
god class. You’ll see more about stubbing in a later chapter.
Open-closed
This principle was actually first defined by Bertrand Meyer in his book
Object-Oriented Software Construction.
The software entities of your app: classes, methods, etc. should be open
for extension but closed for modification. This means that you should
design them in such a way that adding new features or modifying
behavior shouldn’t require you to modify too much of your existing code
but instead add new code or create new classes.
This can be accomplished by:
Using class inheritance or interfaces and overriding methods.
Delegating to other classes (dependencies) by using composition

and allowing to easily exchange those classes.
For example, suppose you’re writing an app for a house architect who
tells you that he needs to calculate the total area of each room for a given
blueprint of a house. You may end up with the following solution:
data class Room(val width: Double, val height: Double)
class ArchitectUtils {
...
fun calculateArea(rooms: List<Room>) {
var total = 0
for (room in rooms) {
total += room.width * room.height
}
return total
}
}
The architect is happy with that, however, the next week comes and he
tells you that now he needs to add the area of the yard of the house. The
yard could be circular.
You may rename Room to RectangularSpace and add a CircularSpace class,

such as:
interface Space
data class RectangularSpace(val width: Double, val height: Double): Spa
data class CircularSpace(radius: Double): Space
...
fun calculateArea(spaces: List<Space>) {
var total = 0
for (space in spaces) {
if (space is SquareSpace) {
total += space.width * space.height
} elseif (space is CircularSpace) {
total += space.radius * space.radius * PI
}
}
return total
}
}
This code above is violating the principle, because it’s not closed for
modification, you’re always modifying existing code to support new
types.
So, to comply with the principle, you should do the following:

interface Space {
fun area(): Double
}
data class RectangularSpace(val width: Double, val height: Double): Spa

override fun area() = width * height
}
data class CircularSpace(radius: Double): Space {

override fun area() = radius * radius * PI
}
...
fun calculateArea(spaces: List<Space>) {
var total = 0
for (space in spaces) {
total += space.area()
}
return total
}
}
As you can see, if you need to support new types, you can just create a
new class that implements the Space interface with its area() method.
You won’t need to modify anything else! This is what "closed for
modification but open for extension" means.
Following this principle will give you a strong code base that almost
never changes but enables extension. This is even more noticeable if
you’re writing a library because you can’t change your interface with the
clients of your library. In that case, the clients won’t need to change
anything in their code and at the same time allow them to use new
features.
When using TDD, you’ll write a new test to check the new feature or
change an existing test to verify some behavior that now has to deal with
more use cases. While writing the test you may notice that it starts to
become too complex. That may be a sign you need to introduce a new
class that inherits from the old one or use composition to handle each
use case.
Liskov substitution
Also called design by contract, was initially introduced by Barbara Liskov
in a 1987 conference keynote titled Data abstraction and hierarchy.
Basically, it states that an app that uses an object of a base class should
be able to use objects of derived classes without knowing about that and
continue working. Therefore, your code should not be checking the
subtype. In the subclass you can override some of the parent methods as
long as you continue to comply with its semantics and maintain the
expected behavior. As you can see, if you respect the contract, the app
should continue to work.
A good example, by Uncle Bob (Martin), of a violation of this principle is

that in mathematics, a Square is a Rectangle. You may be tempted to
implement this having the Square class inherit from a Rectangle class.
Then, anywhere in your code where you expect a Rectangle, you could
pass a Square. The problem is that in a Rectangle you can change the
width or the height independently, but you cannot do that in a Square. If
you want to change the width of a Square, you should override the
setWidth method to also change the height to the same value. To comply
with the essence of a Square, the same would apply if you want to change
the height. Therefore, this implementation would be violating the
principle because you would be changing the expected behavior defined
in the base type, which in this case is a reshapable-rectangle.
In your tests created using TDD, everything that you verified for your
base class should also be verified for your new child class.
In the Square/Rectangle example mentioned, by creating your tests first,

you would realize that you cannot model a Square inheriting from a
Rectangle, because the area tests you write will pass for a Rectangle but
not for a Square or vice versa:
private const val WIDTH = 4
private const val HEIGHT = 3
private fun assertArea(rectangle: Rectangle) {

Assert.assertTrue(WIDTH * HEIGHT, rectangle.area())
}
@Test
fun testAreaRectangle() {
val rectangle = Rectangle()
rectangle.width = WIDTH
rectangle.height = HEIGHT
assertArea(rectangle) // This test will pass

}
@Test
fun testAreaSquare() {
val square = Square()
square.width = WIDTH
square.height = HEIGHT // This will also set square.width to HEIGHT
assertArea(square) // Therefore, this test will fail, because area is

}
Another example, in this case not violating the principle:
interface Repository {
fun findContactOrNull(id: String): Contact?
}
class InMemoryRepository: Repository {

private lateinit var cachedContacts: Map<String, Contact>
...
fun findContactOrNull(id: String): Contact? {
return cachedContacts[id]
}
}
class SqlRepository: Repository {

fun findContactOrNull(id: String): Contact? {
val contact = // Implementation to get it from a SQL DB
return contact
}
}
As you can see, the base interface declares a method that indicates that it
would return a Contact object by id or null if it doesn’t find it. Later, the
implementations, an in-memory DB and a Sql DB do what they have to
do to return the Contact. Neither of them change the semantic of the
interface. If instead, for example, an implementation removes a contact
and then returns it, it would be violating the principle because you
wouldn’t be maintaining the expected behavior.
Interface segregation
This principle encourages you to create fine grained interfaces that are
client specific. Suppose you have a class with a few methods, one part of
your app may only need to access a subset of your methods and other
part may need to access another subset. This principle encourages you to
create two interfaces. Clients should have access to only what they need
and nothing more.
For example, suppose you have an app where the user has to register and
login to use it. You may have the following interface:
interface Membership {
fun login(username: String, password: String): User
fun logout(user: User)
fun register(username: String, password: String)
fun forgotPassword(username: String)
}
You may have a screen that, after login, only deals with showing the user
data and enables them to logout.
You may have another screen to register and finally another one to let
the user recover their password if it was forgotten.
So instead of all those screens using the fat Membership interface, it’s
better to segregate it into the following interfaces:
interface Login {
fun login(username: String, password: String): User
fun logout(user: User)
}
The screen that handles login and shows user data, enables to logout
should use this interface.
interface Register {
fun register(username: String, password: String)
}
The screen that handles registration should use this interface.
interface Forgot {
fun forgotPassword(username: String)
}
The screen that handles recovering the password should use this last
interface.
You may then have a class that implements all of these interfaces if it
needs to. But if it doesn’t, each screen should use the corresponding
interface. Another example where it would be good to segregate
interfaces is the following: suppose you’re writing an app that will allow
you to send a file to a printer, scan a document to use it in your app and
send a file to an email address. You may implement it like this:
interface Printer {
fun print(file: File)
fun scan(): Bitmap
fun sendTo(file: File, email: String)
}
To represent a printer that can do everything, you may have:
class FullPrinter: Printer {

override fun print(file: File) {
// Implementable logic
}
override fun scan(): Bitmap {
}
override fun sendTo(file: File, email: String) {
}
}
However, a mobile phone could only implement scan() and sendTo() so

you would have to write:
class MobileDevice: Printer {
throw UnsupportedOperationException()
}
}
}
}
You’re forcing the mobile phone to implement methods that it doesn’t

support. Now you know, you should segregate into the following
interfaces:
interface Printer {
fun print(file: File)
}
interface Scanner {
fun scan(): Bitmap
fun sendTo(file: File, email: String)
}
And implement them accordingly:
class FullPrinter: Printer, Scanner {

}
}
}
}
class Mobile: Scanner {

}
}
}
When writing tests using TDD, if your class under test has a dependency,
it’s easier if you have to stub just the methods of a fine grained interface.
TDD enforces writing more client-focused interfaces, because it makes
you think from the client perspective — you avoid exposing those
methods that won’t be used by the client.
Dependency inversion
This principle states that a concrete class A should not depend on a
concrete class B, but an abstraction of B instead. This abstraction could
be an interface or an abstract class.
For example, think of a Presenter (or ViewModel) that needs to request

data from an API. The Presenter (or ViewModel) should require an object
that is able to request data from the API. Something like the following:
class ApiDataFetcher {
fun fetch(): Data {
// Implementation that retrieves data from an API
}
}
class MyPresenter(private val apiDataFetcher: ApiDataFetcher) {

fun getData(): Data {
...
return apiDataFetcher.fetch()
}
...
}
Here, the presenter is depending on a concrete class, ApiDataFetcher. It’s

not following the dependency inversion principle. What if later you need
to fetch the data from SharedPreferences or a database using Room?
You should rewrite this to comply with the dependency inversion

principle:
interface DataFetcher {
fun getData(): Data
}
class ApiDataFetcher: DataFetcher {

fun fetch(): Data {
// Implementation that retrieves data from an API
}
}
class MyPresenter(private val dataFetcher: DataFetcher)
Now, when you create the Presenter, you can still pass the ApiDataFetcher
as a parameter. However, the presenter doesn’t know about it, it just
depends on an abstraction, the DataFetcher interface. So it will be easy to
change it to a SharedPreferencesDataFetcher or a RoomDataFetcher class as
long as those classes implement the DataFetcher interface.
When writing tests using TDD, instead of passing real collaborators

(dependencies) to a class under test, it’s easier to pass fake objects that
conform to the same interface. These fake objects could be forced to
always behave the same to reproduce a specific scenario to test. You’ll
create this kind of tests in Chapter 7, "Introduction to Mockito."
Key points
Use software architecture to communicate development standards
between team members.
It’s not uncommon to reuse software architecture on different

projects.
When starting a new project, one of the first decisions is to decide

on its software architecture.
Proper software architecture helps with testing.
Support your software architecture using design patterns and the

SOLID principles.
Design patterns are classified into three categories: creational,
structural and behavioral.
The dependency injection pattern is the key to having a testable

architecture.
There are other user interface design patterns such as MVC, MVP
and MVVM.

In the next chapter, you’ll continue writing tests using Mockito and the
MVVM architecture that is suggested by Google, which uses ViewModel
and LiveData.
Review the following references to learn more about software

architecture:
Google suggested architecture:

https://developer.android.com/jetpack/docs/guide
Design Principles and Design Patterns, by Robert C. Martin, where

he introduced the SOLID principles:
http://staff.cs.utu.fi/staff/jouni.smed/doos_06/material/DesignPrinci
plesAndPatterns.pdf
The Principles of Object Oriented Design, by Robert C. Martin,

where he explains the SOLID and other principles:
http://butunclebob.com/ArticleS.UncleBob.PrinciplesOfOod
To learn about more design patterns, check the book Design

Patterns: Elements of Reusable Object-Oriented Software, by the
Gang of Four: https://www.oreilly.com/library/view/design-
patterns-elements/0201633612/
GUI Architectures, by Martin Fowler:

https://www.martinfowler.com/eaaDev/uiArchs.html
MVC/MVP/MVVM presentation by Florina Muntenescu:

https://youtu.be/QrbhPcbZv0I
Chapter 7: Introduction to Mockito
You'll often find yourself in situations in which you want to write a test
for a method of a class that requires collaboration from another class.
Unit Tests normally focus on a single class, therefore you need a way to
avoid using their actual collaborators. Otherwise, you'd be doing
integration testing, which you'll see in Chapter 8, "Integration."
In this chapter, you'll:
Learn what mocking and stubbing are and when to use these
techniques.
Write more unit tests using the test-driven development (TDD)

pattern to continue testing state, and a way to also verify behavior.
Why Mockito?
If you remember from a previous chapter, whenever you create a test, you
must:
First, configure what you're going to test.
Second, execute the method that you want to test.
Finally, verify the result by checking the state of the object under
test. This is called state verification or black-box testing. This is
what you've done using JUnit.
However, to perform state verification, sometimes the object under test

has to collaborate with another one. Because you want to focus on the
first object, in the configuration phase, you want to provide a test double
collaborator to your object under test. This fake collaborator is just for
testing purposes and you configure it to behave as you want. For
example, you could make a mock so that calling a method on it always
returns the same hardcoded String. This is called stubbing a method.
You'll use Mockito for this.
There's another type of verification called behavior verification or white-
box testing. Here you want to ensure that your object under test will call
specific collaborator methods. For example, you may have a repository
object that retrieves the data from the network, and before returning the
results, it calls a collaborator object to save them into a database. Again,
you can use Mockito to keep an eye on a collaborator and verify if specific
methods were called on it.
Note: Using white-box testing allows you to be more precise in your

tests, but often results in having make more changes to your tests
when you change your production code.
Setting up Mockito
Open the application's build.gradle file and add the following
dependency:
dependencies {
...
testImplementation 'com.nhaarman.mockitokotlin2:mockito-kotlin:2.1.
}
Mockito-Kotlin is a wrapper library around Mockito. It provides top-level

functions to allow for a more Kotlin-like approach and also solves a few
issues with using the Mockito Java library in Kotlin.
Creating unit tests with Mockito

Later, in the UI, you'll show the user a question with two options. You'll
want the user to click on one, and somehow your Game class will handle
that answer, delegate to the Question class, increment the score if the
answer was correct and, finally, return the next question.
Mocking and verifying

Start by adding the following test to the GameUnitTests.kt file:
@Test
fun whenAnswering_shouldDelegateToQuestion() {
// 1
val question = mock<Question>()
val game = Game(listOf(question))
// 2
game.answer(question, "OPTION")
// 3
verify(question, times(1)).answer(eq("OPTION"))
}
Note: When importing mock, verify, times and eq, you should choose
the options starting with com.nhaarman.mockitokotlin2.*.
In this test:
1. You set up the test. The answer() method of Game will call answer on
the Question, so you create a mock, which you can later verify
against.
2. Call the answer() method of Game, passing the Question mock as a

parameter.
3. Verify the method answer() was called on the Question mock. You
used the times(1) verification mode to check that the answer()
method was called exactly one time. You also used the eq argument
matcher to check that the answer() method was called with a String
equal to OPTION.
You can omit times(1) as it's the default. So modify the code to the
following:
verify(question).answer(eq("OPTION"))
Note: Other verification modes, like times(), are: never(), atLeast(),

atMost(). Other argument matchers, like eq(), are: same(), any().
Try to compile and run the test, you'll see an error because the Game class
doesn't have the answer() method yet.
So, open the Game class and create the answer() method:
fun answer(question: Question, option: String) {

// TODO
}
Run the test. You'll see that it doesn't pass:
This is because Kotlin classes and methods are final by default. Mockito
won't work with final classes/methods out of the box. To fix this you have
the following options:
Use a mock-maker-inline extension to allow Mockito mock final

classes/methods.
Add the open keyword to classes and methods that you'll mock.
Create an interface and have the class implement the interface.

Then, just mock the interface (interfaces are open by default).
Using mock-maker-inline extension

Go to the project window and change to Project. Create a resources
directory under app ‣ src ‣ test. Inside resources, create a directory called
mockito-extensions and a file called org.mockito.plugins.MockMaker
with the following content:
Now, you can go back and run the last test you created and see that it
still doesn't pass, but this time with another error:
Here it states that it was expecting an invocation to the answer() method

of the Question class.
So now, fix the answer() method with the correct implementation:

question.answer(option)
}
Now, run the test and see that it passes:
Stubbing methods
The Game should increment the current score when answered correctly, so
add the following test:
@Test
// 1
whenever(question.answer(anyString())).thenReturn(true)
// 2
// 3
}
In the above, you:
1. Mocked the Question class again. Using

whenever/method/thenReturn you're stubbing the question.answer()
method to always return true. Notice here you used the anyString()
argument matcher as you don't care which specific String you need
to stub the call.
Note: You could choose to use a specific String matcher here, which
would make the test stronger.
1. Call the answer() method of Game.
2. Check that the game score was incremented.
Run the test, and you will see that it fails. Add the following code to the
answer() method of the Game class:

question.answer(option)
incrementScore()
}
Now, run the test again and you will see that it passes.
You are also going to want to check that it doesn't increment the score
when answering incorrectly. To do that, add the following test:
g y g
@Test
fun whenAnsweringIncorrectly_shouldNotIncrementCurrentScore() {
whenever(question.answer(anyString())).thenReturn(false)
}
Here, instead, you are stubbing the answer() method to always return
false.
Run the test and you will see that it fails. It's a good thing you checked
for that boundary condition! To fix this, replace your answer() method
with the following:

val result = question.answer(option)
if (result) {
incrementScore()
}
}
This adds a check to only increment the score if the answer is correct.
Now, run both tests and you will see them pass.
Refactoring
Open the Game class. Notice that this class knows about the score and a
list of questions. When requesting to answer a question, the Game class
delegates this to the Question class and increments the score if the
answer was correct. Game could also be refactored to delegate the logic of
incrementing the current score and highest score to a new class, Score.
Create a Score class in the same package as the Game class with the
following content:
class Score(highestScore: Int = 0) {
var current = 0
private set
var highest = highestScore

private set
fun increment() {
current++
if (current > highest) {
highest = current
}
}
}
Now, update the Game class to use this new class:
class Game(private val questions: List<Question>,

highest: Int = 0) {
private val score = Score(highest)
val currentScore: Int

get() = score.current
val highestScore: Int

get() = score.highest
private var questionIndex = -1
score.increment()
}
...
Run the tests again and verify that everything is still working.
With that change, however, take another look at the following unit tests
from GameUnitTests.kt:
@Test
val game = Game(emptyList(), 0)
Assert.assertEquals(
"Current score should have been 1",
1,
game.currentScore)
}
@Test
}
@Test
fun whenIncrementingScore_belowHighScore_shouldNotIncrementHighScore()
}
When calling game.incrementScore(), game.highestScore, or

game.currentScore because you refactored to internally delegate to a
dependent class, Score, you are now performing integration tests. You'll
see and learn more about them in Chapter 8, "Integration."
In order to keep your tests at the unit level, remove these tests from
GameUnitTests.kt and create a new file called ScoreUnitTests.kt with the
following content:
class ScoreUnitTests {
@Test
val score = Score()
score.increment()
Assert.assertEquals(
"Current score should have been 1",
1,
score.current)
}
@Test
fun whenIncrementingScore_aboveHighScore_shouldAlsoIncrementHighScore
val score = Score()
score.increment()
Assert.assertEquals(1, score.highest)
}
@Test
fun whenIncrementingScore_belowHighScore_shouldNotIncrementHighScore(
score.increment()
Assert.assertEquals(10, score.highest)
}
}
This gets your tests back to the unit level because you test the methods
of the Score object without dependent classes.
Run them to check that they pass.
With that refactor, the only method that is still using the
incrementScore() method in your Game class is the answer() method. Let's
simplify this. Remove the incrementScore() method and change the
answer() method as follows:
if (result) {
score.increment()
}
}
Now, because you removed the public scoreIncrement() method, the only
way to increment the score in your Game class is by answering questions.
Next, open GameUnitTests.kt and have a look at the following tests:
@Test
}
@Test
}
You may have guessed that now these are integration tests. This is
because you are asserting game.currentScore that internally depends on a
Score class from your refactor. To convert them to unit tests, you will
need to change them to verify that the increment() method on the Score
class was or wasn't called. To do that, replace them with the following:
@Test
val score = mock<Score>()
val game = Game(listOf(question), score)
verify(score).increment()
}
@Test
val game = Game(listOf(question), score)
verify(score, never()).increment()
}
You'll see that it doesn't compile now, because you're passing a list of
questions and a score to the Game class constructor, but it doesn't support
that yet. To fix that, open your Game class and change the constructor to
the following:

val score: Score = Score(0)) {
Once that is done, remove the old score, currentScore and highestScore
properties as they are not needed anymore. Your modified Game class
should be the following:
val score: Score = Score(0)) {
private var questionIndex = -1
fun nextQuestion(): Question? {

if (questionIndex + 1 < questions.size) {
questionIndex++
return questions[questionIndex]
}
return null
}

if (result) {
score.increment()
}
}
}
Run the tests and everything should now pass. Congratulations, you have
successfully refactored your tests and kept them at the unit level.
Verifying in order
To save and retrieve the high score, you'll need to add functionality to a
repository. From the Project view, create a new package common ‣
repository under app ‣ src ‣ test ‣ java ‣ com ‣ raywenderlich ‣ android ‣
cocktails. Create a new file called RepositoryUnitTests.kt and add the
following code:
class RepositoryUnitTests {
@Test
fun saveScore_shouldSaveToSharedPreferences() {
val api: CocktailsApi = mock()
// 1
val sharedPreferencesEditor: SharedPreferences.Editor =
mock()
val sharedPreferences: SharedPreferences = mock()
whenever(sharedPreferences.edit())
.thenReturn(sharedPreferencesEditor)
val repository = CocktailsRepositoryImpl(api,
sharedPreferences)
// 2
val score = 100
repository.saveHighScore(score)
// 3
inOrder(sharedPreferencesEditor) {
// 4
verify(sharedPreferencesEditor).putInt(any(), eq(score))
verify(sharedPreferencesEditor).apply()
}
}
}
Going over each step in turn:
1. You're going to save the score into the CocktailsRepository using

SharedPreferences, so you need to mock this dependency and instruct
to return an Editor mock whenever an editor is requested.
2. Execute the saveHighScore() method.
3. Use inOrder to check that the subsequent verifications are executed

in the exact order.
4. Verify that the score is saved correctly.
In order for this code to compile, add a saveHighScore() method to your

CocktailsRepository interface.
interface CocktailsRepository {
...
fun saveHighScore(score: Int)
}
Then modify your CocktailsRepositoryImpl constructor to take in

SharedPreferences as a parameter and override the saveHighScore()
method:
class CocktailsRepositoryImpl(
private val api: CocktailsApi,
private val sharedPreferences: SharedPreferences)
: CocktailsRepository {
override fun saveHighScore(score: Int) {

// TODO
}
Run the test and see that it fails. To fix it, add the following code to the
CocktailsRepositoryImpl class:
private const val HIGH_SCORE_KEY = "HIGH_SCORE_KEY"
...

val editor = sharedPreferences.edit()
editor.putInt(HIGH_SCORE_KEY, score)
editor.apply()
}
This is adding logic to your saveHighScore method to save it in

sharedPreferences. Run the test again it will pass.
You are also going to want to have a way to read the high score from the
repository. To get started, add the following test:
@Test
fun getScore_shouldGetFromSharedPreferences() {
val api: CocktailsApi = mock()
val sharedPreferences: SharedPreferences = mock()
val repository = CocktailsRepositoryImpl(api,

sharedPreferences)
repository.getHighScore()
verify(sharedPreferences).getInt(any(), any())
}
Next, add the getHighScore() method to CocktailsRepository and

CocktailsRepositoryImpl:
interface CocktailsRepository {
...
fun getHighScore(): Int
}
...
override fun getHighScore(): Int = 0
Run the test, see it fail, and then add the following code to the
CocktailsRepositoryImpl class to see it pass:
override fun getHighScore()

= sharedPreferences.getInt(HIGH_SCORE_KEY, 0)
If you look at these two tests, you may notice that you have some code
that is repeated in both of them. Let's DRY this up by refactoring your
RepositoryUnitTests so that it looks like the following:
private lateinit var repository: CocktailsRepository
private lateinit var api: CocktailsApi
private lateinit var sharedPreferences: SharedPreferences
private lateinit var sharedPreferencesEditor: SharedPreferences.Edito
@Before
fun setup() {
api = mock()
sharedPreferences = mock()
sharedPreferencesEditor = mock()
repository = CocktailsRepositoryImpl(api, sharedPreferences)

}
@Test
fun saveScore_shouldSaveToSharedPreferences() {
val score = 100
repository.saveHighScore(score)
inOrder(sharedPreferencesEditor) {
verify(sharedPreferencesEditor).putInt(any(), eq(score))
verify(sharedPreferencesEditor).apply()
}
}
@Test
fun getScore_shouldGetFromSharedPreferences() {
repository.getHighScore()
verify(sharedPreferences).getInt(any(), any())
}
}
Run the tests again to check everything is still working.
Spying
Suppose you want to only save the high score if it is higher than the
previously saved high score. To do that, you want to start by adding the
following test to your RepositoryUnitTests class:
@Test
fun saveScore_shouldNotSaveToSharedPreferencesIfLower() {
val previouslySavedHighScore = 100
val newHighScore = 10
val spyRepository = spy(repository)
doReturn(previouslySavedHighScore)
.whenever(spyRepository)
.getHighScore()
spyRepository.saveHighScore(newHighScore)
verify(sharedPreferencesEditor, never())
.putInt(any(), eq(newHighScore))
}
In this test you are stubbing the getHighScore() method but you also need
to call the real saveHighScore() method on the same object, which is a real
object, CocktailsRepositoryImpl. To do that you need a spy instead of a
mock. Using a spy will let you call the methods of a real object, while also
tracking every interaction, just as you would do with a mock. When
setting up spies, you need to use doReturn/whenever/method to stub a
method. Try running the test and you will see that it fails.
To make the test pass, modify the saveHighScore() method of the

CocktailsRepositoryImpl so that it is as follows:

val highScore = getHighScore()
if (score > highScore) {
val editor = sharedPreferences.edit()
editor.putInt(HIGH_SCORE_KEY, score)
editor.apply()
}
}
Run the test again and it will pass.
In order to make games for a user, you'll need a factory to build a Game
with questions, which will map the cocktails returned by the API. Create
a CocktailsGameFactoryUnitTests.kt file under app ‣ src ‣ test ‣ java ‣
com ‣ raywenderlich ‣ android ‣ cocktails ‣ game ‣ factory. Add the
following code:
class CocktailsGameFactoryUnitTests {

private lateinit var factory: CocktailsGameFactory
@Before
fun setup() {
repository = mock()
factory = CocktailsGameFactoryImpl(repository)
}
@Test
fun buildGame_shouldGetCocktailsFromRepo() {
factory.buildGame(mock())
verify(repository).getAlcoholic(any())
}
}
With this test, you are checking that buildGame is calling getAlcoholic
from the repository.
Create the following interface and class to make it compile, under app ‣
src ‣ main ‣ java ‣ com ‣ raywenderlich ‣ android ‣ cocktails ‣ game ‣
factory:
interface CocktailsGameFactory {
fun buildGame(callback: Callback)
interface Callback {
fun onSuccess(game: Game)
fun onError()
}
}
class CocktailsGameFactoryImpl(
private val repository: CocktailsRepository)
: CocktailsGameFactory {
override fun buildGame(callback: CocktailsGameFactory.Callback) {

// TODO
}
}
Run the test and see that it fails. To make the test pass, add the following
code to the buildGame() method:

repository.getAlcoholic(
object : RepositoryCallback<List<Cocktail>, String> {
override fun onSuccess(cocktailList: List<Cocktail>) {
// TODO
}
override fun onError(e: String) {

// TODO
}
})
}
This is adding a call to the getAlcoholic method with stubbed callbacks

for onSuccess and onError. Run the test again and it will pass.
Stubbing callbacks
Create a new test that verifies that the callback is called when the
repository returns successfully with a list of cocktails:
private val cocktails = listOf(
Cocktail("1", "Drink1", "image1"),
Cocktail("4", "Drink4", "image4")
)
@Test
fun buildGame_shouldCallOnSuccess() {
val callback = mock<CocktailsGameFactory.Callback>()
setUpRepositoryWithCocktails(repository)
factory.buildGame(callback)
verify(callback).onSuccess(any())
}
private fun setUpRepositoryWithCocktails(

repository: CocktailsRepository) {
doAnswer {
// 1
val callback: RepositoryCallback<List<Cocktail>, String>
= it.getArgument(0)
callback.onSuccess(cocktails)
}.whenever(repository).getAlcoholic(any())
}
In setUpRepositoryWithCocktails, you are using doAnswer to stub the

repository.getAlcoholic() method to always return success with a list of
cocktails. The doAnswer closure returns an InvocationOnMock type, with
which you can spy on its arguments. You then get the first argument of
the method (which is the callback), and call onSuccess() on it.
Run the test and it will fail. Now, modify the code to the onSuccess
callback of the buildGame() so that buildGame() looks like the following:
callback.onSuccess(Game(emptyList()))
}

// TODO
}
})
}
Run your test again and it will pass. Now, let's do the same with the
onError case to ensure you test the error path as well as success. First, add
the following test:
@Test
fun buildGame_shouldCallOnError() {
val callback = mock<CocktailsGameFactory.Callback>()
setUpRepositoryWithError(repository)
factory.buildGame(callback)
verify(callback).onError()
}
private fun setUpRepositoryWithError(

repository: CocktailsRepository) {
doAnswer {
val callback: RepositoryCallback<List<Cocktail>, String>
= it.getArgument(0)
callback.onError("Error")
}.whenever(repository).getAlcoholic(any())
}
Here setUpRepositoryWithError() is stubbing the getAlcoholic() method to

always answer with an error. Run the test and it will fail.
Now, add the following implementation to the onError callback of your

buildGame function so that buildGame looks like the following:
override fun buildGame(
callback: CocktailsGameFactory.Callback
) {
callback.onSuccess(Game(emptyList()))
}

callback.onError()
}
})
}
Run the test and it will pass.
The following tests are similar to what you've been writing, they ensure
that CocktailsGameFactoryImpl builds a Game using the high score and maps
the list of Cocktail objects to Question objects. They are here to give you
more practice, but if you are really anxious to move on you can skip to
the next section "Testing ViewModel and LiveData".
Create the following tests that verify the factory creates a Game using the
repository.getHighScore() method:
@Test
fun buildGame_shouldGetHighScoreFromRepo() {
factory.buildGame(mock())
verify(repository).getHighScore()
}
@Test
fun buildGame_shouldBuildGameWithHighScore() {
val highScore = 100
whenever(repository.getHighScore()).thenReturn(highScore)
factory.buildGame(object : CocktailsGameFactory.Callback {
override fun onSuccess(game: Game)
= Assert.assertEquals(highScore, game.score.highest)
override fun onError() = Assert.fail()

})
}
As you should always do, run the tests once to make sure that they fail.
To make them pass, modify your buildGame() method so that it is as
follows:

val score = Score(repository.getHighScore())
val game = Game(emptyList(), score)
callback.onSuccess(game)
}

callback.onError()
}
})
}
Run the tests and they will pass.
Now, create the following test that verifies the factory creates a Game
mapping a list of cocktails to a list of questions:
@Test
fun buildGame_shouldBuildGameWithQuestions() {
override fun onSuccess(game: Game) {
cocktails.forEach {
assertQuestion(game.nextQuestion(),
it.strDrink,
it.strDrinkThumb)
}
}
override fun onError() = Assert.fail()

})
}
private fun assertQuestion(question: Question?,

correctOption: String,
imageUrl: String?) {
Assert.assertNotNull(question)
Assert.assertEquals(imageUrl, question?.imageUrl)
Assert.assertEquals(correctOption, question?.correctOption)
Assert.assertNotEquals(correctOption,
question?.incorrectOption)
}
Here, you are asserting that the image of the question that will be shown
in the UI corresponds to the cocktail image, the correct option
corresponds to the name of the drink, and also that the incorrect option
is not the name of the drink.
If you run this, the test will not compile, so add the imageUrl property to
the Question class:

val incorrectOption: String,
val imageUrl: String? = null) {
...
Now run the test, which compiles but now fails. To make it pass, replace
your buildGame() method with the following:
val questions = buildQuestions(cocktailList)
val score = Score(repository.getHighScore())
val game = Game(questions, score)
}

callback.onError()
}
})
}
private fun buildQuestions(cocktailList: List<Cocktail>)

= cocktailList.map { cocktail ->
val otherCocktail
= cocktailList.shuffled().first { it != cocktail }
Question(cocktail.strDrink,
otherCocktail.strDrink,
cocktail.strDrinkThumb)
}
This adds in a buildQuestions method that creates a series of questions

for the list of cocktails. This is called in your onSuccess callback in
buildGame with the result passed to Game. Run the test again and it will
pass.
Testing ViewModel and LiveData

To update the UI with questions, the score, and also to enable the user to
interact with the question options, you're going to use ViewModel and
LiveData from Android Architecture Components. To get started, add the
following dependencies in your build.gradle within the app module:
dependencies {
...
implementation 'androidx.lifecycle:lifecycle-extensions:2.0.0'
testImplementation 'androidx.arch.core:core-testing:2.0.1'
}
Next, create a package called viewmodel under app ‣ src ‣ test ‣ java ‣
com ‣ raywenderlich ‣ android ‣ cocktails ‣ game. Now, create a
CocktailsGameViewModelUnitTests.kt file under this viewmodel
directory you just created with the following code:
class CocktailsGameViewModelUnitTests {
@get:Rule
val taskExecutorRule = InstantTaskExecutorRule()
}
You may have noticed @get:Rule. This is a test rule. A test rule is a tool to
change the way tests run, sometimes adding additional checks or
running code before and after your tests. Android Architecture
Components uses a background executor that is asynchronous to do its
magic. InstantTaskExecutorRule is a rule that swaps out that executor and
replaces it with synchronous one. This will make sure that, when you're
using LiveData with the ViewModel, it's all run synchronously in the tests.
Now that you have your test scaffolding, add the following to your test
file:
private lateinit var factory: CocktailsGameFactory
private lateinit var viewModel: CocktailsGameViewModel
private lateinit var game: Game
private lateinit var loadingObserver: Observer<Boolean>
private lateinit var errorObserver: Observer<Boolean>
private lateinit var scoreObserver: Observer<Score>
private lateinit var questionObserver: Observer<Question>
@Before
fun setup() {
// 1
repository = mock()
factory = mock()
viewModel = CocktailsGameViewModel(repository, factory)
// 2
game = mock()
// 3
loadingObserver = mock()
errorObserver = mock()
scoreObserver = mock()
questionObserver = mock()
viewModel.getLoading().observeForever(loadingObserver)
viewModel.getScore().observeForever(scoreObserver)
viewModel.getQuestion().observeForever(questionObserver)
viewModel.getError().observeForever(errorObserver)
}
In the above:
1. Your ViewModel will require a CocktailsRepository to save the

highscore and a CocktailsGameFactory to build a game. These are
dependencies, so you need to mock them.
2. You'll use a Game mock to stub some of its methods and verify you call
methods on it.
3. You need a few mocked observers because the Activity will observe
LiveData objects exposed by the ViewModel. In the UI, you'll show a
loading view when retrieving the cocktails from the API and an error
view if there's an error retrieving the cocktails, score updates and
questions. Because there's no lifecycle here, you can use the
observeForever() method.
Note: Ensure to import androidx.lifecycle.Observer.
To make the test compile, create a class under app ‣ src ‣ main ‣ java ‣
com ‣ raywenderlich ‣ android ‣ cocktails ‣ game ‣ viewmodel called
CocktailsGameViewModel with the following content:
class CocktailsGameViewModel(
private val repository: CocktailsRepository,
private val factory: CocktailsGameFactory) : ViewModel() {
private val loadingLiveData = MutableLiveData<Boolean>()

private val errorLiveData = MutableLiveData<Boolean>()
private val questionLiveData = MutableLiveData<Question>()
private val scoreLiveData = MutableLiveData<Score>()
fun getLoading(): LiveData<Boolean> = loadingLiveData

fun getError(): LiveData<Boolean> = errorLiveData
fun getQuestion(): LiveData<Question> = questionLiveData
fun getScore(): LiveData<Score> = scoreLiveData
}
Next, add the following methods to

CocktailsGameViewModelUnitTests.kt:
private fun setUpFactoryWithSuccessGame(game: Game) {

doAnswer {
val callback: CocktailsGameFactory.Callback =
it.getArgument(0)
}.whenever(factory).buildGame(any())
}
private fun setUpFactoryWithError() {

doAnswer {
val callback: CocktailsGameFactory.Callback =
it.getArgument(0)
callback.onError()
}.whenever(factory).buildGame(any())
}
You'll use these methods to stub the buildGame() method from the
CocktailsGameFactory class.
Now, add the following test:

@Test
fun init_shouldBuildGame() {
viewModel.initGame()
verify(factory).buildGame(any())
}
Here, you're verifying that calling initGame on the ViewModel will call
buildGame from the factory.
Finally, add the corresponding implementation to your

CocktailsGameViewModel to make the test compile:
fun initGame() {
// TODO
}
Run the test and it will compile but won't pass.
To make it pass, replace initGame() in CocktailsGameViewModel with the

following:
fun initGame() {
// TODO
}

// TODO
}
})
}
You are going to want to show a loading view and remove the error view
while building the game. To get started with that, add the following
tests:
@Test
fun init_shouldShowLoading() {
verify(loadingObserver).onChanged(eq(true))
}
@Test
fun init_shouldHideError() {
verify(errorObserver).onChanged(eq(false))
}
In both tests, you verify that initGame publishes the correct data. When
the program posts a value to a LiveData, the object calls onChanged() with
the value. This is the function you are checking for.
Note: There are multiple ways you can verify the result. For
example, instead of using
verify(loadingObserver).onChanged(eq(true)), you could replace it
with Assert.assertTrue(viewModel.getLoading().value!!) instead to
achieve the same result. This alternative compares the last value of
the LiveData to the expected one instead of making sure a method
was called with that data.
As always, you run your new tests to ensure that they fail. To fix them,
modify your initGame() method by adding the following two lines as
follows:
fun initGame() {
loadingLiveData.value = true
errorLiveData.value = false
factory.buildGame(...)
}
Run the tests again and they will pass.
You are also going to want to show the error view and stop showing the
loading view when there's a problem building the game. To get started
add the following tests:
@Test
fun init_shouldShowError_whenFactoryReturnsError() {
setUpFactoryWithError()
verify(errorObserver).onChanged(eq(true))
}
@Test
fun init_shouldHideLoading_whenFactoryReturnsError() {
setUpFactoryWithError()
verify(loadingObserver).onChanged(eq(false))
}
Run the tests to ensure that they fail. To fix them, modify your onError()
callback in initGame() as follows:

loadingLiveData.value = false
errorLiveData.value = true
}
Run the tests and check that they pass.
Another scenario that you will want to cover is to hide the error and
loading views when the factory builds a game successfully. To get started,
add these tests:
@Test
fun init_shouldHideError_whenFactoryReturnsSuccess() {
setUpFactoryWithSuccessGame(game)
verify(errorObserver, times(2)).onChanged(eq(false))
}
@Test
fun init_shouldHideLoading_whenFactoryReturnsSuccess() {
verify(loadingObserver).onChanged(eq(false))
}
Here, you check the error is set to false two times. The first false value is
before calling the repository to build the game, and the second one is set
when the game couldn't be built because of an error.
Run the tests to ensure that they fail. To fix these tests, modify your
onSuccess() callback in initGame as follows:

}
Run the tests again and they will pass.
Another requirement is to show the score when the game is built. Start
by adding the following test:
@Test
fun init_shouldShowScore_whenFactoryReturnsSuccess() {
whenever(game.score).thenReturn(score)
verify(scoreObserver).onChanged(eq(score))
}
Run it to make sure it doesn't pass. Now, modify your onSuccess()
callback in initGame() as follows:

scoreLiveData.value = game.score
}
Run the test and check that it passes.
You are going to want to show the first question when the game is built.
Start by adding this test:
@Test
fun init_shouldShowFirstQuestion_whenFactoryReturnsSuccess() {
whenever(game.nextQuestion()).thenReturn(question)
verify(questionObserver).onChanged(eq(question))
}
Run it to make sure that if fails. Now, modify the onSuccess() callback of
initGame as follows:

questionLiveData.value = game.nextQuestion()
}
Run the test and make sure that it passes.
You are going to want to show the next question when calling
nextQuestion. Once again, you will start by adding a test as follows:
@Test
fun nextQuestion_shouldShowQuestion() {
val question1 = mock<Question>()
val question2 = mock<Question>()
whenever(game.nextQuestion())
.thenReturn(question1)
.thenReturn(question2)
viewModel.nextQuestion()
verify(questionObserver).onChanged(eq(question2))
}
Here, you can see you're stubbing the nextQuestion() method from a Game
to first return question1 and then question2.
To make it compile add the nextQuestion() method method to your

ViewModel as follows:
fun nextQuestion() {
// TODO
}
Now run your test to make sure that it fails. To fix it, replace your
nextQuestion() with the following implementation:
fun nextQuestion() {
game?.let {
questionLiveData.value = it.nextQuestion()
}
}
Then, inside your onSuccess() in initGame() modify it as follows:

[email protected] = game
nextQuestion()
}
Finally, add the game variable to the class:

private var game: Game? = null
Now, run your test and it will pass.
You have one more piece of functionality to implement. Answering a

question should delegate to the answer() method of the Game, save the
high score, and show the next question and score — in that order. Start
off by adding this test:
@Test
fun answerQuestion_shouldDelegateToGame_saveHighScore_showQuestionAndSc
whenever(game.score).thenReturn(score)
viewModel.answerQuestion(question, "VALUE")
inOrder(game, repository, questionObserver, scoreObserver) {

verify(game).answer(eq(question), eq("VALUE"))
verify(repository).saveHighScore(any())
verify(scoreObserver).onChanged(eq(score))
verify(questionObserver).onChanged(eq(question))
}
}
Notice, here, that you're using inOrder() again to check the methods are
called exactly in the specified order.
Add the answerQuestion() method, to make it compile:
fun answerQuestion(question: Question, option: String) {

}
Now, run the test to make sure that it fails. Finally, add the
corresponding implementation:
fun answerQuestion(question: Question, option: String) {
game?.let {
it.answer(question, option)
repository.saveHighScore(it.score.highest)
scoreLiveData.value = it.score
questionLiveData.value = question
}
}
Run the test and check it passes.
Mockito annotations
Instead of calling the mock() and spy() methods, you can use annotations.
For example, open RepositoryUnitTests.kt and modify the class
definition, variable definitions and setup functions to look like the
following:
@RunWith(MockitoJUnitRunner::class)
@Mock
private lateinit var api: CocktailsApi
@Mock
private lateinit var sharedPreferences: SharedPreferences
@Mock
private lateinit var sharedPreferencesEditor:
SharedPreferences.Editor
@Before
fun setup() {
repository = CocktailsRepositoryImpl(api, sharedPreferences)

}
Note: Be sure to import org.mockito.junit.MockitoJUnitRunner when

asked.
The @RunWith(MockitoJUnitRunner::class) annotation is to instruct that

you are going to write tests using Mockito. Now, you can annotate using
@Mock every property that you'll later use as mocks. Notice that in the
setup() method, you removed the calls to mock for each property.
Run the tests and they will still pass.
You've been doing a lot of work getting logic correct in your app. To see it
work in the UI, un-comment the commented implementation in
CocktailsGameActivity.kt, CocktailsGameViewModelFactory.kt and
CocktailsApplication.kt and run the app.
You now have a well tested working cocktail game with the help of TDD.
Challenge
Challenge: Writing another test
When answering incorrectly three times, it should finish the game.
When answering correctly three times sequentially, it should start
giving double score.
Write a test for each one and add the corresponding functionality
progressively to make each test pass.
Key points
With JUnit you can do state verification, also called black-box
testing.
With Mockito you can perform behavior verification or white-box

testing.
Using a mock of a class will let you stub methods simulate a

particular situation in a test. It'll also verify if one or more methods
were called on that mock.
Using a spy is similar to using a mock, but on real instances. You'll

be able to stub a method and verify if a method was called just like a
mock, but also be able to call the real methods of the instance.
Remember: Red, Green, Refactor

Awesome! You've just learned the basics of unit testing with Mockito.
Check the materials for the final and challenge versions of the code of
this chapter.
Check the following references to know more about the topic:
Test Double Patterns by Gerard Meszaros:

http://xunitpatterns.com/Test%20Double%20Patterns.html
Mocks Aren't Stubs by Martin Fowlder:

https://martinfowler.com/articles/mocksArentStubs.html
Chapter 8: Integration
As mentioned in Chapter 4, "The Testing Pyramid," integration
tests perform checks on how different parts of your app interact
together. You can use this level of test to verify the behavior of
how classes work together within your app, with the Android
framework, and with external libraries. It’s the next level up from
unit tests.
Unit tests are great for ensuring that all your individual pieces
work. Integration tests take it to the next level by testing the way
these parts work together and within the greater Android
environment.
In this chapter, you’ll:
Learn what an integration test is and where are the best

places to use them.
Understand the dependency many integration tests have on

the Android framework and how to handle it.
Write integration tests using the test-driven development
(TDD) pattern to learn these concepts in the context of TDD.
Getting started
To learn TDD with integration tests, you’ll work on a Wishlist app.
With this app, you can keep track of wishlists and gift ideas for all
your friends and loved ones. You will continue working on this
app in the next chapter.
Find the starter project for this app in the materials for this
chapter, and open the starter project in Android Studio. Build and
run the app. You'll see a blank screen with a button to add a list
on the bottom. Clicking the button, you see a field to add a name
for someone's wishlist. Enter all you want right now, but it won't
save yet! You'll be able to see it when you finish the
implementation. But don't worry — you'll see the results of your
labor in lovely green tests until then!
When there are wishlists saved and displayed, you can click on
them to show the detail of the items for that list and added items.
You will write tests for the ViewModel of this detail screen in this
chapter. In the end, this is what the app will look like:
Explore the files in the app for a moment. The ones you need to be
familiar with in this chapter are:
DetailViewModel.kt: This contains the logic for the detail

screen, and is the class you will be testing.
Repository.kt: This is the interface for the data repository in

this app.
RepositoryImpl.kt: This is the implementation of the

Repository interface.
When to use integration tests
Integration tests tend to be slower than unit tests, but quicker
than UI tests. Because of this, you want to first put everything you
can into unit tests. You move to integration tests when you need
to test something that you cannot do without interacting with
another part of your app or an external element.
Generally, when you want to create a unit test but can't quite test
it in isolation, you want to use an integration test. Sometimes you
can get away with extracting the logic out so it can be unit tested
(and is preferable), but the integration test is inevitable at times.
While balancing the lean towards more unit tests, you also need
to rely on integration tests to make sure that each of your
thoroughly tested units works well together with the others. If
your database works perfectly, as does your view model, it's still
useless if the code linking them fails!
Testing with the Android framework

One of the most frequent dependencies that force you to use an
integration test is the Android framework. This does not
necessarily mean it uses the screen; it can be any component of
the SDK. When your code ends up interacting with Android, you
can't get away with unit tests. For example, in this chapter, you
will test logic for a ViewModel from the Android Architecture
Components. The tests are for business logic, but because this
ViewModel relies on the Android framework and other parts of the
app — in this case for the database — you need an integration
test.
Note: You may know that it is possible to test a ViewModel as a
unit test. You saw an example in Chapter 7, “Introduction to
Mockito.” For the purpose of this exercise, you are
specifically testing the ViewModel in an integration fashion —
testing that it works with its dependencies rather than
mocking them. If you want to see an example of how to test a
similar ViewModel as a unit test, take a look at
MainViewModelTest.kt in the project for this chapter.
When running tests that require the Android framework you have
two options:
1. Run them on an Android device or emulator.
2. Use Robolectric.
Robolectric is a framework that allows you to run Android-

dependent tests in a unit-test way. It creates a sandbox in which
to run your tests; the sandbox acts like an Android environment
with Android APIs. A benefit to using Robolectric is that it is
faster, running on the JVM; using a device/emulator, however,
more accurately shows you how your code will behave when
installed onto a device, having more Android features.
This chapter will show you how to use either a device/emulator or

Robolectric. However, note that the final sample project is taking
the emulator approach.
Creating a test class

Create a file called DetailViewModelTest.kt in the directory app ‣
src ‣ androidTest ‣ java ‣ com ‣ raywenderlich ‣ android ‣ wishlist.
The key here is that it mimics the location of DetailViewModel.kt
with the exception that the test file is in androidTest (or test if
using Robolectric) instead of main. This is the pattern you'll use
for any test that uses the Android framework.
Using the Create Test shortcut

There's also a shortcut to create test files as an alternative to
manually creating them. Open DetailViewModel.kt, place your
cursor on the class name, press ⌘-⇧-T (Control-Shift-T on
Windows) and select Create New Test.... Android Studio then
shows you a dialog to create this file for you.
Select Ok to accept the defaults, then, select the androidTest

option (or test when using Robolectric for writing a unit test), and
the editor does the rest.
Setting up the test class
If it's not there already, make sure you have the empty test class in
your file:
class DetailViewModelTest {
}
Great! Now, make sure you have what you need to run your
ViewModel test. Add the following test rule to your test class:
@get:Rule
var instantTaskExecutorRule = InstantTaskExecutorRule()
This is the same TestRule you used in Chapter 7, "Introduction to

Mockito" to make sure that when you're using LiveData with the
ViewModel it's all run synchronously in the tests.
Two more things before you're able to write a test: 1. You need an
instance of the DetailViewModel to perform the tests on; 2. You'll
need the dependency to create it. Add the following to your test
class:
// 1
private val wishlistDao: WishlistDao = Mockito.spy(
Room.inMemoryDatabaseBuilder(
InstrumentationRegistry.getInstrumentation().context,
WishlistDatabase::class.java).build().wishlistDao())
// 2
private val viewModel = DetailViewModel(
RepositoryImpl(wishlistDao))
In the above, you:
1. Create a spy of WishlistDao to use to create the Repository

dependency for the DetailViewModel. You're using Mockito to
create the spy, as described in Chapter 7, "Introduction to
Mockito." Don't worry about understanding how you are
creating the WishlistDao here. You'll learn about that in the
next chapter, Chapter 9, "Testing the Persistence Layer." You
could mock the repository here, instead, but in this example
you will test their interaction. To see an example of what a
test would look like with the Repository mocked, take a look
at MainViewModelTest.kt.
1. Create a DetailViewModel, with a RepositoryImpl created from

your spy.
Note: You may have caught that you are spying on a final
class, which generally does not work with Mockito. While in
Chapter 7, "Introduction to Mockito" you used the mock-
maker-inline extension, in this project, you're using
dexmaker-mockito-inline to make this possible on Android.
You can learn more about this at
https://github.com/linkedin/dexmaker. The dependency is
already added for you in the build.gradle file:
com.linkedin.dexmaker:dexmaker-mockito-inline:2.21.0.
Because of the use of Dexmaker, you need to run the tests on
a device or emulator with Android P or higher. There are
ways to get around this restriction so you can test on other
devices, but you will not use them in this chapter. Take a
moment to create an emulator with the latest version of
Android if you do not yet have an emulator or device with
Android P or higher.
Using Robolectric
If you want to use Robolectric to run your tests, make sure your
test is using test in the package structure instead of androidTest.
Note: The materials included for this chapter use the

device/emulator approach rather than use Robolectric. All
Robolectric steps are optional and are there to help you if
you're interested in using it.
Then, add the following code in app ‣ build.gradle, then sync:

android {
testOptions {
unitTests {
includeAndroidResources = true
}
}
}
dependencies {
testImplementation 'org.robolectric:robolectric:4.2'
}
In DetailViewModelTest.kt, right above the class declaration add:
@RunWith(RobolectricTestRunner::class)
And before you call .build() to create your wishlistDao, add

.allowMainThreadQueries() to the chain. It now looks like this:
private val wishlistDao: WishlistDao = Mockito.spy(

Room.inMemoryDatabaseBuilder(
WishlistDatabase::class.java)
.allowMainThreadQueries()
.build().wishlistDao())
One last thing: Add a file called org.mockito.plugins.MockMaker

to app ‣ src ‣ test ‣ resources ‣ mockito-extensions with this line
like you did in Chapter 7, "Introduction to Mockito":
mock-maker-inline
Dexmaker only works on device, and this will allow you to spy on
that final class while using Robolectric off device.
Writing a failing integration test

In the fashion of TDD, write some tests before adding the
implementation of DetailViewModel. Starting with the
saveNewItem() function, write a test that verifies that saving a new
item calls the database using the Data Access Object (DAO):
@Test
fun saveNewItemCallsDatabase() {
// 1
viewModel.saveNewItem(Wishlist("Victoria",
listOf("RW Android Apprentice Book", "Android phone"), 1),
"Smart watch")
// 2
verify(wishlistDao).save(any())
}
Here, you:
1. Call the saveNewItem() function with data. You can use your
own name and wishes if you like!
2. Verify that saveNewItem() called the save() function on the

DAO using the same technique learned in Chapter 7,
"Introduction to Mockito."
Build and run your test to see it fail. Remember, you need to run
these tests on a device or emulator with Android P or higher
because of the use of Dexmaker.
This error is saying that the save() function was never called on
the wishlistDao. You know what you need to do to make it pass —
keep moving on!

Next step! The function needs to call save() on the DAO (and only
call save()). Add the following to saveNewItem() in DetailViewModel:
repository.saveWishlist(Wishlist("", listOf()))
Run the test, and see it pass.
Testing the wishlist's save functionality

Repeat the TDD pattern for three more tests:
1. One similar to the above test that makes sure saveNewItem()

saves the correct data.
2. One that getWishList() calls the database using the DAO.
3. One that ensures correct data returns when calling

getWishList().
Add the first test to your test class:

@Test
fun saveNewItemSavesData() {
// 1
val wishlist = Wishlist("Victoria",
listOf("RW Android Apprentice Book", "Android phone"), 1)
// 2
val name = "Smart watch"
viewModel.saveNewItem(wishlist, name)
// 3
val mockObserver = mock<Observer<Wishlist>>()
// 4
wishlistDao.findById(wishlist.id)
.observeForever(mockObserver)
verify(mockObserver).onChanged(
wishlist.copy(wishes = wishlist.wishes + name))
}
With this test, you:
1. Create a new wishlist.
2. Create a new item name for the list and call saveNewItem().
3. Create a mock Observer to use as you've done in Chapter 7,

"Introduction to Mockito."
4. Query the database and ensure that the wishlist you just
saved returns, signaling it saved correctly. When the program
posts a value to a LiveData, the object calls onChanged() with
the value. This is the function you are checking for.
It's the same pattern: creating test data as your set up, calling the
function, then verifying the result.
Build and run the test to make sure it fails before modifying the
DetailViewModel to make it pass.
It's the expected error you see. The function called save(), but it
saved the wrong thing!
Once you see that it's failing, change the body of saveNewItem() to
make it pass:
repository.saveWishlistItem(
Note: You may be wondering why you're making a copy of the

wishlist here instead of mutating it. This is to follow the
safety of immutability. You can learn more about this in
"Functional Programming for Android Developers"
https://medium.freecodecamp.org/functional-programming-
for-android-developers-part-2-5c0834669d1a.
Run it again to see it pass.

Testing the database queries
Add the next test to ensure that getWishlist() calls the database:
@Test
fun getWishListCallsDatabase() {
viewModel.getWishlist(1)
verify(wishlistDao).findById(any())
}
This looks very similar to your first test in this class. Run it and
see it fail.
There's the message that says you need to implement that call to
the wishlistDao.
Once you've run your failing test, change the code in getWishList()
to make it pass:
return repository.getWishlist(0)
Run all your tests that you've written in this chapter. Hooray! All
three tests are passing!
Testing the data returned
Your last test in this chapter is to make sure getWishlist() returns
the correct data. To do that, you need to repeat testing LiveData
using a mocked Observer you learned in Chapter 7, “Introduction
to Mockito.”
Add this final test to your test class:
@Test
fun getWishListReturnsCorrectData() {
// 1
val wishlist = Wishlist("Victoria",
listOf("RW Android Apprentice Book", "Android phone"), 1)
// 2
wishlistDao.save(wishlist)
// 3
val mockObserver = mock<Observer<Wishlist>>()
viewModel.getWishlist(1).observeForever(mockObserver)
// 4
verify(mockObserver).onChanged(wishlist)
}
Taking each part in turn:
1. Again, you set up your test data.
2. Save a wishlist to the database to be retrieved later in this

test.
3. Create a mockObserver, mocking a lifecycle Observer. You use

this observer to observeForever() on the LiveData that
getWishlist() returns.
4. Verify that the function published the correct data.
Run your test to make sure it fails.
This error says that getWishlist() is returning null! Sure, you

mocked it, but only when you use an id of 1, the id you're passing
into getWishlist(). If you look in the DetailViewModel, right now
there is a 0 hardcoded in for the id. Change the body of the
getWishlist() function to use the id that's passed in:
return repository.getWishlist(id)
Run the test again. You can do this by clicking on the Play button
near the name of the test class.
All green! Great job!

Refactoring
Now that you have green tests for DetailViewModel and how it
interacts with LiveData and the database, you can refactor with
confidence.
Take a look at saveNewItem() in DetailViewModel. It's doing a bit of

work to format the Wishlist for saving:
fun saveNewItem(wishlist: Wishlist, name: String) {

repository.saveWishlistItem(
}
One could argue that this responsibility belongs to the repository.

Why not perform that refactoring!
There are three files you need to change to refactor this, and it
won't compile until you've done all three.
1. In DetailViewModel change the contents of saveWishlistItem()

to be:
repository.saveWishlistItem(wishlist, name)
1. In the Repository interface, change the saveWishlistItem()

signature to this:
fun saveWishlistItem(wishlist: Wishlist, name: String)

1. In RepositoryImpl, change the body of saveWishlistItem() to
have the logic you deleted from the DetailViewModel, and the
signature to match the interface:
override fun saveWishlistItem(

wishlist: Wishlist,
name: String
) {
wishlistDao.save(
}
Run the tests to make sure nothing broke during the refactor.
It didn't — congratulations!
Running the app

After all your hard work, you can see your app in action! Build and
run the app, and play around with creating wishlists with items.
Key points
Integration tests verify the way different parts of your app
work together.
They are slower than unit tests, and should therefore only be
used when you need to test how things interact.
When interacting with the Android framework you can rely

on an Android device or emulator, or use Robolectric.
You can use dexmaker-mockito-inline to mock final classes

for Android tests.

You can find the final version of the code in this chapter in the
chapter materials.
If you want to continue exploring integration tests in this app,

take a look at MainViewModelTest.kt at the tests already written
there.
There's much more that you can do with integration tests than
ViewModel tests. In Chapter 9, "Testing the Persistence Layer,"
you'll learn how you can test your persistence layer, and Chapter
10, "Testing the Network Layer," introduces the network layer.
For more about integration testing, you can look at the Android
documentation:
https://developer.android.com/training/testing/fundamentals
#medium-tests
https://developer.android.com/training/testing/integration-
testing/
Chapter 9: Testing the
Persistence Layer
In most apps you'll build, you will store data in one way or
another. It might be in shared preferences, in a database, or
otherwise. No matter which way you're saving it, you need to be
confident it is always working. If a user takes the time to put
together content and then loses it because your persistence code
broke, both you and your user will have a sad day.
You have the tools for mocking out a persistence layer interface
from Chapter 7, "Introduction to Mockito." In this chapter you will
take it a step further, testing that when you interact with a
database, it behaves the way you expect.
In this chapter you will learn:
How to use TDD to have a well tested Room database.
Why persistence testing can be difficult.
Which parts of your persistence layer should you test.
Note: It is helpful, but not necessary to have a basic

understanding of Room. To brush up on the basics, check out
our tutorial, "Data Persistence with Room":
https://www.raywenderlich.com/69-data-persistence-with-
room.
Getting started
To learn about testing the persistence layer you will write tests
while building up a Room database for the Wishlist app. This app
provides a place where you can keep track of the wishlists and the
gift ideas for all your friends and loved ones.
To get started, find the starter project included for this chapter
and open it up in Android Studio. If you are continuing from
Chapter 8, "Integration," notice there are a couple differences
between the projects. It is recommended you continue by using
the starter project for this chapter. If you choose to continue with
your project from Chapter 8, "Integration," copy and override the
files that are different from the starter project for this chapter.
The files you'll need to copy are WishlistDao.kt, KoinModules.kt,
RepositoryImpl.kt and WishlistDatabase.kt.
Build and run the app. You'll see a blank screen with a button to
add a list on the bottom. Clicking the button, you see a field to
add a name for someone's wishlist. However, if you try to save
something right now it won't work! You will implement the
persistence layer to save the wishlist in this chapter.
When there are wishlists saved and displayed, you can click on
them to show the detail of the items for that list, and add items.
By the end of this chapter, this is what the app will look like:
Time to get familiar with the code.
Exploring the project

There are a couple files you should be familiar with before getting
started. Open these files and take a look around:
WishlistDao.kt: Is the database access object. You will work

on defining the database interactions in this class. This is
also the class you will write your tests for.
RepositoryImpl.kt: This class should be familiar to you from

Chapter 8, "Integration." It's the repository that hooks your
app up with the database.
KoinModules.kt: This handles the dependency injection for

the app, specifying how to create any dependencies.
StringListConverter.kt: This is a helper object to convert lists
of Strings to a single String and back again for the purposes
of storing in the database.

As with any test, the first thing you need to do is create the file.
Create WishlistDaoTest.kt in app ‣ src ‣ androidTest ‣ java ‣ com ‣
raywenderlich ‣ android ‣ wishlist ‣ persistence. In it, create an
empty class with the @RunWith annotation. The import you want
for AndroidJUnit4 is androidx.test.ext.junit.runners.AndroidJUnit4:
@RunWith(AndroidJUnit4::class)
class WishlistDaoTest {
}
The AndroidJUnit4 class you're including here is a cross

environment JUnit4 runner for Android tests. Note that this is
only required when using a mix of JUnit3 and JUnit4.
Continuing your set up, add the following test rule to your test
class:
@get:Rule
var instantTaskExecutorRule = InstantTaskExecutorRule()
Android Architecture Components uses an asynchronous

background executor to do its work. InstantTaskExecutorRule is a
rule that swaps out that executor and replaces it with a
synchronous one. This will make sure that, when you're using
LiveData, it's all run synchronously in the tests.
You also need to create the properties to hold your

WishlistDatabase and WishlistDao. Add these properties to your test
class now:
private lateinit var wishlistDatabase: WishlistDatabase
private lateinit var wishlistDao: WishlistDao
The WishlistDao is what you are performing your tests on. To

create an instance of this class, you'll need an instance of the
WishlistDatabase first. You will initialize these in a @Before block in
a moment.
Using an in-memory database

One of the challenges that make writing persistence tests difficult
is managing the state before and after the tests run. You're testing
saving and retrieving data, but you don't want to end your test run
with a bunch of test data on your device or emulator. How can you
save data while your tests are running, but ensure that test data is
gone when the tests finish? You could consider erasing the whole
database, but if you have your own non-test data saved in the
database outside of the tests, that would delete too.
Tests must be repeatable. This means you should be able to run a

test multiple times with the same result. There's also a
requirement that one test cannot influence the outcome of
another. What if you're testing for an empty database but there
are items left over from another test? You will need to clear data
between tests.
You can solve this problem by using an in-memory database.

Room luckily provides a way to easily create one. Add this to your
test class, importing
androidx.test.platform.app.InstrumentationRegistry:
@Before
fun initDb() {
// 1
wishlistDatabase = Room.inMemoryDatabaseBuilder(
WishlistDatabase::class.java).build()
// 2
wishlistDao = wishlistDatabase.wishlistDao()
}
1. Here you're using a Room builder to create an in-memory

WishlistDatabase. Compare this to the database creation in
KoinModules.kt. Information stored in an in-memory
database disappears when the tests finish, solving your state
issue.
2. You then use this database to get your WishlistDao.
Almost done setting up! After your tests finish, you also need to
close your database. Add this to your test class:
@After
fun closeDb() {
wishlistDatabase.close()
}
Now you're ready to start writing tests.
Writing a test
Test number one is going to test that when there's nothing saved,
getAll() returns an empty list. This is a function for fetching all of
the wishlists from the database. Add the following test, using the
imports androidx.lifecycle.Observer for Observer, and
com.nhaarman.mockitokotlin2.* for mock() and verify():
@Test
fun getAllReturnsEmptyList() {
val testObserver: Observer<List<Wishlist>> = mock()
wishlistDao.getAll().observeForever(testObserver)
verify(testObserver).onChanged(emptyList())
}
This tests the result of a LiveData response similar to how you

wrote your tests in Chapter 7, "Introduction to Mockito." You
create a mock Observer, observe the LiveData returned from
getAll() with it, and verify the result is an empty list.
You have one error right now, that getAll() is unresolved. Add the
following to WishlistDao:
fun getAll(): LiveData<List<Wishlist>>
Everything looks pretty good, so try to run the test. Oh no!

There's still a compiler error.
Hmmm. The next minimum thing to make this test run is to add a
@Query annotation. Add an empty query annotation to getAll():
@Query("")
Try running it with this change. If you are familiar with Room,
you may know what's coming.
The compiler enforces that you include a query in the parameter.

That means the next step is to fill in the query as simply as
possible. Fill in your query with the following:
@Query("SELECT * FROM wishlist")
Run your test and it finally compiles! But... it's passing. When
practicing TDD you always want to see your tests fail first. You
never saw a state where the test was compiling and failing. You
were so careful to only add the smallest bits until it compiled.
Room made it really hard to write something that didn't work.
Maybe the real question is "Should you be testing this?" The
answer to this question is important.
Knowing not to test the library

The fact that Room made it hard to write a failing test is a clue.
When your tests align closely with a library or framework, you
want to be sure you're testing your code, and not the third-party
code. If you really want to write tests for that library, you might be
able to contribute to the library, if it's an open source project. :]
Sometimes this is a gray line to try to find, and it's one of the
things that makes testing persistence difficult. It's a part of your
code that likely relies heavily on a framework.
It's a balance to test that your interactions with the framework or

library are correct without testing the library itself. Watch out for
cases like these and use your best judgement. Over time you'll
gain some kind of intuition on which tests are valuable, and what
tests are better left to the library's contributors.
In this case it's not up to you to test the Room framework. It's
those writing Room's responsibility to make sure that when the
database is empty, it returns nothing. Instead you want to test
that your logic, your queries, and the code that depends on them
are working correctly.
With that in mind, you can move on to test other database

interactions.
Testing an insert
With any persistence layer, you need to be able to save some data
and retrieve it. That's exactly what your next test will do. Add this
test to your class, keeping in mind that save() is not yet resolved:
@Test
fun saveWishlistsSavesData() {
// 1
val wishlist1 = Wishlist("Victoria", listOf(), 1)
val wishlist2 = Wishlist("Tyler", listOf(), 2)
wishlistDao.save(wishlist1, wishlist2)
// 2
// 3
val listClass =
ArrayList::class.java as Class<ArrayList<Wishlist>>
val argumentCaptor = ArgumentCaptor.forClass(listClass)
// 4
verify(testObserver).onChanged(argumentCaptor.capture())
// 5
assertTrue(argumentCaptor.value.size > 0)
}
Here you:
1. Create a couple wishlists and save them to the database. At

this point save() does not exist yet, so there will be an error.
2. Use your mock testObserver again to call getAll().
3. Create an ArgumentCaptor to capture the value in onChanged().

Using an ArgumentCaptor from Mockito allows you to make
more complex assertions on a value than equals().
4. Test that the result from the database is a non empty list. At
this point you care that data was saved and not what was
saved, so you're checking the list size only.
Great! Next, to make it compile and run, you need to add a save()
function to the WishlistDao:
@Delete
fun save(vararg wishlist: Wishlist)
You need to have a database interaction annotation in order for

this to compile, as you learned earlier in this chapter. You also
want to see this test failing, so you're using the wrong one,
@Delete. Run your test and see it fail.
You know the drill, time to make this test green!
Making your test pass

This one is simple enough to make it pass. Just change the @Delete
annotation with an @Insert. Your save() signature should now look
like this:
@Insert(onConflict = OnConflictStrategy.REPLACE)
fun save(vararg wishlist: Wishlist)
Using OnConflictStrategy.REPLACE allows the the database to

override an entry that already exists.
Run your tests, and they should all be green.
Testing your query

Now that you have a way to save data in your database, you can
test your getAll() query for real! Add this test:
@Test
fun getAllRetrievesData() {
val wishlist1 = Wishlist("Victoria", emptyList(), 1)
val wishlist2 = Wishlist("Tyler", emptyList(), 2)

val listClass =
ArrayList::class.java as Class<ArrayList<Wishlist>>
val argumentCaptor = ArgumentCaptor.forClass(listClass)
verify(testObserver).onChanged(argumentCaptor.capture())
val capturedArgument = argumentCaptor.value
assertTrue(capturedArgument
.containsAll(listOf(wishlist1, wishlist2)))
}
This is almost the same as your previous test with the exception
of the final line. In that line you're testing that the list result
contains the exact wishlists you expect.
Build and run your tests. It may come as a surprise, but they
failed! Why is that? Insert a debugger breakpoint on the assertion
line and inspect the capturedArgument at that point when you run it
again, using the debugger. Huh! Somehow there is a list with an
empty string in it.
Good investigating! You found the bug before it reached

production. Time to solve the problem.
Fixing the bug

How could this happen? StringListConverter holds the key. Take a
look at the object. In stringToStringList() when there is an empty
String saved in the database, as is the case for an empty list, the
split function used returns a list with an empty string in it! Now
that you know the problem, you can solve it. Replace the body of
stringToStringList() with:
if (!string.isNullOrBlank()) string?.split("|")?.toMutableList()
else mutableListOf()
Now run those tests again and see them pass!
Note: All over while testing you'll be performing verifications

that rely heavily on the equals() method to perform
comparisons, containsAll() being one of them. Much of the
time in these cases you are looking for data equality (the
properties of both objects are exactly the same) rather than
object equality (they reference the same object in memory).
Because of this, you want to make sure your equals()
performs the way you expect. In Kotlin, this is often as simple
as making your class a data class. This overrides equals() and
hashcode() for you to compare the properties. Take caution for
the times where this isn't enough! For example, Kotlin may
not compare Lists the way you expect. For that reason
equals() and hashcode() are overridden for Wishlist in this
app. You can see this in Wishlist.kt.
Testing a new query

Moving on. In your database you also need the ability to retrieve
an item by id. To create this functionality, start by adding a test
for it:
@Test
fun findByIdRetrievesCorrectData() {
// 1
val wishlist1 = Wishlist("Victoria", emptyList(), 1)
val wishlist2 = Wishlist("Tyler", emptyList(), 2)
// 2
val testObserver: Observer<Wishlist> = mock()
wishlistDao.findById(wishlist2.id).observeForever(testObserver)
verify(testObserver).onChanged(wishlist2)
}
Here you:
1. Create and save some wishlists, same as your other tests.
2. Query for a specific wishlist, wishlist2, and verify the result is

correct.
Now to write the minimum code to make it compile. Add this to

the WishlistDao:
@Query("SELECT * FROM wishlist WHERE id != :id")

fun findById(id: Int): LiveData<Wishlist>
Notice it's intentionally incorrect. It's searching for a wishlist

where the id is not the given id. This is again to make sure you see
a failing test.
Run that test and verify it really does fail.

It's the last time you'll do it this chapter: make that test green! All
you need to do is remove that not (!) from the query. It should
now look like this:
@Query("SELECT * FROM wishlist WHERE id = :id")
Ready? Run your tests to see them all pass.
Creating test data

You have a working database with reliable tests but there's more
that you can do. There is something you can do to also help save
set up time as you write other tests. This tool is called test data
creation.
If you look at the tests you've written in this chapter, as well as

Chapter 8, "Integration," you see many lines where you're
manually creating a Wishlist. Not only is this tedious, but you're
only testing that your code works for that specific Wishlist.
One way to abstract this work and make your data random, is by
using a Factory. A Factory object will create instances of your data
class with random values for the properties. This will make your
tests stronger and easier to write!
Start by creating a WishlistFactory object in your test directory. As

this is specific to your persistence right now, a good place to put it
is app ‣ src ‣ androidTest ‣ java ‣ com ‣ raywenderlich ‣ android ‣
wishlist ‣ persistence ‣ WishlistFactory.kt:
object WishlistFactory {
}
If you use this in other places as well, you can move this Factory
to a more convenient location. Since you're only using it in this
one test right now, this location works great.
You need a way to create random values for your data class
properties. A simple way to do this is to create helper methods to
create them, one for each type of property you need. Again, you
could place these in a reusable location, but because you're only
using them here right now, they can share the WishlistFactory.
In your Wishlist you need two types of data: String and Int. Add
these methods to your WishlistFactory:
// 1
private fun makeRandomString() = UUID.randomUUID().toString()
// 2
private fun makeRandomInt() =
ThreadLocalRandom.current().nextInt(0, 1000 + 1)
These are simple, built in ways to create random values. You can
use similar ways to create helpers for Long, Boolean, etc.
Now, to finish the Factory, add a method to create a Wishlist:
fun makeWishlist(): Wishlist {

return Wishlist(
makeRandomString(),
listOf(makeRandomString(), makeRandomString()),
makeRandomInt())
}
You use the random value methods you just created to set the
properties, knowing you will likely have a completely different
Wishlist every time you create one. They won't look anything like
what's on your wishlist, but they will be unique. Well, the Wishlist
won't have what you want unless you want a UUID for your
birthday.
Using a Factory in your test
You now have an easy way to create test data, so why not use it?
Refactor your tests so that each time you create a Wishlist, you
use the factory instead. It should look like this in each of your
tests:
val wishlist1 = WishlistFactory.makeWishlist()

val wishlist2 = WishlistFactory.makeWishlist()
So clean! Run your tests to make sure they still pass.
Hooking up your database

You now have beautiful, tested database interactions, so surely
you want to see them in action! Before you run the app, open up
RepositoryImpl and change the body of the functions to match the
following:
override fun saveWishlist(wishlist: Wishlist) {

wishlistDao.save(wishlist)
}
override fun getWishlists(): LiveData<List<Wishlist>> {

return wishlistDao.getAll()
}
override fun getWishlist(id: Int): LiveData<Wishlist> {

return wishlistDao.findById(id)
}
override fun saveWishlistItem(

wishlist: Wishlist,
name: String
) {
wishlistDao.save(
}
All this is doing is hooking up the repository to call the newly
implemented methods on the WishlistDao. You're ready to build
and run the app!
Note: If the RepositoryImpl looks familiar to you, that's

because you saw it in Chapter 8, "Integration." The
implementation was removed for the start of the chapter as
the WishlistDao interface was emptied so you could test it.
Play around with your fully functioning app! Create some

Wishlists and add some items to them. You'll always know the
perfect gift to give now.
Handling stateful tests

In this chapter you learned hands on how to handle the
statefulness of your tests using an in-memory database. You need
this set up and tear down to write reliable, repeatable persistence
tests, but how do you handle it when you're using something
other than Room for your persistence?
Unfortunately, many libraries don't have these convenient, built-

in testing helpers. Some do, such as Realm, but often you're left in
the dust. In these cases you're usually left to clear the persisted
data before each test. When doing this, make sure your testing
device doesn't have any data you want to keep for that app!
Key points
Persistence tests help keep your user's data safe.
Statefulness can make persistence tests difficult to write.
You can use an in-memory database to help handle stateful

tests.
You need to include both set up (@Before) and tear down

(@After) with persistence tests.
Be careful to test your code and not the library or framework

you're using.
Sometimes you need to write "broken" code first to ensure

that your tests fail.
You can use Factories to create test data for reliable,

repeatable tests.
If the persistence library you're using doesn't have built in

strategies for testing, you may need to delete all persisted
data before each test.
You now know how to get started testing your persistence layer in
your app. Keep these strategies in mind whenever you're
implementing this layer.
As with all of programming, there are often many ways to do the

same thing. For another example of how to test RoomDB, take a
look at "Room DB: Advanced Data Persistence"
https://www.raywenderlich.com/5686-room-db-advanced-data-
persistence. You may even start to think how you can use the
strategy in this tutorial for the tests you wrote in Chapter 8,
"Integration." ;]
For an example of how to test an SQLite database, take a look at

"Testing persistence in the Android ecosystem"
https://blog.novoda.com/testing-persistence-in-the-android-
ecosystem/.
Moving on to an equally important layer, in the next chapter,

Chapter 10, "Testing the Network Layer," you'll learn strategies for
testing code that relies on API network calls.
Chapter 10: Testing the Network
Layer
You live in a time of the internet, and it's likely that your app
does, too. Most apps connect to an API over the network in one
way or another, giving them a network layer. As this is often a
critical part of your app, it stands to reason that you should test
it. This is what you'll learn to test in this chapter! Along the way
you'll learn:
Tools for testing your network layer.
How to provide reliable test data.
Important things to know about maintaining network layer

tests.
One thing to consider when testing your interaction with data

across a network is that, when running your automated tests, you
don't actually want to make a network call. Network calls are
unpredictable. A call can fail because of the network connection,
the server, the internet service provider and so on. You can't
always know the state of any of these components. You need your
tests to be repeatable and predictable, so this dependency on a
wobbly network won't work, here.
There are tools that you can use to test your network layer
without hitting the network, which is what you will focus on in
this chapter. You will look at three tools to add to your testing
toolbox:
MockWebserver to mock the responses from network

requests
Mockito to mock your API responses
Faker for data creation
Getting started
In this chapter, you will work on the network layer for an app
called Punchline. This is an app that will show you a new, random
joke every time you press a button. To start, find the starter
project in the materials for this chapter and open it in Android
Studio.
Run the app, but you won't see much yet:
There will be no UI for you to play with until the end of Chapter
11, "User Interface." Until then, you'll see your progress made in
the form of green tests for the network layer!
The Punchline app has a single call to the network: the one that
fetches a random joke. You will test this request three times, each
using a different tool. There are a couple of files that you should
have on hand before you get started with your tests:
JokeService.kt: This is the Retrofit service that you will

declare your network requests in.
Repository.kt: This file defines RepositoryImpl. It's the glue

that connects the network layer with the rest of the app.
Joke.kt: Your Joke data model lives here. You can see that it
has values for the ID and joke.
Note: Is it helpful, but not required, to have a familiarity with

Retrofit for this chapter. To learn more, go to "Android
Networking Tutorial: Getting Started"
https://www.raywenderlich.com/2-android-networking-
tutorial-getting-started.
Creating your test file

You can't write tests without a place to put them! First off, create
your test file. Create JokeServiceTest.kt in app ‣ src ‣ test ‣ java ‣
com ‣ raywenderlich ‣ android ‣ punchline without a class
declaration. You'll put all three of your test classes in this file for
easy comparison. Notice that this test is under test and not
androidTest. That's right, you don't need the Android framework
to test your network layer when using these tools! They'll run nice
and fast. You can also use MockWebServer in your Android tests if
there's a place you want to use it in an integration test in your
own apps.
Investigating the API
Most of the Retrofit boilerplate is already set up for you. You can
peek at KoinModules.kt if you're interested in seeing that set up.
What you care about before you test is the specific endpoint you
are testing and implementing.
When interfacing with a network, you often have the endpoints

and responses prescribed for you. Even if you are helping with or
have a say in how they look, they ultimately live outside your app.
Just as you often do in your own app, you have API specifications
to stick to here.
The call you are implementing is rather simple. You make a call to
"random_joke.json", and get a JSON response back. The JSON looks
as follows:
{
"id":17,
"joke":"Where do programmers like to hangout? The Foo Bar.",
"created_at":"2018-12-31T21:08:53.772Z",
"updated_at":"2018-12-31T21:36:33.937Z",
"url":"https://rw-punchline.herokuapp.com/jokes/17.json"
}
You only care about the first two properties for this example —
"id" and "joke" — and will ignore the rest.
Now you have all the knowledge you need to get started with your
test writing!
Using MockWebServer
The first tool you will learn is MockWebServer. This is a library
from OkHttp that allows you to run a local HTTP server in your
tests. With it, you can specify what you want the server to return
and perform verifications on the requests made.
The dependency for MockWebServer is already added in the
project for you. You can see it in app ‣ build.gradle as:
testImplementation "com.squareup.okhttp3:mockwebserver:3.12.0"
To start, you need a test class. Add this empty class to your test
file:
class JokeServiceTestUsingMockWebServer {
}
Setting up MockWebServer
MockWebServer has a test rule you can use for your network tests.
It is a scriptable web server. You will supply it responses and it
will return them on request. Add the rule to your test class:
@get:Rule
val mockWebServer = MockWebServer()
Now, you'll set up your JokeService to test. Because you're using

Retrofit, you'll use a Retrofit.Builder to set it up. Add the following
to your test class:
private val retrofit by lazy {

Retrofit.Builder()
// 1
.baseUrl(mockWebServer.url("/"))
// 2
.addCallAdapterFactory(RxJava2CallAdapterFactory.create())
// 3
.addConverterFactory(GsonConverterFactory.create())
// 4
.build()
}
In the above, you:

1. Set the baseUrl on the builder using the mockWebServer. This is
required when using Retrofit. Because you're not hitting the
network, "/" is perfectly valid, here.
1. Add a call adapter. Using an RxJava call adapter allows you to

return RxJava streams in your JokeService, helping you
handle the asynchronous nature of the network calls. Don't
worry; you don't need to be an RxExpert to keep going!
2. Add a converter factory. This is so you can use Gson to

automatically convert the JSON to a nice Kotlin object, Joke.
3. Build it!
You then use retrofit to create your JokeService! Add this code:
private val jokeService by lazy {

retrofit.create(JokeService::class.java)
}
You're all set to start scripting and testing!
Running the MockWebServer

Your JokeService should have a function, getRandomJoke(), that
returns a random joke. To handle the asynchronous nature of
network calls, you will use RxJava. If you're not familiar with
RxJava, this is all you need to know: getRandomJoke() will return a
Single of type Joke. Parties will subscribe to this Single and receive
an event when it emits a Joke. RxJava brings a lot of power, but,
for the sake of this exercise, you can think of it as a way to
perform a callback.
This test for getRandomJoke() will be called

getRandomJokeEmitsJoke() to match the described functionality.
Add the test function to your class:
@Test
fun getRandomJokeEmitsJoke() {
}
There's nothing in it yet, but run the test anyway. There's some
interesting output in the console.
The MockWebServer starts up at the beginning of your test and

closes at the end. Anywhere in-between the beginning and end,
you can script requests, make those requests to get the response,
and perform verifications.
Scripting a response
Now that you have MockWebServer set up to receive requests, it's
time to script something for it to return!
There are two ways to set up the JSON to script a response with.
One way is to pull in the JSON from a file. This is a great option if
you have a long expected response or you have a real response
from a request that you want to copy paste in. You won't use this
first way in this chapter, but know that you can place this JSON in
a *.json file under app ‣ src ‣ test ‣ resources, then use
getJson("file/path.json") from MockWebServer to fetch it. For
example, if you had a app ‣ src ‣ test ‣ resources ‣ joke ‣
random_joke.json, you would call
getJson("joke/random_joke.json").
Because it's such a short response and because it will allow you to
dynamically build it, you will create your JSON String in your test
file.
Start by creating a property with a JSON string at the class level:
private val testJson = """{ "id": 1, "joke": "joke" }"""
Notice the use of triple quotes to create raw strings. By using raw
strings for your JSON Strings, you don't need to worry about
escaping characters such as the quotes around the JSON
properties.
You have to admit, this JSON is pretty boring, but later you'll spice
it up! For now, it's time to learn how to use this JSON to script a
response!
Add this to your empty getRandomJokeEmitsJoke() test:
// 1
mockWebServer.enqueue(
// 2
MockResponse()
// 3
.setBody(testJson)
// 4
.setResponseCode(200))
Going over these step-by-step:
1. Use the mockWebServer that you created before to enqueue a

response.
2. You enqueue a response by building and passing in a

MockResponse object.
1. Use the testJson that you created as the body of the response.
2. Set the response code to 200 — success!
There are many other things you can set on this MockResponse to
test different situations. For example, you can set headers and use
throttleBody() to simulate a slow network!
One other thing to note, as the name suggests, you can enqueue
multiple responses in a row to return with each new request. This
could be helpful when you have an integration test that hits
multiple different endpoints and combines the results.
Writing a MockWebServer test

Phew! With all that set up, it's finally time to finish writing your
test. Add these two lines to the bottom of
getRandomJokeEmitsJoke(). There will be an error at getRandomJoke()
because you haven't created it yet:
// 1
val testObserver = jokeService.getRandomJoke().test()
// 2
testObserver.assertValue(Joke("1", "joke"))
Here, you:
1. Call getRandomJoke() on your jokeService. By chaining test()

you get a TestObserver that you can use to verify the value of
the Single that getRandomJoke() returns.
2. Verify that the value that returns is a Joke object with the
same values you placed in the testJson and enqueued with
MockWebServer.
Next step of the TDD process: Write just enough code so you can
compile and run your test. Add getRandomJoke() to the JokeService
interface with a return value of Single<Joke>:
fun getRandomJoke(): Single<Joke>

Note: When you're writing tests and need to create a method,
property, etc. that doesn't exist yet, you can use the shortcut
Option-Return on Mac or Alt-Enter on Windows to pop up a
dropdown with options to auto create it for you.
Build and run your test! As you may have expected if you went
through Chapter 9, "Testing the Persistence Layer," and have some
familiarity with Retrofit, you'll get an error that Retrofit requires
an HTTP method annotation for your new method:
Your goal is to make this run, so next you add an annotation! Add
the @GET annotation to your JokeService method:
@GET("https://raywenderlich.com")
fun getRandomJoke(): Single<Joke>
The @GET annotation requires a path or URL. When you pass in a

full URL like "https://raywenderlich.com" it uses that URL, but if
you pass in a path like "joke.json" it uses the base URL appended
with "joke.json" for the URL. To make sure you see your test fail,
you're passing in a full URL. Remember, right now you have a
response enqueued for any endpoint given with the base URL, so
giving it something without the base URL will return an empty
response. Run it and see it fail:
That's no joke! No wonder, you're calling the wrong URL. Update

the parameter so the test can pass:
@GET("joke.json")
Run it, and you're all green!
Refactoring your test

You may feel like there's a code-smell in the way you're hard
coding the values for the ID and joke. Thankfully, you can change
that! Because you're creating the JSON String in the test, you can
create it the way you like. Make a constant for the ID and the joke.
By putting them outside the test class at the file level you'll be
able to use them in your other tests too:
private const val id = "6"

private const val joke =
"How does a train eat? It goes chew, chew"
Note: Think about how you might use a factory such as in

Chapter 9, "Testing the Persistence Layer" to make these
random values for each test.
Now use these values to update your testJson:
private val testJson = """{ "id": $id, "joke": "$joke" }"""
And your test assertion:
testObserver.assertValue(Joke(id, joke))
Run your test, and it should still pass!
Maintaining test data

You just set up some test data and wrote some tests. Now,
imagine the JSON response had many more properties and you
had more endpoints to test. Then, imagine the format of the
response changed. Maybe instead of joke as a String, it contained
an object with different translations of the joke. You need to make
sure you update your tests with this change when it happens. If
you don't test the new type of response, your tests are no longer
accurate.
This is one of the difficulties of network tests: How do you make

your tests reliable and maintainable? Do you keep a file of your
responses and swap it whenever there's a change? Do you
dynamically create your responses? This will differ depending on
your needs and may change as you figure out what's right for your
app. You'll learn even more about this in Chapter 16, "Strategies
for Handling Test Data"
Testing the endpoint

You may be having some doubts about that last test. If it will pass
with any endpoint with the same base URL, what is it testing? Is it
testing that the response is correctly parsed into a Joke object?
While it's important to know your data is represented correctly,
MockWebServer does help you test the endpoint too! Next you'll add
a test that the endpoint is correct.
Add the following test to your test class:

@Test
fun getRandomJokeGetsRandomJokeJson() {
// 1
mockWebServer.enqueue(
MockResponse()
.setBody(testJson)
.setResponseCode(200))
// 2
val testObserver = jokeService.getRandomJoke().test()
// 3
testObserver.assertNoErrors()
// 4
assertEquals("/random_joke.json",
mockWebServer.takeRequest().path)
}
Here's what's happening:
1. Enqueue the response, same as before.
2. Call getRandomJoke() as before, getting a reference to a

TestObserver.
3. You can also use the testObserver to make sure there were no
errors emitted.
4. Here's what you're writing this for! You can use the
mockWebServer to get the path that was requested to compare it
to what you expect. There are many other things other than
the request path you can test this way!
Run the test, and you're right! It fails!
Update the @GET annotation once more to make this pass:

@GET("random_joke.json")
Build and run your test. It passes! You now know your JokeService
uses the correct endpoint. That's all you'll use of MockWebServer for
this chapter, but you can see how it is a powerful and robust tool
for testing network requests! But what if you don't need that
much detail? That's what you'll learn how to do next.
Mocking the service

Depending on your app and your team, it may be enough to know
that the service methods are available and you're using them
correctly. This can be done using Mockito, which you first learned
in Chapter 7, "Introduction to Mockito." In this test you will also
be concerned with getRandomJoke(), but your test will worry more
about its interaction with the respository.
To start, create a new test class to write your Mockito tests in. This
class can be in the same file as your MockWebServer test if you like:
class JokeServiceTestMockingService {
}
You technically don't need to put your new tests in a new test
class (as long as the test methods themselves have different
names). However, by creating this separate class, it helps keep the
topics divided and allows you to run the new tests without the
server running behind them.
Next, you need to set up your test subject. Add this to your
JokeServiceTestMockingService class. When prompted, import
com.nhaarman.mockitokotlin2.mock:
private val jokeService: JokeService = mock()

private val repository = RepositoryImpl(jokeService)
Here, you're mocking the JokeService, which you then pass into
the constructor of RepositoryImpl.
Now, the test you've been waiting for! Add this test method:
@Test
// 1
val joke = Joke(id, joke)
// 2
whenever(jokeService.getRandomJoke())
.thenReturn(Single.just(joke))
// 3
val testObserver = repository.getJoke().test()
// 4
testObserver.assertValue(joke)
}
Here, you:
1. Create the Joke object to use in this test. Notice that it's using
the constant ID and joke you set up while writing the
MockWebServer tests.
2. Set up your JokeService mock to return a Single with that Joke

when getRandomJoke() is called.
3. The signature of getJoke() on the repository also returns a

Single. Use the same TestObserver strategy you used before
here.
4. Assert that the repository's getJoke() emits the same Joke that
comes from the JokeService you mocked.
You can see this is testing the interactions with the network layer
instead of the network calls themselves.
Run your test and see it fail:

What else now but to make it pass! Open up Repository.kt.
Change the body of getJoke() in RepositoryImpl to be the
following:
return service.getRandomJoke()
Now, you're calling the JokeService as expected. Run that test and
see it pass this time!
With this strategy, you can't (and don't need to) test for the
endpoint. That means you get to move onto the next tool to learn!
Using Faker for test data

Throughout this chapter, you've been using the same boring old
test data:
private const val id = "6"

private const val joke =
"How does a train eat? It goes chew, chew"
This is quite a change from the last chapter when you learned to
use Factories to create random test data! How would you like to
learn another way to generate test data? Now's the time! Your test
will look very similar to the one you just wrote, but with more fun
test data.
One library that helps with this is Faker. With this library, you can
generate data from names and addresses to Harry Potter and
Hitchhiker's Guide to the Galaxy. You can see the full list at
https://github.com/DiUS/java-faker#fakers. The library is already
added to the project for you. You can see it in app ‣ build.gradle
as:
testImplementation 'com.github.javafaker:javafaker:0.16'
To get started, add a test class to your file just as before:
class JokeServiceTestUsingFaker {
}
Then, your setup is like the one for the last test with one addition:
var faker = Faker()

private val jokeService: JokeService = mock()
private val repository = RepositoryImpl(jokeService)
You include an instance of Faker to use to create your test data

this time.
Final test for this chapter! Add this to your new test class:
@Test
val joke = Joke(
faker.idNumber().valid(),
faker.lorem().sentence())
whenever(jokeService.getRandomJoke())
val testObserver = repository.getJoke().test()
testObserver.assertValue(joke)
}
Everything is the same as the last test you write except for the
creation of your Joke object at the start. For this, you're using
faker, which you instantiated above. Many of the Faker methods
return an object that's part of the library. It's from these objects
that you can get a value. For example, the IdNumber object that's
returned from idNumber() has methods to get both valid and
invalid IDs, as well as other forms of ID such as SSN.
Play around with what values you can get from Faker for your
Joke. You don't need to stick with the ID and Lorem ipsum
prescribed here.
Run the test, and it passes! Of course that's because you

implemented this feature with the previous test. Take the
responsibility to revert the changes you made so you can see this
test fail.
Deciding what tools to use

You've learned so many tools in this chapter and in the previous
chapters. How do you decide which to use? When is a unit test
best or when is it time for an integration test? Hopefully, you're
starting to think about how you can mix and match some of these
things, such as how you used both Faker and Mockito for your last
test.
There are some patterns, many of which you see in this book, that
guide your testing decisions. There are also some restrictions
from the libraries themselves on how they can be used.
Ultimately, you have to figure out what works best for your needs
and for your team. Don't be afraid to try out new things or think
again when something isn't working. Where are the holes in your
tests where you aren't catching the bugs? What tests are brittle
and hard to maintain? On the other hand, what tests have been
consistently saving you from deploying buggy code? Watching for
these things will help you grow to understand how to make
testing work for your app.
Key points
To keep your tests repeatable and predictable, you shouldn't
make real network calls in your tests.
You can use MockWebServer to script request responses and

verify that the correct endpoint was called.
How you create and maintain test data will change depending
on the needs for your app.
You can mock the network layer with Mockito if you don't
need the fine-grained control of MockWebServer.
By using the Faker library you can easily create random,

interesting test data.
Deciding which tools are right for the job takes time to learn
through experiment, trial, and error.

You've come so far! From unit tests to integration, testing so
many parts of your apps. You still have a very important part to
learn how to test: the User Interface! All the things you've been
learning are leading up to this point. In Chapter 11, "User
Interface," you'll learn how to automate those clicks on your
screen and to verify what the user sees.
Chapter 11: User Interface
You’ve made it to the third and final part of the testing pyramid: User
Interface (UI) tests, also known as end-to-end tests.
Almost all Android apps have a UI, and subsequently, an essential layer
for testing. UI testing generally verifies two things:
1. That the user sees what you expect them to see.
2. That the correct events happen when the user interacts with the
screen.
With UI tests, you can automate some of the testing you might
otherwise need to do with tedious, manual click-testing. A step up from
integration tests, these test your app most holistically.
Because UI tests highly rely on the Android framework, you need to

install the APK and test instrumentation runner onto a device or
emulator before you can run them. Once installed, you can run the tests
that use the screen to display and perform actions that verify the
behavior. Because of the work involved, UI tests are the slowest and
most expensive to run, which means you’re less likely to run them,
losing the benefit of quick feedback.
Note: With AndroidX Test, it’s possible to run these tests without a
device or emulator and instead run them with Robolectric. This
chapter will not elaborate on the specifics as the technique is the
same as described in Chapter 8, “Integration.”
Following the TDD process requires running your tests frequently while
building, so you won’t want to lean too heavily on UI tests. The length
of time it takes to run them will increase the time it takes to write them.
Test the things you need to test with UI tests, and push what you can
into integration or unit tests. A good rule of thumb is the 10/20/70 split
mentioned in Chapter 4, “The Testing Pyramid,” which explains that
10% of your tests should be UI tests. The idea is that you test for the
main flows, putting whatever logic you can into classes that you can
verify using a faster test.
Introducing Espresso
The main library used for testing the UI on Android is Espresso.
Manually click-testing all parts of your app is slow and tedious. With
Espresso, you can launch a screen, perform view interactions and verify
what is or is not in view. Because this is common practice, Android
Studio automatically includes the library for you when generating a new
project.
Note: Google’s motivation behind this library is for you “to write
concise, beautiful and reliable Android UI tests.”
Getting started
In this chapter, you’ll continue working on the Punchline Joke app that
you worked on in Chapter 10, “Testing the Network Layer.” This is an
app that shows you a new, random joke each time you press a button.
Open the project where you left off in Android Studio, or find the starter
project in the materials for this chapter and open that.
Build and run the app. There’s not much to see yet because it’s your job
to add the UI in this chapter.
Getting familiar with the project

In this chapter, you’ll write tests and implementation for MainActivity.
Find the following files, so you’re all set to go:
activity_main.xml: This is the layout file. At the moment, it’s

sparse, but it won’t be when you’re done with it.
MainActivity.kt: This file is where you set up the view. Notice in

onCreate() that it’s subscribing to LiveData, and then handling the
results in render(). It’s using UiModel to hold the data that you need
to display.
Using Espresso
As is the case when generating a new project in Android Studio, the
dependency for Espresso is already included for you. Open app ‣
build.gradle, and you’ll see the following testing dependency alongside
the other testing dependencies:
androidTestImplementation 'androidx.test.espresso:espresso-core:3.2.0'
This is the primary library that you’ll be using in this chapter. You’ll be
using this alongside some of the libraries and techniques you used from
previous chapters.
You’re all set to go now, so it’s time to dive in!
What makes up Espresso?

There are three main classes you need to know when working with
Espresso: ViewMatchers, ViewActions and ViewAssertions:
ViewMatchers:Contain methods that Espresso uses to find the view

on your screen with which it needs to interact.
ViewActions: Contain methods that tell Espresso how to automate

your UI. For example, it contains methods like click() that you can
use to tell Espresso to click on a button.
ViewAssertions: Contain methods used to check if a view matches a

specific set of conditions.

To get started, inside app ‣ src ‣ androidTest ‣ java ‣ com ‣ raywenderlich
‣ android ‣ punchline ‣ app, create a file named MainActivityTest.kt.
Add to it, an empty test class using
androidx.test.ext.junit.runners.AndroidJUnit4 and
org.koin.test.KoinTest for imports:
class MainActivityTest: KoinTest {
}
You should be familiar with AndroidJUnit4 from previous chapters. As a

reminder, it is only required when using a mix of JUnit3 and JUnit4.
You‘re extending from KoinTest because this project uses the Koin
dependency injection framework. You don’t need to know much about
Koin to see the power of using dependency injection to help with your
UI tests.
Using dependency injection to set mocks

In previous chapters, you used Mockito mocks to stub out some
functionality. For example, when you didn’t want to hit the network
layer. In many of these cases, you can introduce these mocked classes by
passing them through the constructor. But how would you do that for an
Activity? You don’t have the same luxury, because the Android
framework instantiates the class for you. This is why dependency
injection is helpful when it comes to testing.
In this test, you’ll mock the repository so that you’re not hitting the
network layer. This helps with speed and stability.
In KoinModules.kt, a Retrofit service is defined which you need to

override. Do this by adding a @Before block in MainActivityTest.kt with
the following call:
@Before
fun setUp() {
declareMock<Repository>()
}
By using declareMock(), you’re overriding the provided dependency

injection Repository with a Mockito mock.
You’ll need a reference to this repository so that you can stub methods
onto it later. Luckily, Koin will deliver it to you — all you need to do is
ask. Add this property to your class, importing org.koin.test.inject:
private val mockRepository: Repository by inject()
By delegating the property instantiation to inject(), this sets

mockRepository to the mock that Koin passes to the ViewModel used in the
Activity.
One last thing to set up before writing tests. You’ll use the Faker library
you learned in Chapter 10, “Testing the Network Layer” to generate
random test data. Add the property code to prepare this:
private var faker = Faker()
Great! Now you’re all set up for writing tests.
Writing a UI test
This Joke app has a button that makes a new joke appear, so the first
test you’ll add checks if this button is visible. Following the usual
pattern, this test will have setup, actions and verifications.
Start with the setup. Create a new test function with the following stub:
@Test
fun onLaunchButtonIsDisplayed() {
whenever(mockRepository.getJoke())
.thenReturn(Single.just(Joke(
faker.lorem().sentence())))
}
Here, you’re stubbing out the repository so that you don’t hit the
network layer. It’s building on the skills you learned in the previous
chapters using Mockito to stub a function that returns an RxJava Single.
You’re also using Faker to generate random test data for you. Have fun
with the different values you can generate. Just make sure the type is
correct for creating your Joke.
Next, you need to open the activity and perform your verification. Add
these lines to the bottom of your new test. Use the suggested
androidx.test.espresso.* imports, and know that buttonNewJoke will be
unresolved in the beginning:
// 1
ActivityScenario.launch(MainActivity::class.java)
// 2
onView(withId(R.id.buttonNewJoke))
There are a few new things here, so here’s how it works:

1. You use ActivityScenario to launch MainActivity. This comes from
the AndroidX Test library imported as
androidx.test.ext:junit:1.1.1 in the app ‣ build.gradle file. Here,
you’re only using it for one thing, but you can use ActivityScenario
for several things, including driving the Activity’s lifecycle state.
2. This is where you use the Espresso library. You’re passing the
ViewMatcher, withId(), to onView to find the view with the ID
buttonNewJoke. You haven’t created it yet, so there’s an error here.
Then, you’re using the ViewAssertion matches() to assert that this
view is also matched by the ViewMatcher isDisplayed().
You can almost run this test, but you need to write enough code to make
it compile. Add the button to activity_main.xml, inside the
ConstraintLayout tag:
<Button
android:id="@+id/buttonNewJoke"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:visibility="gone"
/>
Android Studio might complain that you haven’t added constraints yet,
but now’s not the time to make it pretty. You’re adding just enough code
to make it compile and run. Notice that you did include
android:visibility="gone". This is so you can see the test fail first, which
lets you know that your test is working.
Before you run your test, turn off animations on your testing device.
Whether it’s an emulator or physical device you’re working with, go to
Settings ‣ Developer options and set all of the following to off:
Window animation scale.
Transition animation scale.
Animator duration scale.
While you’re there, make sure you have “Don’t keep activities” disabled;
otherwise, the ActivityScenario cannot run.
Build and run the test. Android Studio might prompt you to pick a
device on which to run the test, so take your pick. After you run it, you’ll
see a long error that includes something like this:
Expected: is displayed on the screen to the user

Got: "AppCompatButton{id=2131165226, res-name=buttonNewJoke...
This indicates the test found a button with the ID buttonNewJoke, but it’s
not being displayed. It’s an easy fix to make it pass. Remove the
visibility attribute from the XML:
android:visibility="gone"
Run the test again, and it passes. You can now move on to making sure
the joke shows up.
Note: If you’re snooping through all of the Android Studio options,

you may notice there’s a Record Espresso Test option.
Yes, you can use this to automatically create Espresso tests by
clicking through your app and entering details into a wizard.
However, the result is brittle, hard-to-read tests. While it can be
useful for setting up the starting boilerplate or learning how to
match something you’re unsure of, it’s best to avoid using it.
Testing for text

When the app is first launched, you expect to see the first joke. In this
test, you’ll make sure there’s a view that displays that joke right away.
Add this test to your test class:

@Test
fun onLaunchJokeIsDisplayed() {
// 1
val joke = Joke(
// 2
onView(withId(R.id.textJoke))
.check(matches(withText(joke.joke)))
}
This test is similar to the first test with some small but significant
differences:
1. You’re keeping a reference to the Joke you’re creating to stub the

repository. You want to know what the joke was later, to make sure
it’s on the screen.
2. This verification uses many of the same elements you saw before,
but this time you’re using withText() instead of isDisplayed() to
match the text of the joke. withText() accepts both a String literal
and a String reference ID.
Deciding which Matcher to use
Did you notice how many autocomplete options appeared after you
typed “with” of withText() or withId()? With so many options, how
do you know which to choose?
First, you must watch for accuracy and brittleness. Then, make sure
what you’re matching can only match the one view you want to
test.
This could tempt you to be very specific, but you also want to make
sure your tests aren’t breaking with any little change to the UI. For
example, what if you’re matching with the String literal "Okay" and
it’s later changed to "OK" and then changed again to "Got it"? You’d
have to update the test with each change.
This is why matching using IDs is common. Once you have an ID
set, it’s likely the only view on screen with that ID and unlikely to
frequently change — unless it’s a collection.
Here’s Google’s handy Espresso Cheat Sheet for possible Matchers,
Actions and Assertions:
https://developer.android.com/training/testing/espresso/cheat-
sheet
You’re almost ready to run this test. There’s no view with the ID
textJoke yet, so add that to the XML immediately below your button:
<TextView
android:id="@+id/textJoke"
/>
Once again, you have the bare minimum to make it compile and run.
Everything will be scrunched into the corner, but don’t worry, you’ll fix
that soon.
Run the test, and you’ll see an error similar to before:
Expected: with text: is "Dolores quia consequatur quos."

Got: "AppCompatTextView{id=2131165357, res-name=textJoke, text=,
It found the right view, but there’s no text. Because it’s failing, you
know you’re following the right TDD steps.
Open MainActivity.kt and find showJoke(). render() already calls this for
you when a Joke is loaded, so you don’t need to worry about the logic,
only the UI (which makes sense with this being a chapter about UI
testing).
With everything set up, all you need to do is connect the data to the
view. Add this line to showJoke() using the suggested synthetic import:
textJoke.text = joke.joke
Run your test to make sure it passes.
Refactoring
Run the app to see how it’s looking so far. It may not be pretty, but it
sure is testable!
Now that you have some UI tests in place, you can take a break from test
writing and do some refactoring in activity_main.xml.
Add these attributes to the TextView:
style="@style/TextAppearance.AppCompat.Title"
android:gravity="center_horizontal"
android:padding="16dp"
app:layout_constraintBottom_toTopOf="@+id/buttonNewJoke"
app:layout_constraintEnd_toEndOf="parent"
app:layout_constraintStart_toStartOf="parent"
app:layout_constraintTop_toTopOf="parent"
app:layout_constraintVertical_chainStyle="packed"
Add these to the Button:
android:text="@string/new_joke"
app:layout_constraintBottom_toBottomOf="parent"
app:layout_constraintTop_toBottomOf="@id/textJoke"
This adds a little style, along with some descriptive text and constraints.
Run it again to see the changes.
Run your tests to make sure everything still works.

Note: Feel free to modify the view to make it look the way you
want. Just make sure you don’t break your tests while you do it.
Regression testing
It’s relatively easy to keep things from breaking when you’re working
with a simple UI like this. But these tests are extremely helpful when
you’re working with a complicated UI with nested, reused views.
Because you want to limit your UI tests, you may fall into a pattern of
introducing regression tests.
Regression tests are tests that you add when something goes wrong. You
can start by adding a couple tests for the happy path in the UI layer. If
you find something that's broken, write a test for it to make sure it
doesn't break, or regress, again. You can use this at any layer of the
testing pyramid, and it might look like this when paired with TDD:
1. Something broke, a bug is reported, or there’s a crash.
2. You write a test for the expected behavior that broke. It’s likely that
there’s not already a test for this; otherwise, it would have caught
the issue ahead of time.
3. Fix the issue. Do what you need to do to fix the bug, make the test
pass, and keep all other tests green.
4. You now have a regression test to make sure the same issue doesn’t
come up again. No zombie bugs coming back to life here.
These types of tests are especially helpful when working with legacy
systems that aren’t well tested. It’s a way to start introducing valuable
tests alongside your fixes. While you’re in there writing the regression
test, you might take a few minutes to add other tests for vulnerable
nearby areas.
Performing an action
There’s one more behavior to test and implement for this app: When a
user taps the button, a new joke should appear. This is the final and
most complex test you’ll write for this chapter.
Add this last test to your test class:
@Test
fun onButtonClickNewJokeIsDisplayed() {
// 1
.thenReturn(Single.just(Joke(
faker.lorem().sentence())))
// 2
val joke = Joke(
// 3
onView(withId(R.id.buttonNewJoke))
.perform(click())
// 4
onView(withId(R.id.textJoke))
.check(matches(withText(joke.joke)))
}
This is a longer one, so take it step-by-step:
1. Stub the repository response and launch MainActivity as you’ve

done before.
2. Because you want a new joke to show when the button is pressed,
stub the repository with a new Joke.
3. This is where you use a ViewAction for the first time. You locate the
view using a ViewMatcher, then pass in the ViewAction click() to
perform() to perform that action.
4. Finally, verify the new joke is shown in the TextView.
Phew! That was a lot. Everything is already compiling, so go ahead and

run your test. You’ll see a familiar error:
Expected: with text: is "Error ut sed doloremque qui."
Got: "AppCompatTextView{id=2131165357, res-name=textJoke, ...
text=Laudantium et quod dolor.,
It looks like the new joke never showed up! That’s because nothing is
happening when you click the button. Don’t worry, you can fix that.
In MainActivity, add this click listener to the bottom of onCreate():
buttonNewJoke.setOnClickListener {
viewModel.getJoke()
}
Again, all of the logic is already there for you to fetch the new joke. Call
showJoke() when it’s finished. Run your test to see it pass.
You made it! You finished Punchline with fully functioning UI tests. Run
your app and play around with it.
You can now refactor the UI and make it as visually appealing as you’d
like. Just remember to run the tests periodically. Oh yeah, and do your
best to remember the jokes — you may need them for your next party! :]
Using sharedTest (optional)

In Chapter 8, “Integration,” you learned that you could run Android
tests on either a device or locally using Robolectric. For this to work,
your test must be in the correct test/ or androidTest/ directory. With a
small configuration change and a new sharedTest/ directory, you’ll be
able to run your tests both ways without needing to move the file.
Note: If you drag and drop a test from test/ or androidTest/ into
sharedTest/, Android Studio might have some problems running it
because of caching issues.
The first step to setting up shared tests is modifying app ‣ build.gradle

so that it pulls the shared tests into the test/ and androidTest/ source
sets. Add this to the android block of app ‣ build.gradle:
sourceSets {
String sharedTestDir = 'src/sharedTest/java'
test {
java.srcDir sharedTestDir
}
androidTest {
}
}
You then need to make sure that any libraries you use in your tests are
in your dependencies list with both androidTestImplementation and
testImplementation. To save you some work, this is done for you. You’ll
see the duplicates if you open app ‣ build.gradle. Just remember to do a
Gradle Sync while your there.
The only thing left is learning how to run your shared tests. By default,
you don’t have the local vs. device control you want. You can only run
the shared tests as Android tests. There are two ways you can get this
control:
Running the tests from the command line.
Creating a run configuration.
Add this small test to sharedTest/ in a newly created file named

JokeTest.kt. This way you’ll have something to run:
class JokeTest {
private val faker = Faker()
@Test
fun jokeReturnsJoke() {
val title = faker.book().title()
val joke = Joke(faker.code().isbn10(), title)
assert(title == joke.joke)
}
}
To be honest, that was a bit of a joke! :]

Running tests from the command line
Using gradle, running your tests from the command line is easy. Open a
terminal and navigate to the root directory of your project, or use the
terminal view in Android Studio.
Run the following command:
./gradlew test
This runs all of the tests you have in test/ and sharedTest/. You can use
this even if you don’t have shared tests as it will run the tests you have
in test/ only.
Now, try running this one:
./gradlew connectedAndroidTest
Likewise, this one will run the tests you have in androidTest/ and
sharedTest/. This also works if you don’t have any shared tests.
Creating a run configuration

Android Studio also supports creating run configurations, which are
presets you create and run that inform Android Studio of how you want
things to run.
To start, select Edit Configurations... from the run drop-down.
Then, select "+" to create a new run configuration, selecting Android

JUnit for the type.
You now have a few things to fill out in the editor:
Name: Pick a name that makes sense to you. Something like

Robolectric Shared Tests works well.
Test kind: Select "All in directory" to run all of your tests. You can
change this to be more specific if you want even more control over
which tests run.
Directory: Select the path to src/ if you want to run both test/ and
sharedTest/, or sharedTest/ if you only want to run the ones in that
directory.
Use classpath of module: Select app here.
Click OK.
This option is now available in the drop-down if you want to run all of
your shared tests with Robolectric from Android Studio.
Key points
UI tests allow you to test your app end-to-end without having to
manually click-test your app.
Using the Espresso library, you're able to write UI tests.
You can run Android tests on a device and locally using

Roboelectric.

Now that you know the basics of UI testing with Espresso, you can
explore and use everything else the library has to offer.
You can find out about Espresso’s helper libraries and more by reading
their documentation:
https://developer.android.com/training/testing/espresso/
If you want more practice, you can look at Espresso Testing and Screen
Robots: Getting Started: https://www.raywenderlich.com/949489-
espresso-testing-and-screen-robots-getting-started
To take your UI tests to the next level, you can learn how to use Kakao
for even more elegant UI tests by reading UI Testing with Kakao Tutorial
for Android: Getting Started: https://www.raywenderlich.com/1505688-
ui-testing-with-kakao-tutorial-for-android-getting-started
Finally, this chapter uses Espresso and ActivityScenario, which are both
a part of AndroidX Test. To learn more about this suite of testing
libraries, you can watch Getting Started with AndroidX Test:
https://vimeo.com/334519652
This is the end of this section, and you’ve learned a lot about how to test
new apps and features. But what if you’re working on an existing app?
In the next section, you’ll learn some techniques for working with
legacy code.
Section III: TDD on
Legacy Projects
Now that you have an understanding of TDD
and have different tools at your disposal, you’ll
learn how to apply these techniques to projects
that weren't created using TDD and do not have
sufficient test coverage. You’ll work through
Furry Coding Companion Finder on your way to
becoming a TDD guru.

chapters:
Chapter 12: Common Legacy App Problems:

Automated tests can help catch bugs in all
applications, including legacy applications.
Learn how to work around technical debt and
apply testing techniques.
Chapter 13: High-Level Testing with Espresso:

Get started with the Coding Companion Finder
app as you work through a legacy application
and write UI tests using Espresso.
Chapter 14: Hands-On Focused Refactoring:
Refactoring a legacy application doesn't have to
be scary once you have some tests for it. Learn
how to take small steps and move slowly as you
keep your test suite green as you make changes
to your app.
Chapter 15: Refactoring Your Tests: Your tests

are code too. In this chapter, you'll learn how to
refactor both your app code and test code to
make your tests reliable and maintainable.
Chapter 16: Strategies for Handling Test Data:

Learn multiple ways to handle test data in your
Android tests by learning by tutorials. In this
chapter, you'll explore a variety of ways of using
test data in your tests.
Chapter 17: Continuous Integration & Other

Related Tools: Explore what it means to have
continuous integration for your tests and the
tools you can use to achieve it. You'll learn the
pros and cons of different CI strategies in this
chapter.
Chapter 18: Testing Around Other Components:
Set boundaries in your test and use strategies to
help to interact with other libraries and parts of
the Android framework.
Chapter 19: Other Related Techniques: Learn

about other techniques that can complement
your TDD process to make your team more
collaborative and therefore effective.
Chapter 12: Common Legacy App
Problems
In an ideal world, your team will write a new Android app that will use
TDD development techniques from the beginning. In this world, your
app will have great test coverage, an architecture that is set up for
TDD, and code with a high level of quality. Because of that, the team
will be very confident in its ability to fearlessly refactor the code, and
they will ship code regularly.
In the real world, many, if not most, apps have technical debt that you
will need to work around. In this chapter, you will learn about some of
the more common issues you will run into with legacy applications.
Then, in subsequent chapters, you will learn how to address these
when working on legacy projects.
A brief history of TDD in Android

While there are many technical aspects that affect TDD on a platform,
the culture surrounding the development community has a big impact
on its adoption.
Android was first released on the G1 in 2008. In those days, the

primary development language was Java and, as a result, many early
developers came from other Java domains, such as server-side
development. At that point, TDD as we know it today had only been
around for nine years and was just beginning to see adoption in some
software development communities and organizations. During that
time, the Rails framework, which was four years old, arguably had the
largest percentage of projects being developed using TDD. This was
due, in part, because many signatories of the Agile Manifesto were
evangelizing it.
Java was 12 years old at that time and pre-dated TDD. It had become a
mature technology that large conservative enterprises were using to
run mission-critical software. As a result, most Java developers, with
the exception of those who were working at cutting edge Agile
development shops, were not practicing TDD.
This enterprise Java development culture affected the early days of

Android development. Unlike Rails, the initial versions of Android
supported testing as more of an afterthought, or not at all. As a result,
most new Android developers had not come from an environment
where testing was important or even known; it was not a primary
concern of the framework and most developers focused on learning
how to write apps, not on testing. During those early days, many of the
concepts we take for granted today were just beginning to be
developed. Over time the platform evolved, devices became more
powerful, and apps became more complex.
Eventually, tools like Robolectric and Espresso were introduced and

refined. TDD became more of an accepted practice among Android
developers. But even today, it is not uncommon to be at an Android
developer meetup where fewer than half of the developers in the
audience are actively writing tests or practicing TDD on a daily basis.
Lean/XP technical practice co-dependency

TDD is one of the key development practices of Lean/XP.
Before Lean/XP became popular, most development organizations used

a Waterfall development approach with steps that often provided
marginal value for projects.
Some traits of a Waterfall project include:
1. All known requirements for the entire project (i.e., a project that
will take months of development effort) are gathered and written
in a requirements document before coding begins.
2. Certain documents, such as software architecture design

documents, are often created after the requirements have been
created, but before development begins.
3. Test plans need to be created before testing begins.
While all of this sounds logical in theory, in reality:
1. At some point, usually mid-project, new requirements are

discovered that require a change.
2. If the project is really strict with its process (usually the

exception), it must modify all artifacts upstream and downstream.
This leads to a project timeline that is extended.
3. If the business cannot afford to have the timeline slip because of

the process (again, what usually happens), phases are skipped, the
documents end up not reflecting the reality of the project, and the
project gradually becomes chaotic.
Lean/XP fun historical facts

In software, the techniques we use were built on the shoulders of the
giants, often from other industries. TDD/XP has roots in
manufacturing through a subset of Six Sigma techniques, which are
called Lean. To learn more about Lean, XP and its relationships, this
Wikipedia article on Lean software development is a great place to
start: https://en.wikipedia.org/wiki/Lean_software_development.
Lean, as the word implies, eliminated many of the unnecessary aspects
of the development process and only kept the things that were deemed
necessary. The practices that are kept are often highly interdependent.
Because of this, if one practice is not followed, another one becomes
more difficult or may not be done as well. When an app is developed
with TDD principles at the unit and integration layer, the practice
guides the project towards good design.
Let's explore some of the co-dependent legacy code issues.
No unit or integration tests

This is the biggest issue you will likely run into when working on a
legacy project. It happens for a variety of reasons. The project may be
several years old and the development team may have chosen not to
write unit tests. For example, it is not uncommon to find teams with
technically strong Android developers who do not know how to write
tests. Of course, if you are on a team that does not know how to
practice TDD, this book would make a great gift, especially for
holidays, birthdays or "just because." :]
Difficult to test architecture

One of the reasons why MVVM, MVP and MVI are popular is because
they tend to drive a structure that is easier to test. But, if the app was
initially developed without any unit tests, it is likely there is no
coherent architecture. While you may get lucky with a design that is
testable, it is more common to find an untested app with an
architecture that is, in fact, difficult to test.
Components that are highly coupled

When an app's components are highly coupled, they are highly
interdependent.
For example, let's say that you want to be able to search for any pets
that share traits with a specific pet — in this case you will use the name
of the pet. One implementation might look like this:
class Cat(
val queenName: String,
val felineFood: String,
val scratchesFurniture: Boolean,
val isLitterTrained: Boolean)
class Dog(
val bestFriendName: String,
val food: String,
val isHouseTrained: Boolean,
val barks: Boolean)
fun findPetsWithSameName(petToFind: Any): List<Any> {

lateinit var petName: String
if (petToFind is Cat){
petName = petToFind.queenName
} else if (petToFind is Dog) {
petName = petToFind.food
}
return yourDatabaseOrWebservice.findByName(petName)
}
The functionality that should be truly unique to the

findPetsWithSameName method is the call to the
yourDatabaseOrWebservice.findByName() call. But, this method also has
code to get data from these objects before doing a find on them. To test
this, you could write a test that only creates an instance of a Cat and
passes it into the method. For example, let's say that you have a cat
named Garfield and your test data has two other animals with the same
name. Your test would look something like this:
@Test
fun `find pets by cats name`() {
val catNamedGarfield = Cat("Garfield", "Lasagne", false, false)
assertEquals(2, findPetsWithSameName(catNamedGarfield).size)
}
This test will pass. The problem is that there is a bug in your
implementation, because the code for the dog is retrieving the search
name from the wrong field. To get full code coverage, and find this
issue, you need to write an additional test that also passes in a Dog.
There is also a boundary condition that has not been addressed if
someone passes in different class, like a Lion:
@Test
fun `find pets by dogs name`() {
val dogNamedStay = Dog("Stay", "Blue Buffalo", false, false)
assertEquals(5, findPetsWithSameName(dogNamedStay).size)
}
@Test
fun `find pets by lions name`() {
val lionNamedButterCup = Lion("Buttercup", "Steak", false, false)
assertEquals(2, findPetsWithSameName(lionNamedButterCup).size)
}
In total, you needed three unit tests for this method. A less coupled
implementation might look like this:
open class Pet(var name: String, var food: String)
class Cat(
name: String,
food: String,
var scratchesFurniture: Boolean,
var isLitterTrained: Boolean): Pet(name, food)
class Dog(
name: String,
food: String,
var isHouseTrained: Boolean,
var barks: Boolean): Pet(name, food)
fun findPetsWithSameName(petToFind: Pet): List<Pet> {

return yourDatabaseOrWebservice.findByName(petToFind.name)
}
With the new version of this code, you only need to write one test to
test the functionality of findPetsWithSameName(petToFind: Pet):
@Test
fun `find pets by cats name`() {
val catNamedGarfield = Cat("Garfield", "Lazagne", false, false)
assertEquals(2, findPetsWithSameName(catNamedGarfield).size)
}
You could refine this test further to mock a pet, taking away any
dependency on implementations of Cat. Highly coupled code can also
lead to situations where one component changes or you do a small
refactoring in one area that leads to large changes throughout the app
(or even the tests). While an app's architecture doesn't guarantee that
this won't happen, the less consistent the architecture, the more likely
you are to see this.
Components with low cohesion

One important tenant for Object-Oriented Design is to have
components that focus on doing one thing well. This is also referred to
as cohesion. For example, let's say you have a class called Car:
class Car {
val starter = Starter()
val ignition = Ignition()
val throttle = Throttle()
val engineTemperature = Temperature()
var engineRPM = 0
var oilPressure = 0
var steeringAngle = 0L
var leftDoorStatus = "closed"
var rightDoorStatus = "closed"
fun startCar() {
ignition.position = "on"
starter.crank()
engineRPM = 1000
oilPressure = 10
engineTemperature = 60
}
fun startDriving() {
if(leftDoorStatus.equals("closed") &&
rightDoorStatus.equals("closed")) {
steeringAngle = 0L
setThrottle(5)
}
}
private fun setThrottle(newPosition: Int) {

if (ignition.position.equals("on") && engineRPM > 0 &&
oilPressure > 0) {
throttle.position = newPosition
engineRPM = newPosition * 1000
oilPressure = newPosition * 10
}
}
For this implementation, you have two publicly callable methods,

startCar() and startDriving():
startCar()method: Turns on the ignition, cranks the starter and

updates the status of the engine RPM, oil pressure and engine
temperature.
startDriving() method: Checks that the doors are closed, sets the
steering wheel to an angle of 0 and calls a private method called
setThrottle() that begins to move the car.
setThrottle() has to check that the ignition position is on and that the
engineRPM and oilPressure are above 0. If they are, it sets the throttle
position to the value passed in, and sets the engine RPM and oil
pressure to a multiplier of the throttle position.
With a car, you could have multiple engine choices. For example, your
current car runs on fuel, but what if you wanted to switch the current
care out for an electric one? Things such as the engineRPM and
oilPressure would not be needed — these are really details of the
engine. As a result of this, your class currently has low cohesion.
Since this is an incomplete car, before it'll be usable, you will need to
add things such as brakes and tires, which will make Car a very big (and
complex) class.
Now, take a look at the same example with high cohesion:

class Engine {
val starter = Starter()
val ignition = Ignition()
val throttle = Throttle()
val engineTemperature = Temperature()
var engineRPM = 0
var oilPressure = 0
fun startEngine() {
ignition.position = "on"
starter.crank()
engineRPM = 1000
oilPressure = 10
}
fun isEngineRunning(): Boolean {

return ignition.position.equals("on") && engineRPM > 0 &&
oilPressure > 0
}
fun setThrottle(newPosition: Int) {

if (isEngineRunning()) {
throttle.position = newPosition
}
}
}
class Car {
val engine = Engine()
var steeringAngle = 0L
var leftDoorStatus = "closed"
var rightDoorStatus = "closed"
fun startCar() {
engine.startEngine()
}
fun startDriving() {
if (leftDoorStatus.equals("closed") &&
rightDoorStatus.equals("closed")) {
steeringAngle = 0L
engine.setThrottle(5)
}
}
}
Here, you have the same functionality, but classes have more of a
single purpose.
If you've been in enough legacy code-bases, you will run across
components like the first example in which you have large classes that
are doing a lot of different things. The more lines of code that a class
has, the more likely it is to have low cohesion.
Other legacy issues

A large codebase with many moving parts
Many legacy systems, over time, become large apps that do a lot of
things. They may have hundreds of different classes, several
dependencies, and may have had different developers — with different
development philosophies — working on the project. Unless you were
the original developer on the project, there will be sections of the app
code that you don't fully understand. If you are new to the project, you
may be trying to figure out how everything works.
Your sample projects for this book will not be that large, but you will be
using the same approach that you will want to use for these projects —
namely, focusing on one section of the app and working through the
others over time.
Complex/large dependent data

In some domains, you may have an app that creates and consumes a
large amount of data. Those projects may have a large number of
models with several unique data fields. Taming this beast as you test
can easily look like a insurmountable task, so stay tuned for tricks on
how to address this.
Libraries/components that are difficult to test

A lot of libraries and components have very important functionality
that is easy to use, but did not consider automated unit tests as part of
the design. One example is using Google Maps in a project with custom
map markers. If you had to create this functionality, you would have to
write a lot of code. But, integration tests can be very challenging. In
some instances, the best solution may be not to test these components
because the value added by the tests are lower than the effort to create
tests.
Google Location Services Components are another example of this.

Stay tuned for an example where we look at ways to work around these
kinds of libraries.
Old libraries
This happens a lot: A developer needs to add functionality to an app.
Instead of reinventing the wheel, they use a library from the Internet.
Time passes and the library version included with the app is not
updated. If you are lucky, the library is still being supported and there
are no breaking changes when you update it.
The library has a new version with breaking changes

If a library version in a project has not been updated in a few years, and
it is being actively maintained, there is a good chance that a new
version with breaking changes has been introduced.
Issues that may introduce include:
1. Features you currently use may have been removed.
2. You may have a significant number of touch-points in the app that

require a significant amount of refactoring.
3. Core functionality may have changed.
Even if you used TDD with the initial version of the library, there is not
much you can practically do to prevent this, outside of timing when
you do your upgrade.
The library is no longer being maintained

This happens for a variety of reasons, including:
1. It was an open-source side project for a developer who ended up

getting busy with other endeavors.
g g y
2. A new library or approach has been developed, which leads to
projects migrating from the old library.
3. If a company created the library, they may have gone out of

business or stopped supporting a product.
If the library is open source, you could decide to take over maintenance
of it. Alternatively you will need to migrate to a new library. If your app
already has a lot of unit tests, this will break them as you add support
for the new library.
Wrangling your project
The app is working well with no tests

After seeing all of the issues you can run into with a legacy project, you
may be asking if you should start to add tests to the project. If you plan
on continuing to maintain a project for a while, the short answer is yes,
but with a few caveats.
1. If your project has more than one developer, TDD will not add as
much value to the project unless the entire development team is
dedicated to practicing it.
2. Unless your project is small, the first passes at TDD will take a
non-trivial amount of effort to set up.
3. Rome wasn't built in a day; neither will your test suite be.
4. You probably will not have the luxury of stopping new feature
development for several months to add test coverage to your entire
project.
You consider rewriting the entire app

This can be a very tempting option when working with a legacy app,
especially if you were not the original author. If your app is small or
truly a mess, that may be the best solution. But, if your project is large,
and you are planning on keeping most of these features, however
tempting a rewrite may be, it could be a job-killing move.
Most legacy apps have a significant number of undocumented features

and edge cases. For these projects, a rewrite will often take several
months. In addition to that, the business will likely want to maintain
the existing app and add new features to it. While a rewrite may still be
the best solution, before heading down that path, a better option
would be to break the app up into components and refactor things a
component at a time. This is called the Strangler pattern.
Lore has it that the Strangler pattern got its name from a vine called
the strangler vine. These vines seed themselves in fig trees. Over time,
they grow into their own plants surrounding and killing the tree.
Likewise, you components will surround the initial implementation,
eventually killing off the original one.
Key points
Lean and Extreme Programming have a lot of interdependencies.
Many legacy applications have little or no automated tests.
Lack of automated tests can make it difficult to get started with

TDD and may require getting started with end-to-end Espresso
tests.
Rewriting a legacy application should generally be considered as a
last resort option.

Beyond the techniques you will be learning in this book, the book
Working Effectively With Legacy Code by Michael Feathers does a great
job of talking about legacy code problems. You can check out this book
at https://www.oreilly.com/library/view/working-effectively-
with/0131177052/.
Chapter 13: High-Level Testing With
Espresso
For many — arguably most — legacy projects, it is easier (and often
provides quicker wins) by starting with UI tests using Espresso.
There are a number of reasons for this, including:
1. Most non-TDD apps have not been decomposed for testability.
2. If you are new to the project and it is fairly large, it can take you a
while to get your head around the architecture of the system as a
whole.
3. UI tests test the large parts of the app working together.
A holistic approach allows you to get functionality of the app under test
before making code changes. This gives you a number of quick wins,
including:
1. The ability to get a section of the app under test before adding
features or refactoring the code.
2. Quickly getting the benefits of automated testing, even if you do not

have time to start to make architectural changes.
3. Providing you with test coverage for when you do start to refactor
your code for testability at a lower level.
Getting started
To explore UI tests, you are going to work on an app called Coding
Companion Finder.
The story
This app was created by a developer who practices pair programming, in
which two developers work side by side on a problem at the same time.
This technique has a number of benefits, including helping to improve
the quality of the software you are working on. Unfortunately, this
person often works from home where it may not always be possible to
have a human partner to pair up with.
Other developers have been in the same situation, and they had
developed a technique called purr programming, in which a human
development pair is substituted with a cat. The developer loved cats,
adopted one, began to regularly purr program, and noticed that the
quality of their software written at home dramatically improved. Beyond
the benefits of pair programming, the developer also gained a loving
companion, and soon realized that it could be used with dogs as well.
Whenever the developer was at meetup groups or work, they'd tell
friends about the benefits of purr programming and a related practice
called canine coding.
One day, when the developer was at the pet store, they noticed that a
local pet shelter was hosting an "Adopt a Pet" day and immediately
thought: "What if there was an app to help match fellow programmers to
these pets?" They decided to partner with the shelter to create the
Coding Companion Finder.
The app has been successful at placing companions into loving homes,
but many pets are still without homes and many developers have yet to
discover this technique. After getting feedback from users, the shelter
has some ideas for the app, but the original developer is too busy, so they
have reached out to you!
Setting up the app

This app uses an API from a website called Petfinder, which requires a
developer key.
To get started, go to the Petfinder registration page at

https://www.petfinder.com/user/register/ and create a new account.
If you are not in the US, Canada or Mexico, choose the United States for
your location and 30354 as your zipcode. Once your account is created,
go here https://www.petfinder.com/user/login/, log in, and then create a
API key by entering a Application Name, Application URL, accept the
Terms of Service and click the GET A KEY button.
Once you request a key, you will be redirected to a page that will show
you an API Key and an API Secret. Copy the API key value.
Now, import the starter project and open up MainActivity.kt. At the top
of this file you will see the following:
val apiKey = "replace with your API key"
val apiSecret = "replace with your API secret"
Replace the string with the key that you've just copied. Run the app.
A tour of the app

The app will briefly present you with a splash screen; then, if you pasted
in the correct key, it will bring up a page showing you a Featured
Companion.
Tap on Find Companion. You will be taken to a search screen where you
can search for companions in the United States, Canada or Mexico. Enter
a location and tap the FIND button to find companions close to that
location. Then, tap on one of them to see more information about a
companion.
Your first assignment
Users have really liked the app, but it is difficult to find the contact
information for a companion in the details screen. As your first task, the
shelter has asked you to add contact information to the companion
details screen.
Understanding the app architecture

Before adding tests and features, you need to understand how the app is
put together. Open up the starter project and open the app level
build.gradle. In addition to the normal Kotlin and Android dependencies,
you have the following:
// Glide
implementation("com.github.bumptech.glide:glide:4.9.0") {
exclude group: "com.android.support"
}
kapt 'com.github.bumptech.glide:compiler:4.9.0'
// carouselview library
implementation "com.synnapps:carouselview:0.1.5"
// retrofit
implementation "com.squareup.okhttp3:logging-interceptor:3.11.0"
implementation 'com.squareup.retrofit2:retrofit:2.5.0'
implementation 'com.squareup.retrofit2:converter-gson:2.5.0'
implementation 'com.jakewharton.retrofit:retrofit2-kotlin-coroutines-adap
Our project depends on Glide, Carouselview and Retrofit. Now, open up

MainActivity.kt. In your onCreate method, you will see the following:
if (petFinderService == null) {
val logger = HttpLoggingInterceptor()
logger.level = HttpLoggingInterceptor.Level.BODY
val client = OkHttpClient.Builder()
.addInterceptor(logger)
.connectTimeout(60L, TimeUnit.SECONDS)
.readTimeout(60L, TimeUnit.SECONDS)
.addInterceptor(AuthorizationInterceptor(this))
.build()
petFinderService = Retrofit.Builder()
.baseUrl("http://api.petfinder.com/v2/")
.addCallAdapterFactory(CoroutineCallAdapterFactory())
.client(client)
.build().create(PetFinderService::class.java)
}
This sets up Retrofit with a hard-coded URL. Looking at onResume, you

will see a few more hints about how things come together:
val navHostController = Navigation.findNavController(this,

R.id.mainPetfinderFragment)
val bottomNavigation =
findViewById<BottomNavigationView>(R.id.bottomNavigation)
NavigationUI.setupWithNavController(bottomNavigation, navHostController
This is using the Jetpack Navigation Library to set up your
BottomNavigationView and hook it up to a fragment element in your
activity_main.xml layout. Open up that file and you will see the
following:
<fragment
android:id="@+id/mainPetfinderFragment"
android:name="androidx.navigation.fragment.NavHostFragment"
android:layout_width="match_parent"
android:layout_height="0dp"
app:defaultNavHost="true"
app:layout_constraintBottom_toTopOf="@id/bottomNavigation"
app:navGraph="@navigation/nav_graph"
/>
<com.google.android.material.bottomnavigation.BottomNavigationView
android:id="@+id/bottomNavigation"
app:layout_constraintVertical_bias="1"
app:menu="@menu/bottom_navigation_menu"/>
Your BottomNavigationView has a menu that is set up in

bottom_navigation_menu.xml and your fragment references a navGraph
with:
app:navGraph="@navigation/nav_graph"
Now, open up bottom_navigation_menu.xml and you will see the

following:
<menu xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:app="http://schemas.android.com/apk/res-auto">
<item
android:id="@id/randomCompanionFragment"
android:enabled="true"
android:icon="@drawable/ic_featured_pet_black_24dp"
android:title="@string/featured_pet"
app:showAsAction="ifRoom" />
<item
android:id="@id/searchForCompanionFragment"
android:enabled="true"
android:icon="@drawable/ic_search_black_24dp"
android:title="@string/find_pet"
app:showAsAction="ifRoom" />
</menu>
This is populating your bottom menu items. Finally open up

nav_graph.xml and you will see the following navigation definition:
<navigation
xmlns:android="http://schemas.android.com/apk/res/android"
xmlns:app="http://schemas.android.com/apk/res-auto"
xmlns:tools="http://schemas.android.com/tools"
android:id="@+id/nav_graph"
app:startDestination="@id/randomCompanionFragment">
<fragment
android:id="@+id/randomCompanionFragment"
android:name="com.raywenderlich.codingcompanionfinder.randomcompanio
android:label="fragment_random_pet"
tools:layout="@layout/fragment_random_companion"/>
<fragment
android:id="@+id/searchForCompanionFragment"
android:name="com.raywenderlich.codingcompanionfinder.searchforcompa
android:label="fragment_search_for_pet"
tools:layout="@layout/fragment_search_for_companion"/>
</navigation>
The call to NavigationUI.setupWithNavController matches the ID of your

menu items with the IDs of your nav_graph and instantiates either your
randomCompanionFragment or searchForCompanionFragment
depending on which menu item you select.
When you have multiple screens in your app, there are three main ways
that you might choose to do it:
1. Multiple activities, one for each screen. When your user navigates to
another screen, a new activity is shown.
2. One activity with multiple fragments. When you user navigates to a

new screen, you show a new fragment.
3. A hybrid of #1 and #2.
Because this is using the Jetpack Navigation Component, it probably is

going to be using a one-activity, multiple-fragment approach. To verify
that, take a look at your project view.
Other than one additional activity that is used for the splash screen, that
assumption appears to be correct. Since you're going to be working on
the search functionality, open up SearchForCompanionFragment.kt and
have a look around.
In your onActivityCreated method, you are using findViewById to get a
reference to your search button. This probably means that the app does
not use data binding or a helper library such as Butterknife to get
references to objects in your view.
When the search button is tapped, a call is made to a local method called
searchForCompanions().
At the top of this method are additional findViewById calls.

This is calling your PetFinderService which is provided by Retrofit.
The results are then passed into a CompanionAdapter that is part of a

RecyclerView.
In summary, your app has the following traits:

1. It does not follow a MVC, MVP, MVVM, or MVI type of pattern.
2. Looking at the app as a whole, it has one main external dependency

on the Petfinder service.
3. Like many legacy apps, it has some coding/architectural issues that

you will need to work around to get it under test.
Determining your system boundaries

When you are adding automated testing to an app using Espresso, you
want to have tests that are repeatable and fast. Using Espresso, you are
performing a form of integration testing, but you still need to have some
system boundaries for your tests. The boundary determines what you are
testing and allows you to control the inputs to your app.
A rule of thumb when writing Espresso tests is to not have tests that
make network requests or access external resources. In the case of your
app, your boundary is the Petfinder service. Pets are constantly being
added and removed from the service, and some calls, such as the one for
a featured pet, provide different data every time you call it. Beyond the
network latency, those changes would make it very difficult to create
meaningful repeatable tests. As you get the app under test, you will be
adding a mock of this to address that.
Preparing your app for testing

To get started, open your app level build.gradle file and add the
following:
androidTestImplementation "androidx.test:rules:1.2.0"
androidTestImplementation "androidx.test.ext:junit:1.1.1"
androidTestImplementation "android.arch.navigation:navigation-testing:1
androidTestImplementation 'com.squareup.okhttp3:mockwebserver:3.12.0'
androidTestImplementation "androidx.test.espresso:espresso-contrib:3.2.
This is adding libraries for rules, junit, navigation-testing,

mockwebserver and espresso-contrib.
Next, rename ExampleInstrumentedTest in the androidTest source set to
be FindCompanionInstrumentedTest.
Now, paste the following into your renamed test class:

lateinit var testScenario: ActivityScenario<MainActivity>
companion object {
private lateinit var startIntent: Intent
// 1
val server = MockWebServer()
// 2
val dispatcher: Dispatcher = object : Dispatcher() {
@Throws(InterruptedException::class)
override fun dispatch(
request: RecordedRequest
): MockResponse {
return CommonTestDataUtil.dispatch(request) ?:
MockResponse().setResponseCode(404)
}
}
@BeforeClass
@JvmStatic
fun setup() {
// 3
server.setDispatcher(dispatcher)
server.start()
// 4
startIntent =
Intent(ApplicationProvider.getApplicationContext(),
MainActivity::class.java)
startIntent.putExtra(MainActivity.PETFINDER_URI,
server.url("").toString())
}
}
This is doing the following:
1. Creating an instance of your MockWebServer.
2. Creating a dispatcher that will intercept requests to the

MockWebServer and respond with a MockResponse that will be sent
out to the server.
3. Passing the dispatcher to your MockWebServer and starting it when

the class is first instantiated.
4. Creates an intent to pass in the URL for the MockWebServer.
Fix all the required imports, selecting okhttp3.mockwebserver.Dispatcher as

the import for the Dispatcher if you get multiple options. There is an
"unresolved reference" error with CommonTestDataUtil. To fix that, create a
new file in your androidTest directory called CommonTestDataUtil.kt,
and paste in the following:
object CommonTestDataUtil {
fun dispatch(request: RecordedRequest): MockResponse? {
when (request.path) {
else -> {
return MockResponse()
.setResponseCode(404)
.setBody("{}")
}
}
}
}
This is the beginning of a helper method that will look at the request
coming in, and respond based on the request parameters. Don't worry
about the warning that the when clause can be simplified for now.
Adding test hooks

MockWebServer spins up a local web server that runs on a random port
on an Android device. In order to use it, your app will need to point your
Retrofit instance at this local server instead of the one at petfinder.com.
Since your app sets up Retrofit in your MainActivity, you are going to add
some logic to allow this to be passed in.
Open open MainActivity.kt and look for the following:
.baseUrl("http://api.petfinder.com/v2/")
.client(client)
}
Replace it with the following:

val baseUrl = intent.getStringExtra(PETFINDER_URI) ?:
"http://api.petfinder.com/v2/"
.baseUrl(baseUrl)
.client(client)
}
This checks the intent for a value stored under PETFINDER_URI, and if one
is present uses it instead of your hard-coded URI.
Now, add the following to the top of the class:
companion object {
val PETFINDER_URI = "petfinder_uri"
val PETFINDER_KEY = "petfinder_key"
}
This creates two constants for your Intent keys. Finally, paste the
following into the top part of your onCreate function:
intent.getStringExtra(PETFINDER_KEY)?.let{
apiKey = it
}
This looks for an API key being passed into your MainActivity via an
Intent, and if there is one, sets your key to that value instead of your
hard-coded one.
Adding legacy tests

When adding tests to a legacy app with no test coverage, the first step is
to add tests around the functionality where you are going to be adding a
feature.
To get started, you are going to add tests around the "search for
companion" section of your app. When you first start the app, you are
taken to the Featured Companion page.
Pressing the Find Companion button takes you to the find page.
For your first test, you are going to open up the app, press the Find
Companion bottom item and verify that you are on the "Find
Companion" page.
To get started, make sure that the app is running on your device and use
the Layout Inspector in your Android Studio Tools menu to get a
snapshot of the screen. Now, highlight that menu item to find the ID of
that menu item.
Your SearchForCompanionFragment is the entry point for your search page. It
has a view called fragment_search_for_companion. Open it up and get the ID
of your Find button:
<com.google.android.material.button.MaterialButton
android:id="@+id/searchButton"
android:text="Find"
app:layout_constraintBottom_toBottomOf="@+id/searchField"
app:layout_constraintStart_toEndOf="@id/searchField"
app:layout_constraintTop_toTopOf="@id/searchField" />
Next, look for the ID of your text input field:
<com.google.android.material.textfield.TextInputEditText
android:id="@+id/searchFieldText"
android:hint="Enter US Location"
android:textColor="@color/primaryTextColor" />
Putting this all together, when you enter the app, you are going to click
on a button with an ID of searchForCompanionFragment. Then, you will
check that items with the ID of searchButton and searchFieldText are
visible, in order to verify that your app does indeed go to the next screen.
Open up your FindCompanionInstrumentedTest, remove the function called

useAppContext, and paste in the following function:
@Test
fun pressing_the_find_bottom_menu_item_takes_the_user_to_the_find_page()
testScenario = ActivityScenario.launch(startIntent)
onView(withId(R.id.searchForCompanionFragment))
.perform(click())
onView(withId(R.id.searchButton))
onView(withId(R.id.searchFieldText))
testScenario.close()
}
Add all the imports, picking

androidx.test.espresso.assertion.ViewAssertions.matches when presented
with choices for the matches() method. Next, run the test by clicking the
green arrow on the left of the code — and you should have green
(passing) test!
Note: You may have noticed that you are not really mocking
anything required for the Featured Companion screen. This is
because you are not testing this screen, and it is not necessary to
populate data for it to test your Find Companion screen.
Now that you have a passing test, you are going to want to have a basic
test to exercise performing a search and tapping on a result.
Understanding your API

Your SearchForCompanionFragment is making a call to your getAnimals
function in your PetFinderService when you tap the Find button, which is
implemented like this using Retrofit:
@GET("animals")
fun getAnimals(
@Header("Authorization") accessToken: String,
@Query("limit") limit: Int = 20,
@Query("location") location: String? = null
) : Deferred<Response<AnimalResult>>
That call passes in a accessToken and location. The accessToken is retrieved

via a OAUTH call to a oauth2/token endpoint passing in the apiKey and
apiSecret you received when you registered for the Petfinder API. In
order to mock out this test data, you are going to need to figure out what
real data might look like! The calls to getAnimals happen to be GET
requests, and the data returned is in a JSON format. But in order to test
out the calls, you will need to make a POST call to get your accessToken
and then pass that in the header of your GET request.
To get an idea of how the data looks, you can use a tool such as Postman,
found here https://www.getpostman.com/, to explore the Petfinder
documents located here https://www.petfinder.com/developers/v2/docs/.
Looking at the output from Postman with the list of pets contracted, you
will see a list with 20 animals.
For your test, you will only need to have one or two pets. Each pet record
will also have several photos.
These photos reference URLs on the web. For the time being, you are not
going to test the photo loading and are going to want to have records
without photos. A relatively straight-forward way to do this is to copy the
fully expanded formatted text into a text file, edit out the data you don't
want, and save it. To save you some time, we have included a file called
search_30318.json in app/src/androidTest/assets.
uses a RecyclerView to display the list of

SearchForCompanionFragment
results. Open the CompanionViewHolder.kt. At the bottom of this file,
you will see:
private fun setupClickEvent(animal: Animal){

view.setOnClickListener {
val viewCompanionFragment = ViewCompanionFragment()
val bundle = Bundle()
bundle.putSerializable(ViewCompanionFragment.ANIMAL, animal)
viewCompanionFragment.arguments = bundle
val transaction =
fragment.childFragmentManager.beginTransaction()
transaction.replace(R.id.viewCompanion,
viewCompanionFragment).addToBackStack("companionView")
.commit()
}
}
When you tap on a companion in that list, a ViewCompanionFragment is

created, and the Animal object for that record is passed into it via Bundle
arguments. Now open your ViewCompanionFragment and you will see that
the only data inputs for this fragment are via the arguments Bundle. So
you do not have to mock out any other calls!
animal = arguments?.getSerializable(ANIMAL) as Animal

Setting up your mock data
Now that you have the data from your API, you are going to need to tell
your test Dispatcher how to retrieve it in order to mock out the API
response. Open CommonTestDataUtil.kt and add in the following:
@Throws(IOException::class)
private fun readFile(jsonFileName: String): String {
val inputStream = this::class.java
.getResourceAsStream("/assets/$jsonFileName")
?: throw NullPointerException(
"Have you added the local resource correctly?, "
+ "Hint: name it as: " + jsonFileName
)
val stringBuilder = StringBuilder()
var inputStreamReader: InputStreamReader? = null
try {
inputStreamReader = InputStreamReader(inputStream)
val bufferedReader = BufferedReader(inputStreamReader)
var character: Int = bufferedReader.read()
while (character != -1) {
stringBuilder.append(character.toChar())
character = bufferedReader.read()
}
} catch (exception: IOException) {
exception.printStackTrace()
} finally {
inputStream.close()
inputStreamReader?.close()
}
return stringBuilder.toString()
}
This is opening up your file, reading it, and returning it as a string. Next,
replace your dispatch function with the following:

return when (request.path) {
"/animals?limit=20&location=30318" -> {
.setBody(readFile("search_30318.json"))
}
else -> {
MockResponse().setResponseCode(404).setBody("{}")
}
}
}
This is adding a when condition for the request that is made when you do
a search for the zipcode 30318.
Writing your next test

For this test, you are going to perform a search, click on a result, and see
details for the companion you tapped on.
Open up FindCompanionInstrumentedTest.kt and add the following:

@Test
fun searching_for_a_companion_and_tapping_on_it_takes_the_user_to_the_c
// 1
.perform(click())
// 2
.perform(typeText("30318"))
.perform(click())
// 3
// 4
onView(withText("KEVIN")).perform(click())
// 5
onView(withText("Rome, GA")).check(matches(isDisplayed()))
}
Here is what this is doing:
1. Clicks the bottom menu item to find a companion.
2. Enters 30318 in searchTextField and clicks the "Find" button.
3. It makes sure that you are still on the Find Companion screen.
4. Clicks on the record for a cat named KEVIN.
5. Verifies that you are on the new page by looking for a data item that
was not in the list of items. In this case, the city of Rome, GA.
Now run the test.

Unfortunately, something is not right. It is not finding results for your
search.
IdlingResources
When using Espresso, your test suite is running in a different thread
from your app. While there are some things related to the lifecycle of
your activity that Espresso will be aware of, there are other things that it
is not. In your case, in order to make your tests and test data readable,
you are putting your data in a separate JSON file that needs to be read in.
While this is not a big hit on the performance of your tests, a file read is
slower than the execution time of your Espresso statements.
Because of this, Espresso is evaluating your statement to press the

button for Rocket the cat before your app has finished reading in the data
and populating your RecyclerView. One clumsy-but-effective way to work
around this is to put a Thread.sleep(5000) before the command for the
button press.
There are, however, a number of problems with that idea, including:
Your tests may not need the amount of time you specified on a
device, and will thus run slower than needed.
Your tests may need more time than you specified on other devices
which makes them unreliable.
This is where IdlingResource comes in. The idea behind using

IdlingResource is to create a mechanism that allows you to send a
message to Espresso telling it that the app is busy doing something and
another message to tell it when it is done. That way your test is only
waiting for a longer running command when it actually should.
To get started, create a new Kotlin file in your test directory called
SimpleIdlingResource.kt and enter the following:
class SimpleIdlingResource : IdlingResource {
// 1
@Nullable
@Volatile
private var callback: IdlingResource.ResourceCallback? = null
// 2
// Idleness is controlled with this boolean.
var activeResources = AtomicInteger(0)
override fun getName(): String {

return this.javaClass.name
}
// 3
override fun isIdleNow(): Boolean {
return activeResources.toInt() < 1
}
override fun registerIdleTransitionCallback(

callback: IdlingResource.ResourceCallback
) {
this.callback = callback
}
// 4
fun incrementBy(incrementValue: Int) {
if (activeResources.addAndGet(incrementValue) < 1 &&
callback != null) {
callback!!.onTransitionToIdle()
}
}
}
This class is an implementation of the IdlingResource interface. There is a

lot going on here, so let's break it down:
1. Setting up a ResourceCallback reference to tell Espresso when it is

transitioning to idle.
2. Creating a counter to keep track of the current active resources.
3. Returns the idling status based on the number of active resources.
4. Increments the active resources count by the number passed in and

transitions to idle if this new value is less than 1.
Now that you have your SimpleIdlingResource, you are going to need a
way to trigger it when something happens. You could move this to your
app code, call it from there, and access it from your test. But, there is a
way that is a little bit cleaner using EventBus.
EventBus is a library that makes it easy to subscribe to and publish

messages. If you haven't used it before you can learn all about it at
https://github.com/greenrobot/EventBus.
To get started, add the following to your app level build.gradle:
implementation 'org.greenrobot:eventbus:3.1.1'
EventBus posts and receives messages as data objects (often called Plain
Old Java Objects, or POJOs in Java). These objects need to be in your app.
Under com.raywenderlich.codingcompanionfinder in the main source
set, create a new package called testhooks. In that package, create a
Kotlin file called IdlingEntity.kt, and add the following content:
data class IdlingEntity(

var incrementValue: Int = 0,
var resetValue: Boolean = false
)
Next, open your MainActivity and add the following:
// 1
@Subscribe
fun onEvent(idlingEntity: IdlingEntity) {
// noop
}
// 2
override fun onStart() {
super.onStart()
EventBus.getDefault().register(this)
}
// 3
override fun onStop() {
super.onStop()
EventBus.getDefault().unregister(this)
}
This is doing three things:
1. Adding a subscription, which is required for the EventBus library to

work.
2. Registering this class with EventBus when the activity starts.
3. Unregistering this class with Eventbus when the activity stops.
Now, open your SearchForCompanion fragment in the searchforcompanion

package and go to your searchForCompanions function. Add a post
command to increment your Idling resources before you call your
petfinder service:
EventBus.getDefault().post(IdlingEntity(1))
And another to decrement it once it is done with the call:
EventBus.getDefault().post(IdlingEntity(-1))
Your final method should look like this:

private fun searchForCompanions() {
val companionLocation = view?
.findViewById<TextInputEditText>(R.id.searchFieldText)
?.text.toString()
val noResultsTextView = view?
.findViewById<TextView>(R.id.noResults)
val searchForCompanionFragment = this
GlobalScope.launch {
accessToken = (activity as MainActivity).accessToken
(activity as MainActivity).petFinderService
?.let { petFinderService ->
// increment the IdlingResources
val getAnimalsRequest = petFinderService
.getAnimals(accessToken, location = companionLocation)
val searchForPetResponse = getAnimalsRequest.await()
if (searchForPetResponse.isSuccessful) {
searchForPetResponse.body()?.let {
GlobalScope.launch(Dispatchers.Main) {
if (it.animals.size > 0) {
noResultsTextView?.visibility = INVISIBLE
viewManager = LinearLayoutManager(context)
companionAdapter = CompanionAdapter(it.animals,
searchForCompanionFragment)
petRecyclerView = view?.let {
it.findViewById<RecyclerView>(
R.id.petRecyclerView
).apply {
layoutManager = viewManager
adapter = companionAdapter
}
}
} else {
noResultsTextView?.visibility = VISIBLE
}
}
}
} else {
}
// Decrement the idling resources.
}
}
}
You're almost there! Open up your FindCompanionInstrumentedTest.kt
and add a line to create a SimpleIdlingResource as a property at the class
level:
private val idlingResource = SimpleIdlingResource()
Now add a subscribe function to receive increment/decrement calls:
@Subscribe
idlingResource.incrementBy(idlingEntity.incrementValue)
}
Next, in
searching_for_a_companion_and_tapping_on_it_takes_the_user_to_the_compan
ion_details(), add these two lines after you launch your activity:
IdlingRegistry.getInstance().register(idlingResource)
This registers your test class with EventBus and registers the idling
resources. Finally, add these two lines at the end of that function before
testScenario.close():
IdlingRegistry.getInstance().unregister(idlingResource)
Your final function should look like this:

@Test
// eventbus and idling resources register

.perform(click())
onView(withId(R.id.searchButton)).perform(click())
// eventbus and idling resources unregister.

}
Run your tests and everything will be green.
Now that you have all of this in place, it's time to add the shelter
information to your companion details page. The test for this will be very
similar to the last test you added. You are going to go to your Find
Companion page, search by a location, select a companion, and then
verify that the contact information is correct. The only difference will be
what you are checking for.
DRYing up your tests

One term you will hear when someone speaks about software as a craft is
writing DRY code. DRY stands for Do Not Repeat Yourself. In practical
terms, this means that you should try to avoid multiple lines of duplicate
code in your app.
This is also a good thing to do with tests. That said, your tests, if you are
doing TDD well, provide a form of documentation for the code. If drying
up your tests makes it easier to maintain your tests and makes them
more readable, by all means do it, but if a particular effort to dry out the
tests doesn't add a significant maintainability benefit, or makes them
more difficult to read, it may be better not to do that refactor.
Looking at your current tests, you have the following code at the
beginning of both:
You also have the following at the end of both:
The common code at the beginning of both tests can be moved to a

common @Before method and will make both tests more readable by
moving common set-up information out of the individual tests. To get
started, add the following method to your test class:
@Before
fun beforeTestsRun() {
}
The @Before annotation tells the test suite to run that function before
every test. Now, add the following method:
@After
fun afterTestsRun() {
}
Next, remove the following from your

pressing_the_find_bottom_menu_item_takes_the_user_to_the_find_page() and
ion_details() tests:
And:
Finally, run the tests and everything will still be green.
Your next test is going to use the IdlingRegistry. Having that run before
your pressing_the_find_bottom_menu_item_takes_the_user_to_the_find_page
will not affect that test. It will also make
ion_details and your next test more readable. Move the following two
lines from
ion_details to the end of the beforeTestRun() function:
Next, cut the following two lines from the end of

ion_details:
And add them to the beginning of your afterTestRun() function.
Finally, run the tests and make sure everything is green.
Your next test is going to share a lot of steps with

ion_details. So lets refactor some common functionality.
First, cut the following lines from
ion_details:
.perform(click())
Next, paste them into a new function:
private fun find_and_select_kevin_in_30318(){

.perform(click())
}
Now, add a call to your new function in

searching_for_a_companion_and_tapping_on_it_takes_the_user_to_the_co
mpanion_details:
@Test
find_and_select_kevin_in_30318()
}
Finally, run the tests and make sure everything is still green — it is
important to check your refactors haven't accidentally broken anything.
Writing your failing test

Now, it is time to write your failing test – which you will implement
afterwards. For this new test, you are going to check to make sure that
you can view the correct phone number and email address when you view
the details for Rocket.
To get started, add the following test:
@Test
fun verify_that_companion_details_shows_a_valid_phone_number_and_email(
onView(withText("(706) 236-4537"))
onView(withText("[email protected]"))
}
Now, run the test and make sure that it fails.
Next, open fragment_view_companion.xml and replace this:

<androidx.appcompat.widget.AppCompatTextView
android:id="@+id/breed"
android:text="@string/breed_placeholder"
app:layout_constraintBottom_toTopOf="@+id/age"
app:layout_constraintEnd_toStartOf="@id/city"
app:layout_constraintTop_toBottomOf="@+id/petCarouselView" />
android:id="@+id/city"
android:text="@string/city_placeholder"
app:layout_constraintBottom_toBottomOf="@id/breed"
app:layout_constraintStart_toEndOf="@+id/breed"
app:layout_constraintTop_toTopOf="@+id/breed" />
android:id="@+id/age"
android:text="@string/age_placeholder"
app:layout_constraintBottom_toTopOf="@id/meetTitlePlaceholder"
app:layout_constraintEnd_toStartOf="@id/sex"
app:layout_constraintTop_toBottomOf="@id/breed" />
With:
app:layout_constraintBottom_toTopOf="@+id/email"
android:text="@string/city_placeholder"
android:id="@+id/email"
android:text="email placeholder"
android:textStyle="bold"
app:layout_constraintEnd_toStartOf="@id/telephone"
app:layout_constraintTop_toBottomOf="@+id/breed" />
android:id="@+id/telephone"
android:text="telephone placeholder"
app:layout_constraintBottom_toBottomOf="@id/email"
app:layout_constraintStart_toEndOf="@+id/email"
app:layout_constraintTop_toTopOf="@+id/email" />
android:text="@string/age_placeholder"
app:layout_constraintTop_toBottomOf="@id/email" />
Next, open ViewCompanionFragment.kt and add the following to the
populatePet() function:
populateTextField(R.id.email, animal.contact.email)
populateTextField(R.id.telephone, animal.contact.phone)
Finally, run your tests and everything should be green.
Congratulations! Your app is under test and the shelter has a new feature
that will help them to place more coding companions!
Key points
When working with a legacy app, start by abstracting away external
dependencies.
Don't try to get everything under test at one time.
Focus your testing efforts around a section you are changing.
MockWebServer is a great way to mock data for Retrofit.
When getting a legacy app under test you will probably end up
needing to use IdlingResources.
DRY out your tests when it makes them more readable.
Don't try to refactor a section of your legacy app until it is under

test.

You've done a lot of work in this chapter! If you want to take some of
these techniques further, try writing some tests around more scenarios in
the app. You can also try your hand at adding additional features to the
app. To see what is available with the API check out the Petfinder API
documentation at https://www.petfinder.com/developers/api-docs.
Digging deeper, in the next chapter, Chapter 14, "Hands-On Focused

Refactoring," you will begin to explore how you can use TDD and
refactoring side-by-side.
Chapter 14: Hands-On Focused
Refactoring
In the last chapter, you had a chance to:
1. Get familiar with the Coding Companion app.
2. Add tests around the search functionality.
3. Add a feature to make it easier to find contact information about a

companion.
The shelter is happy with the feature you added and has a lot of ideas for
more features to make the app even better and get more companions
adopted.
Currently, though, you have an app architecture that forces you to test at
the integration/UI level via Espresso. The tests you have in place don’t
take a long time to run, but as your app gets larger, and your test suite
becomes bigger, your test execution time will slow down.
In Chapter 6, ”Architecting for Testing,” you learned about architecting

for testing and why an MVVM architecture helps to make apps more
readable and easier to test at a lower level. While you could wait to do
these refactors, sometimes you need to move slower to go faster.
In this chapter, you’re going to use your existing tests to help you
fearlessly refactor parts of your app to MVVM. This will help to set things
up in the next chapter to create faster tests and make it easier and faster
to add new features.
Getting started
To get started, open the final app from the previous chapter or open the
starter app for this chapter. Then, open
FindCompanionInstrumentedTest.kt located inside the androidTest
source set.
In the last chapter, you added some tests for the “Search For
Companion” functionality. You can find this test inside
FindCompanionInstrumentedTest.kt having the name
ion_details.
This test does the following:
1. It starts the app’s main activity, which takes the user to the Random
Companion screen; this screen is backed by RandomCompanionFragment.
1. Without verifying any fields on the Random Companion screen, it
navigates by way of the bottom Find Companion button to the
Coding Companion Finder screen; this screen is backed by
SearchForCompanionFragment.
1. Staying in SearchForCompanionFragment, it enters a valid United States

zipcode and clicks the Find button.
1. Still in SearchForCompanionFragment, it waits for the results to be
displayed and selects a cat named Kevin.
1. It then waits for the app to navigate to the Companion Details

screen — backed by the ViewCompanionDetails fragment — and
validates the city/state in which the selected companion is located.
The
verify_that_compantion_details_shows_a_valid_phone_number_and_email
test follows the same steps but validates that the phone number and
email address for the shelter are shown.
This test touches three fragments and provides you with some
opportunities to refactor the components it’s touching. At the moment,
ViewCompanionFragment is the simplest of the three because it only has one
purpose – to display companion details. Therefore, you’ll start by
refactoring this test.
Adding supplemental coverage before refactoring

You already have some testing around the “Search For Companion”
functionality, including ViewCompanionFragment. Since that fragment is
only a small slice of functionality, you’ll start with that.
Before you start to refactor, you need to make sure you have tests around
everything that you’re changing. This helps to ensure that your
refactoring doesn’t accidentally break anything. Because you’re changing
things to an MVVM architecture, you’re going to touch all of the data
elements this fragment displays.
Looking at the two tests that test this screen, in

FindCompanionsInstrumentedTest.kt, you’ll see the following:
@Test
}
@Test
fun verify_that_companion_details_shows_a_valid_phone_number_and_email(
}
This is testing some of the fields in the View Companion details, but not
all of them. Because Espresso tests are slow, it’s better to add these
checks to one of your existing tests.
In this case, you’re going to use

ion_details, so paste the following to the end of that test:
onView(withText("Domestic Short Hair")).check(matches(isDisplayed()))

onView(withText("Young")).check(matches(isDisplayed()))
onView(withText("Female")).check(matches(isDisplayed()))
onView(withText("Medium")).check(matches(isDisplayed()))
onView(withText("Meet KEVIN")).check(matches(isDisplayed()))
Your test will now look like this:
@Test
onView(withText("Domestic Short Hair")).check(matches(isDisplayed()))
onView(withText("Young")).check(matches(isDisplayed()))
onView(withText("Medium")).check(matches(isDisplayed()))
onView(withText("Meet KEVIN")).check(matches(isDisplayed()))
}
Run it, and you’ll see the following:

According to this message, the view hierarchy has more than one field
with text containing “Domestic Short Hair”.
Refactoring for testability

With Espresso tests, you’ll often run into a scenario where you have a
matcher for an element that ends up matching more than one element in
the view hierarchy. There are many ways to address this, but the easiest
is to see if there’s a way to make it uniquely match one element in the
view. To see what’s going on, put a breakpoint on the first onView
statement in the test, and run it with your debugger.
Looking at the app screen, you’ll see the following:

The screen only displays “Domestic Short Hair” once. So, what’s
happening here?
The CompanionViewHolder holds some clues. Look at setupClickEvent in this

View Holder, and you’ll see the following line:
transaction.replace(R.id.viewCompanion, viewCompanionFragment).addToBac
Click on R.id.viewCompanion to find it in your view, and you’ll see that it

isn’t displaying viewCompanionFragment in a FrameLayout.
This FrameLayout is on the same level as a ConstraintLayout that has a
RecyclerView, which ultimately displays the search results.
This FrameLayout also has a higher Z value, which makes it display over
the ConstraintLayout.
Look at the setupClickEvent in CompanionViewHolder, and you’ll see that
you’re doing a transaction to replace R.id.viewCompanion with a
ViewCompanionFragment.
transaction.replace(R.id.viewCompanion, viewCompanionFragment)
.addToBackStack("companionView")
.commit()
The issue is most likely that two views show this information — but one
is hiding below the other. One way to fix this problem might be to also
match on the ID of the field.
Currently, in find_pet_list_layout.xml, you’ve given the field that contains

the breeds an ID of breed which is displayed by the CompanionViewHolder in
the RecyclerView.
android:layout_width="0dp"
android:paddingEnd="10dp"
android:paddingStart="10dp"
android:text="Breed"
app:layout_constraintBottom_toBottomOf="@+id/sex"
app:layout_constraintStart_toEndOf="@id/sex"
app:layout_constraintTop_toTopOf="@id/sex" />
It’s also named breed in fragment_view_companion.xml that is displayed

by your ViewCompanionFragment.
You could change the ID of the breed in the ViewCompanionFragment, but a

better approach is to do a full replacement of the fragment, so you don’t
have two simultaneous view hierarchies in the ViewCompanionFragment.
Since you’re already using the Jetpack Navigation Components, this is a

good time to do a refactor to use this. If you’re new to Android
Navigation Components, you can learn more about them at
https://developer.android.com/guide/navigation.
Open nav_graph.xml inside res ‣ navigation and add the following inside
the <navigation> element:
<fragment
android:id="@+id/viewCompanion"
android:name="com.raywenderlich.codingcompanionfinder.searchforcompani
android:label="fragment_view_companion"
tools:layout="@layout/fragment_view_companion" >
<argument
android:name="animal"
app:argType="com.raywenderlich.codingcompanionfinder.models.Animal"
</fragment>
This adds the ViewCompanionFragment to the navigation graph.
Next, replace:
<fragment
tools:layout="@layout/fragment_search_for_companion" />
With the following:
<fragment
tools:layout="@layout/fragment_search_for_companion" >
<action
android:id="@+id/action_searchForCompanionFragment_to_viewCompanion
app:destination="@id/viewCompanion" />
</fragment>
This adds an action to allow you to navigate between the

SearchForCompanion and ViewCompanion fragment.
Looking at clickListener() in the CompanionViewHolder, you’re passing an

animal object to that fragment:
val viewCompanionFragment = ViewCompanionFragment()
val bundle = Bundle()
bundle.putSerializable(ViewCompanionFragment.ANIMAL, animal)
viewCompanionFragment.arguments = bundle
val transaction =
fragment.childFragmentManager.beginTransaction()
transaction.replace(R.id.searchForCompanion,
viewCompanionFragment)
.addToBackStack("companionView")
.commit()
}
To pass the arguments with Jetpack Navigation, you’ll use Safe Args. If
you’ve not used this before, you can learn more about it at
https://developer.android.com/guide/navigation/navigation-pass-
data#Safe-args.
Open CodingCompanionFinder build.gradle and add the following to the

dependencies:
classpath "androidx.navigation:navigation-safe-args-gradle-plugin:2.1.0
Next, open your app level build.gradle and add the following to the top of
the file:
apply plugin: "androidx.navigation.safeargs.kotlin"
This adds in Safe Args support.
Now, open CompanionViewHolder.kt and replace setupClickEvent() with

the following, fixing the imports as needed:
private fun setupClickEvent(animal: Animal) {

val action = SearchForCompanionFragmentDirections
.actionSearchForCompanionFragmentToViewCompanion(animal)
view.findNavController().navigate(action)
}
}
This is using SearchForCompanionFragmentDirections which is generated by

Safe Args to create a navigation action with the animal as a parameter.
You’re then passing the action to the navigate method on the navigation
controller to perform the navigation to the ViewCompanionFragment.
Finally, open ViewCompanionFragment.kt in the same package and add

the following property:
val args: ViewCompanionFragmentArgs by navArgs()
Then replace onCreateView with the following:
override fun onCreateView(

inflater: LayoutInflater, container: ViewGroup?,
savedInstanceState: Bundle?
): View? {
// Inflate the layout for this fragment
animal = args.animal
viewCompanionFragment = this
return inflater.inflate(R.layout.fragment_view_companion,
container, false)
}
This retrieves the arguments passed to the fragment via

ViewCompanionFragmentArgs which is generated by Safe Args.
Open fragment_search_for_companion.xml and remove the FrameLayout

with an android:id of @+id/viewCompanion.
Finally, execute the

ion_details test in FindCompanionInstrumentedTest.kt, and it’ll be
green.
Your first focused refactor

Now that you have proper test coverage around ViewCompanionFragment, it’s
time to refactor it. To get started, open the app level build.gradle and add
the following to the dependencies section:
// Architecture components
def lifecycle_version = "2.0.0"
implementation "androidx.lifecycle:lifecycle-extensions:$lifecycle_vers
kapt "androidx.lifecycle:lifecycle-compiler:$lifecycle_version"
This is adding the Jetpack Lifecycle components. Next, add the following
Android section below buildTypes in the same file to enable data binding:
dataBinding {
enabled = true
}
Following that, create a Kotlin file named ViewCompanionViewModel.kt

in the searchforcompanion package and add the following:
data class ViewCompanionViewModel(

var name: String = "",
var breed: String = "",
var city: String = "",
var email: String = "",
var telephone: String = "",
var age: String = "",
var sex: String = "",
var size: String = "",
var title: String = "",
var description: String = ""
): ViewModel()
This creates a ViewModel for the data.
Next, open fragment_view_companion.xml and add a <layout> tag around

the ConstraintLayout along with a <data> and <variable> tag for the view
model, so it looks like this:
<layout>
<data>
<variable
name="viewCompanionViewModel"
type="com.raywenderlich.codingcompanionfinder.searchforcompanion.V
</data>
<androidx.constraintlayout.widget.ConstraintLayout xmlns:android="http
android:layout_height="match_parent"
android:background="@color/secondaryTextColor"
android:translationZ="5dp"
tools:context=".randomcompanion.RandomCompanionFragment">
.
.
.
</androidx.constraintlayout.widget.ConstraintLayout>
</layout>
This adds the ability to bind data from the ViewCompanionViewModel to this
view.
Now, bind each attribute of the view model to each element with a
corresponding ID by replacing the text with the binding. For example, for
the element with an ID of “name”, you’ll replace the text with
@{viewCompanionViewModel.name}.
Your final fragment_view_companion.xml will look like this:

<layout>
<data>
<variable
name="viewCompanionViewModel"
type="com.raywenderlich.codingcompanionfinder.searchforcompanion.V
</data>
<androidx.constraintlayout.widget.ConstraintLayout xmlns:android="htt
android:background="@color/secondaryTextColor"
android:translationZ="5dp"
tools:context=".randomcompanion.RandomCompanionFragment">
android:id="@+id/petName"
android:layout_marginTop="10dp"
android:layout_marginBottom="5dp"
android:text="@{viewCompanionViewModel.name}"
android:textSize="24sp"
app:layout_constraintBottom_toTopOf="@id/petCarouselView"
app:layout_constraintTop_toTopOf="parent" />
<com.synnapps.carouselview.CarouselView
android:id="@+id/petCarouselView"
android:layout_marginBottom="5dp"
app:fillColor="#FFFFFFFF"
app:layout_constraintBottom_toTopOf="@id/breed"
app:layout_constraintTop_toBottomOf="@id/petName"
app:layout_constraintWidth_percent=".6"
app:pageColor="#00000000"
app:radius="6dp"
app:slideInterval="3000"
app:strokeColor="#FF777777"
app:strokeWidth="1dp" />
android:text="@{viewCompanionViewModel.breed}"
android:text="@{viewCompanionViewModel.city}"
android:id="@+id/email"
android:text="@{viewCompanionViewModel.email}"
app:layout_constraintEnd_toStartOf="@id/telephone"
app:layout_constraintTop_toBottomOf="@+id/breed" />
android:id="@+id/telephone"
android:text="@{viewCompanionViewModel.telephone}"
app:layout_constraintBottom_toBottomOf="@id/email"
app:layout_constraintStart_toEndOf="@+id/email"
app:layout_constraintTop_toTopOf="@+id/email" />
android:text="@{viewCompanionViewModel.age}"
app:layout_constraintTop_toBottomOf="@id/email" />
android:id="@+id/sex"
android:text="@{viewCompanionViewModel.sex}"
app:layout_constraintBottom_toBottomOf="@id/age"
app:layout_constraintEnd_toStartOf="@id/size"
app:layout_constraintStart_toEndOf="@id/age"
app:layout_constraintTop_toTopOf="@id/age" />
android:id="@+id/size"
android:text="@{viewCompanionViewModel.size}"
app:layout_constraintBottom_toBottomOf="@id/age"
app:layout_constraintStart_toEndOf="@id/sex"
app:layout_constraintTop_toTopOf="@id/age" />
android:id="@+id/meetTitlePlaceholder"
android:text="@{viewCompanionViewModel.title}"
app:layout_constraintBottom_toTopOf="@+id/descriptionScroll"
app:layout_constraintTop_toBottomOf="@+id/age" />
<androidx.core.widget.NestedScrollView
android:id="@+id/descriptionScroll"
android:paddingStart="30dp"
android:paddingEnd="30dp"
app:layout_constraintHeight_percent=".25"
app:layout_constraintHorizontal_bias="0.0"
app:layout_constraintVertical_bias="1.0">
android:id="@+id/description"
android:text="@{viewCompanionViewModel.description}" />
</androidx.core.widget.NestedScrollView>
</layout>
Build and run to make sure the app compiles and that there are no errors.
With your view in good shape, go back to ViewCompanionViewModel.kt

and add the following method to the ViewCompanionViewModel class:
fun populateFromAnimal(animal: Animal) {

name = animal.name
breed = animal.breeds.primary
city = animal.contact.address.city + ", " +
animal.contact.address.state
email = animal.contact.email
telephone = animal.contact.phone
age = animal.age
sex = animal.gender
size = animal.size
title = "Meet " + animal.name
description = animal.description
}
This is a helper method for converting an Animal object to your ViewModel.
Now go to ViewCompanionFragment.kt and replace onCreateView with the

following:

): View? {
animal = args.animal
viewCompanionFragment = this
// 1
val fragmentViewCompanionBinding =
FragmentViewCompanionBinding
.inflate(inflater, container, false)
// 2
val viewCompanionViewModel = ViewModelProviders.of(this)
.get(ViewCompanionViewModel::class.java)
// 3
viewCompanionViewModel.populateFromAnimal(animal)
// 4
fragmentViewCompanionBinding.viewCompanionViewModel =
viewCompanionViewModel
// 5
return fragmentViewCompanionBinding.root
}
If you find that the import for FragmentViewCompanionBinding is not
behaving, do a build and then try again.
The code you just added does the following:
1. It inflates the view via a data-binding-generated

FragmentViewCompanionBinding object.
2. Creates an instance of ViewCompanionViewModel via the

ViewModelProviders.
3. Populates the view model from an Animal.
4. Assigns the view model to your view.
5. Returns the root of the view.
Finally, in onResume, replace the call to populatePet() with populatePhotos()
Run your test now, and it’ll be green.
There’s still one other piece of this refactor that you’ll need to do to wrap
things up. With your data binding, you no longer need populatePet() or
populateTextField(...), so delete them.
Your next refactor

Swapping manual view binding for data binding in the
ViewCompanionFragment was a relatively simple refactor. Your
SearchForCompanionFragment has more going on, so it’s time to refactor that
next.
Adding test coverage

Just like you did with the ViewCompanionFragment test, you want to make
ensure that you have enough test coverage for the
SearchForCompanionFragment.
Three things happen in this fragment:
1. It presents the user with a screen to search for a companion.
1. It gets the user’s input and performs a search.

1. It presents the search results and allows navigation to the
Open FindCompanionInstrumentedTest.kt located inside androidTest
source set.
Looking through the tests, two of the three tests are referencing the
following method:
private fun find_and_select_kevin_in_30318() {
.perform(click())
}
This does a good job of testing most of the three scenarios with one
exception: It does not verify all of the data when you list the results of a
search. To fix that, you’ll write a test, but there’s one small change you
need to make to your test data first.
If you look at your search results data, you have two animals that are
both females. But earlier, you learned about it being difficult to match
multiple elements with the same value/ID. To make things easier to test,
you’ll change the sex of one of the companions.
Start by opening search_30318.json, which is located inside assets in the

androidTest source set. Then, find the first instance of the gender
attribute, which is associated with the Shih Tzu named Joy.
Next, change the gender to Male.
Following that, open FindCompanionInstrumentedTest.kt and add the

following test:
@Test
fun searching_for_a_companion_in_30318_returns_two_results() {
.perform(click())
onView(withText("Joy")).check(matches(isDisplayed()))
onView(withText("Male")).check(matches(isDisplayed()))
onView(withText("Shih Tzu")).check(matches(isDisplayed()))
onView(withText("KEVIN")).check(matches(isDisplayed()))
onView(withText("Domestic Short Hair"))
}
This verifies all of the data elements for the search results without
clicking on one like the other tests are doing.
Finally, run the test, and everything will be green.
Note: For the sake of brevity, you’re not breaking these test
conditions before making them pass. Before you move on, however,
a good exercise is to try changing various data elements to ensure
that each assertion breaks before setting the data back to a state
that makes the test pass.
There are two other scenarios that you need to address.
Looking at searchForCompanions() in SearchForCompanionFragment.kt,

there are two instances that can lead to a text view with a message
indicating that no results are available:
// No Results Text View is invisible when results are available.
it.findViewById<RecyclerView>(R.id.petRecyclerView)
.apply {
}
}
} else {
// No Results Text View is visible when results are not
// available.
}
}
}
} else {
// No Results Text View is visible when results are not
// available.
}
This displays by going to the app and searching for companions under an
invalid location.
There are two scenarios for which you need to add coverage:
1. When the user enters a valid location, but there are no results.
2. When the user enters an invalid location.
You’ll start with the first scenario.
Open CommonTestDataUtil.kt inside androidTest and replace dispatch

with the following:
MockResponse()
.setBody(readFile("search_30318.json"))
}
// test data for no response
MockResponse()
.setBody("{\"animals\": []}")
}
else -> {
}
}
}
This adds a mock for a zip code location of 90210 that returns no results.
Next, add the following test to FindCompanionsInstrumentedTest.kt:
@Test
fun searching_for_a_companion_in_90210_returns_no_results() {
.perform(click())
onView(withId(R.id.noResults))
.check(matches(withEffectiveVisibility(Visibility.VISIBLE)))
}
Since it’s a good idea to have a failing test first, go into

SearchForCompanionFragment.kt and comment out the line that sets the
visibility for noResultsTextView:
it.findViewById<RecyclerView>(R.id.petRecyclerView).apply {
}
}
} else {
// Comment out this line
//noResultsTextView?.visibility = VISIBLE
}
Now, run your new test, and it’ll fail.
Finally, uncomment that line, run the test again — and this time, it
passes.
For the second no results scenario, open

FindCompanionInstrumentedTest.kt and add the following test:
@Test
fun searching_for_a_companion_in_a_call_returns_an_error_displays_no_re
.perform(click())
onView(withId(R.id.searchFieldText)).perform(typeText("dddd"))
.check(matches(withEffectiveVisibility(Visibility.VISIBLE)))
}
Run this test without commenting out the implementation, and you’ll
see a failure message that reads:
Test failed to run to completion. Reason: 'Instrumentation run failed d
Test running failed: Instrumentation run failed due to 'Process crashed
Looking at the code in SearchForCompanions in the

SearchForCompanionFragment, you’ll see the following:
// This is a bug, the scope should be at a higher level.
.apply {
}
}
} else {
}
}
}
} else {
// This is running in the wrong thread
}
There’s a bug in the app where a “no results” scenario causes the app to
crash. This happens because it’s trying to set a value in the view outside
of the main thread.
To fix this error, move the GlobalScope.launch(Dispatchers.Main) line to

the outside of your code block below val searchForPetResponse =
getAnimalsRequest.await(). When you're done, it should look like this:
.apply {
}
}
} else {
}
}
} else {
}
}
Now, run your test and it’ll be green.
Refactoring SearchForCompanionFragment
Now that you have adequate coverage for this section, it’s time to do
some refactoring.
To get started, create a new file named

SearchForCompanionViewModel.kt in the searchforcompanion package.
Give it the following content:
class SearchForCompanionViewModel: ViewModel() {

val noResultsViewVisiblity : MutableLiveData<Int> =
MutableLiveData<Int>()
val companionLocation : MutableLiveData<String> =
MutableLiveData()
}
This creates a ViewModel for the fragment with LiveData elements for the
noResults View and companionLocation.
Next, open fragment_search_for_companion.xml and add a <layout> tag

around the ConstaintLayout. Also, add a <data> and <variable> tag for the
ViewModel:
<layout>
<data>
<variable
name="searchForCompanionViewModel"
type="com.raywenderlich.codingcompanionfinder.searchforcompanion.S
</data>
tools:context=".searchforcompanion.SearchForCompanionFragment">
.
.
.
</layout>
Now bind the SearchForCompanion ViewModel's companionLocation to the

searchField text attribute of the <TextInputEditText> with the ID of
@+id/searchFieldText by adding:
android:text="@={searchForCompanionViewModel.companionLocation}"
Also, bind that ViewModel’s noResultsViewVisibility to the visibility

attribute of the <TextView> with the ID of @+id/noResults by replacing:
android:visibility="invisible"
With the following:
android:visibility="@{searchForCompanionViewModel.noResultsViewVisiblit
The final fragment_search_for_companion.xml will look like this:
<?xml version="1.0" encoding="utf-8"?>
<layout>
<data>
<variable
name="searchForCompanionViewModel"
type="com.raywenderlich.codingcompanionfinder.searchforcompanion.S
</data>
tools:context=".searchforcompanion.SearchForCompanionFragment">
<androidx.constraintlayout.widget.ConstraintLayout
android:id="@+id/searchForCompanion"
app:layout_constraintTop_toTopOf="parent">
<com.google.android.material.textfield.TextInputLayout
android:id="@+id/searchField"
app:layout_constraintBottom_toTopOf="@id/petRecyclerView"
app:layout_constraintEnd_toStartOf="@id/searchButton"
app:layout_constraintWidth_percent=".7">
<com.google.android.material.textfield.TextInputEditText
android:id="@+id/searchFieldText"
android:text="@={searchForCompanionViewModel.companionLocatio
android:hint="Enter US Location"
android:textColor="@color/primaryTextColor" />
</com.google.android.material.textfield.TextInputLayout>
<com.google.android.material.button.MaterialButton
android:id="@+id/searchButton"
android:text="Find"
app:layout_constraintBottom_toBottomOf="@+id/searchField"
app:layout_constraintStart_toEndOf="@id/searchField"
app:layout_constraintTop_toTopOf="@id/searchField" />
<androidx.recyclerview.widget.RecyclerView
android:id="@+id/petRecyclerView"
app:layout_constraintTop_toBottomOf="@id/searchField" />
<TextView
android:id="@+id/noResults"
android:text="No Results"
android:textSize="36sp"
android:visibility="@{searchForCompanionViewModel.noResultsView
app:layout_constraintTop_toBottomOf="@id/searchField" />
</layout>
With that done, go back to SearchForCompanionFragment.kt and replace

onCreateView with the following:
private lateinit var fragmentSearchForCompanionBinding:
FragmentSearchForCompanionBinding
private lateinit var searchForCompanionViewModel:
SearchForCompanionViewModel

): View? {
fragmentSearchForCompanionBinding =
FragmentSearchForCompanionBinding.inflate(inflater,
container, false)
searchForCompanionViewModel = ViewModelProviders.of(this)
.get(SearchForCompanionViewModel::class.java)
fragmentSearchForCompanionBinding.searchForCompanionViewModel
= searchForCompanionViewModel
fragmentSearchForCompanionBinding.lifecycleOwner = this
return fragmentSearchForCompanionBinding.root
}
Note: If you find the FragmentSearchForCompanionBinding import not

resolving, perform a build.
Locate searchForCompanions() and replace it with the following:

private fun searchForCompanions() {
// 1
val searchForCompanionFragment = this
accessToken = (activity as MainActivity).accessToken
(activity as MainActivity).petFinderService?
.let { petFinderService ->
// 2
val getAnimalsRequest = petFinderService.getAnimals(
accessToken,
location =
searchForCompanionViewModel.companionLocation.value
)
// 3
searchForCompanionViewModel
.noResultsViewVisiblity
.postValue(INVISIBLE)
it.findViewById<RecyclerView>(
R.id.petRecyclerView
).apply {
}
}
} else {
// 3
.postValue(VISIBLE)
}
}
} else {
// 3
.postValue(VISIBLE)
}
}
}
}
}
This does the following:
1. Removes the findViewById references for the two text elements.
2. Uses the bound value from the ViewModel to pass the location that a
user is searching for into the web API call.
3. Uses the bound value for noResultsViewVisibility to set the visibility

of the No Results text whether or not results are found.
Run the tests in FindCompanionsInstrumentedTest.kt, and they’ll all still

be green. Great refactor!
This is a good first step at refactoring SearchForCompanionFragment, but

there’s still a lot of logic in your controller.
searchForCompanions()has a lot of stuff going on with its calls to Retrofit;

these can be moved to the ViewModel. This allows you to bring the testing
of this component down to a unit level, which you’ll do in the next
chapter.
To get started, open SearchForCompanionViewModel.kt and add the

following:
// 1
val animals: MutableLiveData<ArrayList<Animal>> =
MutableLiveData<ArrayList<Animal>>()
lateinit var accessToken: String
lateinit var petFinderService: PetFinderService
fun searchForCompanions() {
// 2
accessToken,
location = companionLocation.value
)
// 3
animals.postValue(it.animals)
// 3
noResultsViewVisiblity.postValue(INVISIBLE)
} else {
// 3
noResultsViewVisiblity.postValue(View.VISIBLE)
}
}
} else {
// 3
}
}
}
}
This is a refactored version of the controller’s searchForCompanions

method; it does three things:
1. Creates some variables used to pass data between your ViewModel,

View layout and Fragment.
2. Calls petFinderService to make calls to the API.

3. Sets appropriate values in the ViewModel used in your View layout.
Next, open the SearchForCompanionFragment and add the following method:
private fun setupSearchForCompanions() {

// 1
searchForCompanionViewModel.accessToken =
(activity as MainActivity).accessToken
searchForCompanionViewModel.petFinderService =
(activity as MainActivity).petFinderService!!
// 2
companionAdapter = CompanionAdapter(
searchForCompanionViewModel.animals.value ?: arrayListOf(),
this
)
petRecyclerView = fragmentSearchForCompanionBinding
.petRecyclerView.apply {
}
// 3
searchForCompanionViewModel.animals.observe(this, Observer<ArrayList<
companionAdapter.animals = it ?: arrayListOf()
companionAdapter.notifyDataSetChanged()
})
}
This does the following:
1. Passes the Retrofit service and access token into your ViewModel.
2. Sets up the RecyclerView for the list of companions.
3. Observes changes to the list of animals which occurs when results

come back from a search. It also updates the RecyclerView with the
new data when that happens.
Following that, in the same fragment, replace onActivityCreated with the

following:
override fun onActivityCreated(savedInstanceState: Bundle?) {
// 1
fragmentSearchForCompanionBinding.searchButton
.setOnClickListener {
try {
val inputMethodManager = activity?.getSystemService(
Context.INPUT_METHOD_SERVICE) as InputMethodManager?
inputMethodManager!!.hideSoftInputFromWindow(
activity?.getCurrentFocus()?.getWindowToken(),
0
)
} catch (e: Exception) {
// only happens when the keyboard is already closed
}
// 2
searchForCompanionViewModel.searchForCompanions()
}
// 3
setupSearchForCompanions()
super.onActivityCreated(savedInstanceState)
}
This code:
1. Migrates the findViewById call to using the data binding reference to

get your search button.
2. Replaces the call to the fragment’s local searchForCompanions with a

call to the same method name on the searchForCompanionViewModel.
3. Adds a call to the new setupSearchForCompanions in this fragment.
Now that you made those changes, you can remove searchForCompanions()
in the SearchForCompanionFragment.
Finally, call run on all of your tests in

FindCompanionsIntrumentedTest.kt and they’ll remain green.
Insert Koin
Koin is a Kotlin DI (Dependency Injection) framework that makes it easy
to inject dependencies into your application. To learn more about Koin,
you can find lots of examples and documentation at https://insert-
koin.io/.
In the next chapter, you’ll make use of Koin when you refactor some of
your tests. But since you’re refactoring your code now, you can add Koin
now.
To get started, add the following to the app level build.gradle:
// Koin
implementation 'org.koin:koin-android-viewmodel:1.0.1'
androidTestImplementation 'org.koin:koin-test:1.0.1'
This brings the Koin dependencies into your project.
Next, open MainActivity.kt. You need to move this code:
val apiKey = "replace with your API key"

val apiSecret = "replace with your API secret"
And place it into a companion object. You also need to rename them to
API_KEY and API_SECRET:
val API_KEY = "your api ket"

val API_SECRET = "your api secret"
Now, add another value in the companion object named

DEFAULT_PETFINDER_URL:
val DEFAULT_PETFINDER_URL = "http://api.petfinder.com/v2/"
The final companion object should look like:

companion object {
val PETFINDER_URI = "petfinder_uri"
val PETFINDER_KEY = "petfinder_key"
val API_KEY = "your client id"
val API_SECRET = "your client secret"
val DEFAULT_PETFINDER_URL = "https://api.petfinder.com/v2/"
}
Next, remove the Intent String fetch on line 3 from onCreate:
// remove these!!
intent.getStringExtra(PETFINDER_KEY)?.let {
apiKey = it
}
Following that, remove the following line since it’s no longer needed:
var token: Token = Token()
Open AuthorizationInterceptor.kt located inside the retrofit package in

your project, and replace it with the following:
// 1
class AuthorizationInterceptor : Interceptor, KoinComponent {
// 2
private val petFinderService: PetFinderService by inject()
private var token = Token()
override fun intercept(chain: Interceptor.Chain): Response {
var mainResponse = chain.proceed(chain.request())
val mainRequest = chain.request()
if ((mainResponse.code() == 401 ||
mainResponse.code() == 403) &&
!mainResponse.request().url().url().toString()
.contains("oauth2/token")) {
// 3
val tokenRequest = petFinderService.getToken(
clientId = MainActivity.API_KEY,
clientSecret = MainActivity.API_SECRET)
val tokenResponse = tokenRequest.execute()
if (tokenResponse.isSuccessful()) {
tokenResponse.body()?.let {
token = it
val builder = mainRequest.newBuilder()
.header("Authorization", "Bearer " +
it.accessToken)
.method(mainRequest.method(), mainRequest.body())
mainResponse = chain.proceed(builder.build())
}
}
}
return mainResponse
}
This changes the following things:
1. It removes the dependencies that needed to be passed into this class

when it’s created. It also adds a dependency on KoinComponent that
allows you to inject dependencies into the class.
2. It injects PetFinderService and brings the token it needs to

periodically refresh into the interceptor class.
3. It uses the companion object parameters from MainActivity for
CLIENT_ID and CLIENT_SECRET.
Go to MainActivity.kt and remove the this parameter from the following

line in onCreate:
.addInterceptor(AuthorizationInterceptor(this))
Change it to:
.addInterceptor(AuthorizationInterceptor())
Next, open SearchForCompanionViewModel, and add val

petfinderService: PetfinderService as a constructor parameter, like this:
class SearchForCompanionViewModel(
val petFinderService: PetFinderService
): ViewModel() {
Now, remove:
This makes it easier to inject the PetFinderService into

SearchForCompanionViewModel.
Koin also requires a KoinModule that tells it what to inject.
Create a new file in the main project package named KoinModule.kt and
add the following:
const val PETFINDER_URL = "PETFINDER_URL"
val urlsModule = module {

single(name = PETFINDER_URL) {
MainActivity.DEFAULT_PETFINDER_URL
}
}
val appModule = module {

single<PetFinderService> {

.build()
Retrofit.Builder()
.baseUrl(get(PETFINDER_URL) as String)
.client(client)
}
viewModel { ViewCompanionViewModel() }
viewModel { SearchForCompanionViewModel(get()) }
}
The appModule is creating a single instance of PetFinderService and allows

it to be injected as needed. It’s also creating instances of the
ViewModels, which under the hood uses the Jetpack ViewModelFactory
that binds to the lifecycle of the Fragment. The urlsModule creates a
string that references the Petfinder URL and is used in appModule.
In the main project package, com.raywenderlich.codingcompanionfinder

create a file named CodingCompanionFinder.kt and add the following
content:
class CodingCompanionFinder: Application() {

override fun onCreate() {
super.onCreate()
startKoin(this, listOf(appModule, urlsModule))
}
}
This adds some code to initialize Koin when your app is started.
Now, open AndroidManifest.xml and add

android:name=".CodingCompanionFinder" to the application tag so that it
looks like this:
<application
android:name=".CodingCompanionFinder"
android:allowBackup="true"
android:icon="@mipmap/ic_coding_companion"
android:label="@string/app_name"
android:roundIcon="@mipmap/ic_coding_companion_round"
android:supportsRtl="true"
android:usesCleartextTraffic="true"
android:theme="@style/AppTheme">
.
.
.
This tells the app to use the new Application object when starting the
app.
Following that, open SearchForCompanionFragment.kt under the

searchforcompanion package and change:
private lateinit var searchForCompanionViewModel:

SearchForCompanionViewModel
To the following:
private val searchForCompanionViewModel:

SearchForCompanionViewModel by viewModel()
This uses Koin to inject the lifecycle-aware ViewModel.
Finally, in the same Fragment, remove the following line from the
onCreateView since you no longer need it:
searchForCompanionViewModel =
ViewModelProviders.of(this)
.get(SearchForCompanionViewModel::class.java)
Now, remove this from setupSearchForCompanions:

searchForCompanionViewModel.petFinderService =
(activity as MainActivity).petFinderService!!
Run your app, and it’ll still work as it did before.
While your app is working correctly, run the tests. You’ll notice that most
of them are broken.
To fix them, open the FindCompanionInstrumentedTest.kt inside
androidTest and make the test class inherit from KoinTest. It’ll look like
this:
class FindCompanionInstrumentedTest : KoinTest {
Next, add the following method:
private fun loadKoinTestModules() {

loadKoinModules(module(override = true) {
single(name = PETFINDER_URL){server.url("").toString()}
}, appModule)
}
This is creating a function that loads the appModule you defined earlier
and an inline module that replaces urlsModule to reference the URL for
your MockWebServer.
In beforeTestRun(), add a call to stopKoin(), followed by

loadKoinTestModules(), after you launch the ActivityScenario. Your
changes will look like this:
@Before
// Insert them here!!
stopKoin()
loadKoinTestModules()
}
Since Koin starts as part of the app, this stops that instance of Koin, so
you can inject the test Koin modules, which is done in
loadKoinTestModules().
To finish up, add a call to stopKoin() as the second to last line of

afterTestRun:
@After
stopKoin()
}
Run your tests, and they’ll be green again.
Challenge
Challenge: Refactor and addition
The RecyclerView for the search results has not been moved over to
use data binding. Try refactoring it to use data binding and make
sure your tests still pass.
Try adding a new feature with an Espresso test and then refactor it.
Key points
Make sure your tests cover everything that you’re changing.
Sometimes, you’ll need to refactor your code to make it more

testable.
Some refactors require changes to your tests.
Refactor small parts of your app; do it in phases rather doing

everything all at once.
DI provides a cleaner way to add test dependencies.

Keep your tests green.
Move slow to go fast.

You’ve done a lot of work in this chapter to set yourself up to go fast.
Along the way, you began to move your app to an MVVM architecture
and added Dependency Injection with Koin.
TDD is a journey, but there are a lot of homeless coding companions and
pair-less developers counting on you. So, stay tuned for the next chapter,
where you’ll learn how to refactor your tests to start to go fast.
Chapter 15: Refactoring Your Tests
Sometimes you need to slow down to move fast. In development, that
means taking the time to write and refactor your tests so that you can go
fast with your testing. Right now your app is still fairly small, but the
shelters have big plans for it. There are lots of homeless companions and
pairless developers that need to be matched up! In the last chapter you
started with end-to-end UI tests, added some missing coverage, and then
refactored your code to made it easier to go fast.
End-to-end tests usually run in a simulator or on a device. Because of

that, they take longer to build, deploy, and run. In Chapter 4, "The
Testing Pyramid," you learned about how you should aim to have a
pyramid of tests, with your unit tests being the most numerous, followed
by your integration tests, and finally your end-to-end tests. Right now
you have an inverted pyramid where all of your tests are end-to-end.
As your app gets larger, this will slow down your development velocity
because a number of things happen, including:
Your Espresso tests will take longer and longer for the test suite to
run.
Tests that exercise one part of the app will often be exercising other
parts of the app as well. A change to these other parts can (and will)
break many tests that should not be related to what you are testing.
In this chapter you're going to break down your tests into integration and
unit-level. Along the way you will learn some tricks for mocking things
out, breaking things down, and even sharing tests between Espresso and
Robolectric. A lot of people are counting on you, so let's get started!
Note: In a normal development setting, it may be considered

premature optimization to refactor an app the size of your Coding
Companion Finder until it gets larger. That is a trade-off we needed
to make with this book. That said, there is an art to knowing when to
break things down. When you are new to TDD, it is easy to slip into a
rut of not testing enough and not breaking down your tests soon
enough. This is because testing is hard and it is easy to say it is not
worth the effort.
Until you get some experience with TDD, it is better to err on the
side of over-testing and over-optimization. As you get more familiar
with the tools and techniques you will be in a better place to make
that determination. That said, there will always be gray areas that
experienced TDDers will disagree on.
Source sets, Nitrogen and sharedTest

With androidx.test, Robolectric 4.0 and Project Nitrogen, which can be
found here (https://medium.com/androiddevelopers/write-once-run-
everywhere-tests-on-android-88adb2ba20c5), you have the ability to
write tests in Espresso and run them in either Robolectric on the JVM or
in an emulator/real device. One common use case is to run integration
and some end to end tests using the faster Robolectric while working on
your local machine. Then running the same tests using slower, but closer
to real life, Espresso during less frequent Continuous Integration cycles
to find potential issues on specific versions of Android.
Up to this point with your refactoring, you have been focusing on

running your tests in Espresso and putting them in androidTest. This is
how an Android project is configured out of the box. If you want to run
the same test in Robolectric you would need to move that test to the test
source set or create a new test.
This limitation negates that benefit of being able to run the same test in
Espresso and Robolectric (other than the shared syntax). This is a
shortcoming with the current default Android project setup. Luckily,
there is a way to get around this by using a shared source set.
To get started, open the starter project for this chapter or your final
project from the last one. Go to the app ‣ src directory. You will see three
directories there. androidTest, main and test. Delete test, and rename
androidTest to be sharedTest.
Next, open your app level build.gradle and add the following under your
android section:
android {
sourceSets {
String sharedTestDir = 'src/sharedTest/java'
String sharedResources = 'src/sharedTest/assets'
test {
resources.srcDirs += sharedResources
}
androidTest {
resources.srcDirs += sharedResources
}
}
}
This is creating a new source set that maps both your test and
androidTest to your sharedTest directory. It is also nesting an Android
directive under an Android directive so yours should look like this:
android {
.
.
.
android {
sourceSets {
.
.
.
}
}
.
.
.
}
Note: This may look familiar from the sharedTest set up you did in
Chapter 11, "User Interface."
Now, in your main androidTest

com.raywenderlich.codingcompanionfinder package open
CommonTestDataUtil.kt. In the first line of your readFile function get rid
of the /assets in this line:

.getResourceAsStream("/assets/$jsonFileName")
so that it looks like this:

.getResourceAsStream("/$jsonFileName")
Run your tests in Espresso (you might need to sync Gradle first) and they
will be green.
Note: If you find some of the tests are failing, check that
MainActivity.accessToken is set to your token you retrieved in
Chapter 13.
Now that you have your tests moved to a sharedTest source set, there are
a few things you need to do in order to get them working with
Robolectric.
First, open your app level build.gradle and add the following to the
dependencies section:
testImplementation 'androidx.test:runner:1.2.0'
testImplementation 'androidx.test.espresso:espresso-core:3.2.0'
testImplementation "androidx.test:rules:1.2.0"
testImplementation "androidx.test.ext:junit:1.1.1"
testImplementation "android.arch.navigation:navigation-testing:1.0.0-al
testImplementation 'com.squareup.okhttp3:mockwebserver:3.12.0'
testImplementation "androidx.test.espresso:espresso-contrib:3.2.0"
testImplementation 'org.koin:koin-test:1.0.1'
testImplementation 'org.robolectric:robolectric:4.3'
This is adding all of the dependencies that you had for your Espresso
tests at the unit level. It is also including the Robolectric dependencies
that you will need. Next, add the following to the top level android
section of the same file:
testOptions {
unitTests.includeAndroidResources = true
unitTests.returnDefaultValues = true
}
These are telling Robolectric to include Android resources. Because

Robolectric is not an actual emulator or device, many Android system
calls do not actually do anything. The unitTests.returnDefaultValues
makes them return a dummy default value in those instances, instead of
throwing an exception.
Now, go to your app component drop-down at the top of your IDE and
select Edit Configurations.
Select the + button.
Then, Android Junit.
You will be taken to a screen with a fresh configuration.

Under Use classpath or module select your app module.
Then under Test kind select Class.
Now, under the class select the ellipsis .... The following window will pop
up:
Select FindCompanionInstrumentedTest and press OK. Finally, it will
take you to the previous screen. Press OK on that to continue.
Your new test configuration will be highlighted. Go ahead and run it.
Oh no! Something is not right. If you look at the error messages you will
see the following (you may need to scroll down beyond the first couple of
errors):
Looking at your code, your ActivityScenario.launch is being called from

here with an Intent that is being passed in:
@Before
That Intent is set up in your companion object:
@BeforeClass
@JvmStatic
fun setup() {
server.start()
// It is being set right here!
startIntent = Intent(
ApplicationProvider.getApplicationContext(),
MainActivity::class.java)
startIntent.putExtra(MainActivity.PETFINDER_URI,
server.url("").toString())
}
When running Robolectric this doesn't get called before the @Before setup
function. More importantly, this Intent was initially set up to pass in
your mockwebserver URL when running your tests. In the last chapter
you refactored things so that this is not needed anymore, so let's get rid
of it.
To do that, get rid of the last two lines in that function so that it looks
like this:
@BeforeClass
@JvmStatic
fun setup() {
server.start()
}
Then, change the call on the first line of beforeTestRun from:
@Before
To:
@Before
testScenario =
Now run your tests again.
Things are looking better but you still have some failing tests (or perhaps
not!).
Note: depending on the speed of your machine or resources, you

may end up with none, two, or three failing tests. But even if they all
pass for you, there's something wrong here that you should fix.
These are failing with the same error message. At this point, before
reading further, a good exercise is to trace through things to see if you
can figure out what is going wrong here.
If you trace through this you will see that there are two tests that fail
when they try to click on an element with text that contains KEVIN,
which is the last line of the following function:
private fun find_and_select_kevin_in_30318() {

.perform(click())
}
It would appear that data from your mockWebServer is not being loaded.
The odd thing is that if you look at this test...
@Test
.perform(click())
}
It is able to load up the data and works correctly on some machines but
fails on others. You may experience either of these scenarios. This is
something that can cause a lot of frustration. Some tests are working
correctly, other similar ones that should are not — despite the tests
running correctly on Espresso. The problem has to do with how
Robolectric handles threads. Unlike when you are running tests on an
emulator or device, Robolectric shares a single thread for UI operations
and test code.
More importantly, by default, operations run synchronously using this

looper which means that many operations will not happen in the same
order that they would occur on a live device. This has been an issue with
Robolectric for a while, but luckily they've created a fix for it by adding a
@LooperMode(LooperMode.Mode.PAUSED) annotation before your test class.
Add it to the beginning of our test class so that it looks like following:
import org.robolectric.annotation.LooperMode
@LooperMode(LooperMode.Mode.PAUSED)
class FindCompanionInstrumentedTest: KoinTest {
Now run your tests again and all of them will pass.
Note: Your can find out more about this PAUSED Robolectric
LooperMode at http://robolectric.org/blog/2019/06/04/paused-
looper/.
Testing fragments in isolation

Until now your tests have been large end-to-end UI tests. That said,
some of your test cases are actually testing one component that could be
tested in isolation. A good example of that is your ViewCompanionFragment.
This fragment is called via your SearchForCompanionFragment. This happens
after you have searched for a companion and select one to see more
details.
When you refactored this fragment in the last chapter, you modified it so
that all of the data it needs to display, contained in an Animal object, is
passed into it via navigation parameters. That data is then passed into a
ViewModel which binds these attributes to the view.
Your end-to-end test is currently testing all of this, but the shelters have
a lot of changes that they are going to want to make to this page, along
with the SearchForCompanionFragment and other fragments in your end-to-
end test chain. This will begin to make your end-to-end tests fragile, so
now is a good time to move this to a more focused test.
To get started, open up your app level build.gradle and add the following
to your dependencies:
androidTestImplementation "org.robolectric:annotations:4.3"
// Once https://issuetracker.google.com/127986458 is fixed this can be t
// fragmentscenario testing
debugImplementation 'androidx.fragment:fragment-testing:1.1.0-beta01'
debugImplementation "androidx.test:core:1.2.0"
This is adding the AndroidX fragment testing dependencies.
Note: At the time of this writing there is a known issue with this
module. This means that you need to include it as part of your
implementation. To prevent this from going to production you are
limiting this to a debug build.
It is also adding the Robolectric annotations to

androidTestImplementation to ensure that your Looper annotation does
not cause compile issues with the new test you are about to add.
Now, go to your androidTest folder and under the

com.raywenderlich.codingcompanionfinder package add a class called
ViewCompanionTest. Above the class add the following:
This is telling it to run with AndroidJUnit4 and sets the LooperMode for the
Robolectric runs. Inside of your class add the following:
@Before
// 1
val animal = Animal(
22,
Contact(
phone = "404-867-5309",
email = "[email protected]",
address = Address(
"",
"",
"Atlanta",
"GA",
"30303",
"USA"
)
),
"5",
"small",
arrayListOf(),
Breeds("shih tzu", "", false, false),
"Spike",
"male",
"A sweet little guy with spikey teeth!"
)
// 2
val bundle = ViewCompanionFragmentArgs(animal).toBundle()
// 3
launchFragmentInContainer<ViewCompanionFragment>(bundle,
R.style.AppTheme)
}
1. Creating a test Animal object.
2. Creating a Bundle with your animal object.
3. Launching your fragment with the bundle that you just created.
Two and three might seem like a bit of magic, so let's break that down.
The SafeArgs that are used to pass arguments to your fragment through
the Jetpack Navigation components are doing some things under the
hood for you (see the previous chapter for some description on SafeArgs).
In your CompanionViewHolder, you have the following setupClickEvent
method:
private fun setupClickEvent(animal: Animal){
val action = SearchForCompanionFragmentDirections
.actionSearchForCompanionFragmentToViewCompanion(animal)
view.findNavController().navigate(action)
}
}
If you trace into the generated

actionSearchForCompanionFragmentToViewCompanion function you will see
that it is part of the following:
class SearchForCompanionFragmentDirections
private constructor() {
// 2
private data class
ActionSearchForCompanionFragmentToViewCompanion(
val animal: Animal
) : NavDirections {
override fun getActionId(): Int =
R.id.action_searchForCompanionFragment_to_viewCompanion
// 3
@Suppress("CAST_NEVER_SUCCEEDS")
override fun getArguments(): Bundle {
val result = Bundle()
if (Parcelable::class.java
.isAssignableFrom(Animal::class.java)) {
result.putParcelable("animal",
this.animal as Parcelable)
} else if (Serializable::class.java
.isAssignableFrom(Animal::class.java)) {
result.putSerializable("animal",
this.animal as Serializable)
} else {
throw UnsupportedOperationException(
Animal::class.java.name +
" must implement Parcelable or Serializable or" +
" must be an Enum.")
}
return result
}
}
companion object {
// 1
fun actionSearchForCompanionFragmentToViewCompanion(
animal: Animal
): NavDirections =
ActionSearchForCompanionFragmentToViewCompanion(animal)
}
}
This is doing the following things:
1. Calling the private constructor for the class.
2. Creating the new instance of the class.

3. When the navigation call is made it calls the getArguments function
which serializes the arguments, puts them in the bundle and returns
the bundle.
Your ViewCompanionFragmentArgs generated class provides methods to

deserialize and serialize your arguments as well which you can see if you
trace into them. This is what is called behind the scenes by Jetpack
Navigation when you add by navArgs() to an attribute definition in your
fragment.
At the time of this writing, Jetpack Navigation does not have testing
hooks for this scenario. Because of that we needed to understand what
this was doing behind the scenes to create this test. While that created a
bit of extra short term work, in the long term it gives you more
understanding about the framework.
By having a better understanding, when issues pop up while you are

wiring up your navigation safe arguments, you will know how it works
and be able to better understand how to trace it. Ultimately, this will
make it easier and faster for you to fix issues surrounding navigation.
Now that you have that out of the way, add the following test to your
ViewCompanionTest class:
@Test
fun check_that_all_values_display_correctly() {
onView(withText("Spike")).check(matches(isDisplayed()))
onView(withText("Atlanta, GA")).check(matches(isDisplayed()))
onView(withText("shih tzu")).check(matches(isDisplayed()))
onView(withText("5")).check(matches(isDisplayed()))
onView(withText("male")).check(matches(isDisplayed()))
onView(withText("small")).check(matches(isDisplayed()))
onView(withText("A sweet little guy with spikey teeth!"))
onView(withText("404-867-5309")).check(matches(isDisplayed()))
}
Even though this is a more focused test, it still will be run in Espresso. To
reduce test execution time you are verifying all of the expected display
fields in one test instead of breaking that up. Run the test in Espresso
and it will pass.
Finally, following the instructions at the beginning of this chapter, create
an Android JUnit configuration for your ViewCompanionTest and run it to
execute this test in Robolectric. This will also pass.
Now that you have your ViewCompanionFragment test more focused, let's
refactor your SearchForCompanionFragment tests.
Reviewing from the previous chapter, this fragment does the following:
1. It presents the user with a screen to search for a companion.
1. It gets the user's input and performs a search.

1. Presents the search results and allows navigation to the
To get started, create a new file in your test package called
SearchForCompanionTest.kt. Next, create the following class definition:
class SearchForCompanionTest : KoinTest {
private val idlingResource = SimpleIdlingResource()

}
This is inheriting from KoinTest like you did with

FindCompanionInstrumentedTest, adding in the LooperMode for Robolectric
and setting up your IdlingResource. Now, add in the following to the body
of your class:
companion object {
): MockResponse {
}
}
@BeforeClass
@JvmStatic
fun setup() {
server.start()
}
}
private fun loadKoinTestModules(serverUrl: String) {

single<String>(name = PETFINDER_URL) { serverUrl }
}, appModule)
}
@Subscribe
idlingResource.incrementBy(idlingEntity.incrementValue)
}
This methods are the same as the ones with the same name in your
FindCompanionInstrumentedTest.
Normally, at this point, you might want to consider refactoring this into
a shared component (although bear in mind the Three Strikes Rule,
which you can read about here: https://wiki.c2.com/?
ThreeStrikesAndYouRefactor), but there are some things that may
change so you are going to hold off on that. Following that, add in the
following methods:
@Before
launchFragmentInContainer<SearchForCompanionFragment>(
themeResId = R.style.AppTheme,
factory = object : FragmentFactory() {
override fun instantiate(
classLoader: ClassLoader,
className: String
): Fragment {
stopKoin()
GlobalScope.async {
val serverUrl = server.url("").toString()
loadKoinTestModules(serverUrl)
}.start()
return super.instantiate(classLoader, className)

}
})
}
@After
stopKoin()
}
This is launching your fragment, passing in a FragmentFactory. In your

ViewCompanionTest you did not need a FragmentFactory, because you were
launching your MainActivity. The reason you are using a factory here has
to do with Koin. Your ViewCompanionTest did not need to set Koin up and
FindCompanionInstrumentedTest was able to stop Koin and inject your test
modules after the app started. That only worked for those tests because
you are not testing anything on the Featured Companion page rendered
by your RandomCompanion fragment. Because you set up the Koin
dependencies before you instantiated a SearchForCompanionFragment your
test Koin modules were injected.
With your refactored tests you are directly loading your

SearchForCompanionFragment in a test activity. When that starts up, it is
loading up your app-level Koin dependencies. If you want to change
them over to your test modules by stopping Koin and loading your test
modules after things have been injected into your fragment, there is not
an easy way to do that. To solve this problem, you are passing in a factory
that stops Koin itself, and initializes it with your test dependencies
before instantiating your fragment so that you hit your MockWebServer
when making API requests.
Beyond this, your code is setting up your IdlingResource before your tests
run and tearing them down afterwards. Now, add the following test:
@Test
fun pressing_the_find_bottom_menu_item_takes_the_user_to_the_find_page()
}
This is the same test that you have in your FindCompanionInstrumentedTest

except it doesn't have to navigate to your SearchForCompanionFragment
since it is being directly instantiated for this test.
Run the test and it will fail.
Looking at the error message you will see a message that reads
java.lang.RuntimeException: java.lang.ClassCastException:
androidx.fragment.app.testing.FragmentScenario$EmptyFragmentActivity
cannot be cast to com.raywenderlich.codingcompanionfinder.MainActivity.
Looking at your stack trace, the following line in
SearchForCompanionFragment is your problem:
searchForCompanionViewModel.accessToken = (activity as
MainActivity).accessToken
For now, remove this line. It is setting a stored access token that was
being cached. For the eagle-eyed readers: this will result in extra
requests without tokens being made. Later on there will be an exercise
where you can fix this. Next, open up your SearchForCompanionViewModel
and change the following line close to the top of the class from:
lateinit var accessToken: String
To:
var accessToken: String = ""
Run the test via Espresso and it will be green.
Now use your Edit Configurations... to allow all tests in this class to run
in Robolectric and execute all of its tests. They will also be green.
Next, add the following two tests:

@Test
fun searching_for_a_companion_in_90210_returns_no_results() {
onView(withId(R.id.noResults)).check(
matches(
withEffectiveVisibility(
Visibility.VISIBLE
)
)
)
}
@Test
fun searching_for_a_companion_in_a_call_returns_an_error_displays_no_re
onView(withId(R.id.searchFieldText)).perform(typeText("dddd"))
.check(ViewAssertions.matches(isDisplayed()))
.check(ViewAssertions.matches(
withEffectiveVisibility(Visibility.VISIBLE)))
}
These are also the same as your similarly-named tests in

FindCompanionInstrumentedTest minus the navigation to your
SearchForCompanionFragment. Run all of your tests via Robolectric and they
will be green.
Following the process you used for creating a configuration for

Robolectric, create an Espresso one, but select Android Instrumented
Tests instead of Android JUnit. Run all of the tests in your class again
using Espresso and they will also be green.
Now, add the following test to assert that the values being displayed after
doing a search are correct:
@Test
}
Run all of your tests in Espresso and everything will be green.
Now execute your tests in Robolectric and you might see that there is a
problem.
Note: Like the tests earlier, these tests may not reliably fail, and may
even consistently pass on your machine.
Your test that is supposed to return two results is not. Before reading
further, add some debugging to your application to see what the problem
might be. A hint: look at your IdlingResource to see if it is behaving as
you would expect.
Robolectric and IdlingResource limitations

In theory, you should be able to run any Espresso tests on Robolectric
and have them run. The Google testing code lab at
https://codelabs.developers.google.com/codelabs/android-testing/#10
suggests this as does this talk at Google I/O 2019
https://www.youtube.com/watch?v=VJi2vmaQe6w&feature=youtu.be.
The reality is a lot more nuanced.
Earlier you learned about the @LooperMode annotation and how setting it
to PAUSED can make your Robolectric tests simulate how threads would
run on an emulator or device. That does help with a lot of tests but there
is a thing you are using in your tests that does not work with Robolectric.
In your case IdlingResource is a problem. At the time of this writing the
following issue talks about this
https://github.com/robolectric/robolectric/issues/4807.
Back in Chapter 13, "High-Level Testing With Espresso," you added

IdlingResource to deal with the slight delays happening when your tests
make calls to your MockWebServer.
The reason that MockWebServer causes some delays is because it runs in
a separate thread and takes requests through the network stack. Since
you can't use Roblectric and IdlingResoruce together reliably you are
going to need to refactor your test not to use the MockWebServer.
This is where your usage of Koin is going to start to pay some dividends.
If you look at your KoinModule in the main app package you will see the
following:
// 1
val urlsModule = module {
single(name = PETFINDER_URL) {
MainActivity.DEFAULT_PETFINDER_URL
}
}
val appModule = module {
// 2
logger.level = HttpLoggingInterceptor.Level.BODY
.build()
Retrofit.Builder()
.baseUrl(get(PETFINDER_URL) as String)
.client(client)
}
// 3
}
The following items are being injected when you run your test:
1. The URL for the Petfinder service. You override this in your test.
2. Your PetFinderService.
3. Your SearchForCompanionViewModel.
Currently your test has the following:

single<String>(name = PETFINDER_URL) { serverUrl }
}, appModule)
}
In order to get away from the network calls you are going to mock out
your PetFinderService. To get started replace that function with the
following:

// 1
val petFinderService =
Mockito.mock(PetFinderService::class.java)
// 2
Mockito.`when`(
petFinderService.getAnimals(
ArgumentMatchers.anyString(),
ArgumentMatchers.anyInt(),
ArgumentMatchers.contains("30318")
)
// 3
).thenReturn(GlobalScope.async {
getMockResponseWithResults()
})
petFinderService
}
})
This gets rid of your PETFINDER_URL Koin Single object and overrides
everything you need from appModule. The important thing here is the
mock of your PetFinderService. There are three parts to this mock:
1. First you create a mock of your PetFinderService.
2. Then you use the Mockito when function to have it look for an event.
In this scenario you are looking for a call to getAnimals on your mock
with any string for the access token, any integer for the limit and a
location string that contains the zipcode 30318.
3. When your when conditions are met, it returns a async co-routine

that will return a Response object containing your AnimalResults with
the getMockResonseWithResults() function.
getMockResponseWithResults() isn't yet implemented. Add that function

with the following body:
private fun getMockResponseWithResults(): Response<AnimalResult> {

val gson = Gson()
val animalResult =
gson.fromJson<AnimalResult>(readFile("search_30318.json"),
AnimalResult::class.java)
val responseMock =
Mockito.mock(Response::class.java) as Response<AnimalResult>
Mockito.`when`(responseMock.isSuccessful).thenReturn(true)
Mockito.`when`(responseMock.body()).thenReturn(animalResult)
return responseMock
}
Make sure to import retrofit2.Response, not the OkHttp one. Next, open
CommonTestDataUtil.kt in your main test package and take the private
modifier off the readFile function. Finally, run
searching_for_a_companion_in_30318_returns_two_results test in
Robolectric. It will be green.
Now run the same test using Espresso to make sure that you haven't
broken anything.
Oh no! Something is not right! Looking at the third line of your stack
trace you will see the following:
Mocking final classes with Espresso
Mockito has a limitation when running on an Android device or emulator
that prevents it from being able to mock classes that are final. When the
error above happens, though the message is not very descriptive, it can
mean that you are trying to mock a final class. In the function you
defined above you are mocking the Response class. Trace through to its
definition and you will see the following:
/** An HTTP response. */

public final class Response<T> {
As luck would have it, you can actually get rid of the mock and in the
process make the code a little simpler. To do that, replace your
getMockedResponseWithResults() function with the following:
private fun getMockResponseWithResults(): Response<AnimalResult> {

val gson = Gson()
val animalResult =
gson.fromJson<AnimalResult>(readFile("search_30318.json"),
return Response.success(animalResult)
}
Note: In general, if a class contains only data like this one, it is

generally as easy to make a real one that a mock. The benefit of a
real one is that you know there's no danger of you having mocked it
incorrectly!
This returns an actual Response object instead of a mock, and reduces the
size of your function by three lines. Run your test again and it will fail
with the same stack trace.
At this point you are only mocking out PetFinderService, which is an
interface. The problem here is that Koin includes a version of Mockito
which is old, and missing features. To do that, find this line in your app
level build.gradle:
androidTestImplementation 'org.koin:koin-test:1.0.1'
and replace it with:
androidTestImplementation("org.koin:koin-test:1.0.1")
{ exclude(group: "org.mockito") }
androidTestImplementation "org.mockito:mockito-android:2.28.2"
This excludes Mockito from koin-test and instead defines it as its own
deepndency. Run your test again and it will be green.
For now you are going to stick with mocking open classes in Espresso. At
some point you may find that you will need to mock classes that are not
open. One option is to make them open. But, if you would prefer to not
have to do that, this post on Medium shows a great alternative:
https://proandroiddev.com/mocking-androidtest-in-kotlin-
51f0a603d500.
If you try running all of your tests in Espresso you still have some that
are broken. Let's fix the mocks for them. To get started, add the following
functions:
private fun getMockResponseWithNoResults(): Response<AnimalResult> {

val gson = Gson()
val animalResult =
gson.fromJson<AnimalResult>("{\"animals\": []}",
return Response.success(animalResult)
}
private fun getMockResponseFailed(): Response<AnimalResult> {

val gson = Gson()
return Response.error(401,
Mockito.mock(ResponseBody::class.java))
}
The first one is returning a successful response with no results and the
second is returning a call with a 401 response. Because ResponseBody in
the second function is an abstract class you are able to mock it.
Next, add Mockito when clauses that use your new methods in your
loadKoinTestModules function so that it looks like the following:

val petFinderService =
Mockito.mock(PetFinderService::class.java)
Mockito.`when`(
)
getMockResponseWithResults()
})
// 1
Mockito.`when`(
)
getMockResponseWithNoResults()
})
// 2
Mockito.`when`(
ArgumentMatchers.contains("dddd")
)
getMockResponseFailed()
})
petFinderService
}
})
}
The following conditions were added:
1. When "90210" is entered as a location it returns an empty set of

results.
2. When "dddd" is entered as a location a 401 is returned.
Now run all of your tests in Espresso and Robolectric and all of them will
pass.
Breaking out unit tests

Up to this point your tests have had dependencies on Android. But, as we
discussed in Chapter 4, "The Testing Pyramid," you should strive to have
unit tests. Ideally, you will have more unit tests than integration tests
and more integration tests than end-to-end/UI tests.
Some scenarios where unit tests might make sense include testing:
Classes that focus on business logic.
ViewModels that have logic to present, retrieve, or store/post data.
Services that do things.
Classes that can be tested without needing to depend on Android.
Alternatively, in the following scenarios unit testing may not make

sense. These include tests that:
Focus on Activities and Fragments.
Depending on Android components to run.

Cover boiler plate getters and setters, for example in data objects.
Cover basic data marshaling that are covered at other levels of the
pyramid.
Require a significant amount of mocking.
Looking in your searchforcompanion package, there are only two classes

that are candidates for unit testing. They are your ViewCompanionViewModel
and SearchForCompanionViewModel. They are view models that can be tested
in isolation.
To get started open ViewCompanonViewModel.kt and you will see the

following:
data class ViewCompanionViewModel(

var name: String = "",
var breed: String = "",
var city: String = "",
var email: String = "",
var telephone: String = "",
var age: String = "",
var sex: String = "",
var size: String = "",
var title: String = "",
var description: String = ""
) : ViewModel() {
fun populateFromAnimal(animal: Animal) {

name = animal.name
breed = animal.breeds.primary
city = animal.contact.address.city + ", " +
animal.contact.address.state
email = animal.contact.email
telephone = animal.contact.phone
age = animal.age
sex = animal.gender
size = animal.size
title = "Meet " + animal.name
description = animal.description
}
}
The variable definitions that are part of your data class are not good
candidates for tests, but your populateFromAnimal() function could be
tested. To get started, create a new file called
ViewCompanionViewModelTest.kt in your test package. Next, add the
following content to it:
class ViewCompanionViewModelTest {
// 1
22,
Contact(
phone = "404-867-5309",
address = Address(
"",
"",
"Atlanta",
"GA",
"30303",
"USA"
) ),
"5",
"small",
arrayListOf(),
"Spike",
"male",
)
//2
@Test
fun populateFromAnimal_sets_the_animals_name_to_the_view_model(){
val viewCompanionViewModel = ViewCompanionViewModel()
// 3
assert(viewCompanionViewModel.name.equals("foo"))
}
}
This has the following parts:
1. An Animal object. This is the same data that you had in your
ViewCompanionTest animal object.
2. A test to make sure that the animals name is set when the user calls
the populateFromAnimal() function.
3. A focused failing assertion to start out with to ensure that we have a

valid test.
The test will default to run in Espresso. Following the steps from earlier
in this chapter, create a configuration to run this test class with Android
JUnit and then use it to run your test.
Now that you have a failing assertion, correct it to check for the name of
your animal:
assert(viewCompanionViewModel.name.equals("Spike"))
You may have noticed that your test here is much more focused with only
one assertion. This is intentional for a number of reasons including:
Unit tests are intended to be more focused.
They run faster and as such do not take as much time to spin up
dependencies.
The focused assertions lead to individual tests that are not as brittle.
Just like dancing, driving a car, or learning to speak a language,

repetition and practice help to make you a better TDD'er. We could spend
pages and pages writing focused tests for all of these fields. But, this is a
great opportunity for you to practice.
Before you continue on, take some time to do the following:
1. Write a test function for the next field in your ViewModel with one
assert that should fail.
2. Run the test to make sure that it fails.
3. Fix the assert expectation to ensure that it passes.

4. Go back to step one and repeat this until you have done this for all
fields.
Unit testing Retrofit calls

Now that you have your ViewCompanionViewModel under test, let's do the
same for your SearchForCompanionViewModel. To get started, create a new
Kotlin file in your test package called
SearchForCompanionViewModelTest.kt and add the following content to
it:
class SearchForCompanionViewModelTest {
Now open up SearchForCompanionViewModel.kt and you will see the

following:
class SearchForCompanionViewModel(
val petFinderService: PetFinderService
): ViewModel() {
// 1
val noResultsViewVisiblity : MutableLiveData<Int> =
MutableLiveData<Int>()
// 2
val companionLocation : MutableLiveData<String> =
MutableLiveData()
// 3
val animals: MutableLiveData<ArrayList<Animal>> =
MutableLiveData<ArrayList<Animal>>()
var accessToken: String = ""
// 4
accessToken,
location = companionLocation.value
)
animals.postValue(it.animals)
} else {
}
}
} else {
}
}
}
}
At a high level, this has the following testable elements:
1. The visibility for your noResults view.

2. The location that you want to search on.
3. The animals that are returned.
4. A Retrofit call that uses #2 to return data for #3 and either displays
or hides #1.
To start off, do a focused test that enters 30318 as your location and
checks to be sure that two results are returned.
// 1

// 2
): MockResponse {
}
}
// 3
@Before
fun setup() {
server.start()
.build()
.baseUrl(server.url("").toString())
.client(client)
}
// 4
@Test
fun call_to_searchForCompanions_gets_results() {
val searchForCompanionViewModel =
SearchForCompanionViewModel(petFinderService)
searchForCompanionViewModel.companionLocation.value = "30318"
Assert.assertEquals(2,
searchForCompanionViewModel.animals.value!!.size)
}
This test is doing the following:
1. Setting up your MockWebServer.

2. Setting up your MockWebServer dispatcher.
3. Initializing your PetfinderService pointing it to your MockWebServer.
4. Executes a test, which creates a SearchForCompanionViewModel with

your PetFinderService, sets your location value, runs the search, and
checks that the result has only two results.
This is not an Espresso test, but it may try to run as one due to your
sharedTest setup, so use the Edit Configuration option to set it up to run
as an Android Junit test. Then try running your test.
Oh no! Your test is failing with a dreaded NullPointerException when it

tries to set a value on your companionLocation LiveData object. I thought
Kotlin was supposed to help prevent null pointers! The actual null
pointer is thrown when it tries to determine if this task is running on the
main thread.
This is because your test is trying to access the "main" thread — which
does not exist in the unit test. To fix that you are going to need to add an
InstantTaskExecutorRule. This swaps the default executor for your
ViewModel with one that executes everything synchronously on your
current thread. To add this add the following in your app level
dependencies:
testImplementation "androidx.arch.core:core-testing:2.0.1"
androidTestImplementation "androidx.arch.core:core-testing:2.0.1"
Then add this to your test class:
@get:Rule
val instantTaskExecutorRule = InstantTaskExecutorRule()
You will also need to add this import:

import androidx.arch.core.executor.testing.InstantTaskExecutorRule
Now, run your test again.
Oops! It's still failing, but this time for a different reason! Let's track this
down.
Your error message is because your

is null. Looking at your
searchForCompanionViewModel.animals.value
method body of the searchForCompanions method in your ViewModel, there
are two co-routines that you are using.
.
.
.
.
.
.
.
}
}
}
If you debug through the call you will see that the test exits before your
call completes with your getAnimalsRequest. You are going to need to do
something to allow this to execute your threads and wait for it until
execution is done.
To get started, add the following dependencies to the dependencies

section of your app level build.gradle file:
def coroutinesVersion = "1.3.0-M2"
testImplementation "org.jetbrains.kotlinx:kotlinx-coroutines-android:$c
testImplementation "org.jetbrains.kotlinx:kotlinx-coroutines-test:$coro
androidTestImplementation "org.jetbrains.kotlinx:kotlinx-coroutines-and
androidTestImplementation "org.jetbrains.kotlinx:kotlinx-coroutines-tes
This is adding coroutine testing support, which you can read more about
here: https://github.com/Kotlin/kotlinx.coroutines. Next, add the
following at the class level of your test:
private val mainThreadSurrogate =

newSingleThreadContext("Mocked UI thread")
This is setting up a thread context, owned by your test thread that will
make all of your tests run under one thread.
Then add this to your setup method:
Dispatchers.setMain(mainThreadSurrogate)
This tells the system to use this new thread context that you just created.
Now run your tests.
Another problem?! The issue is that the LiveData result is coming back
after you did your assert. To fix that, replace your test with the following:
@Test
fun call_to_searchForCompanions_gets_results() {
// 1
val countDownLatch = CountDownLatch(1)
// 2
searchForCompanionViewModel.animals.observeForever {
countDownLatch.countDown()
}
// 3
countDownLatch.await(2, TimeUnit.SECONDS)
Assert.assertEquals(2,
searchForCompanionViewModel.animals.value!!.size)
}
This is adding a CountDownLatch that waits until your result comes back.
There are three parts to using it:
1. Setting up your latch with an initial latch value; in this case it is one.
The number is how many times countDown needs to be called on it
before it continues after await.
1. Using an observeForever on your LiveData object, and, when a result is

received, incrementing the value of the latch down.
2. A call to await with a timeout of 2 seconds to wait for the result to be

returned. The timeout is important so that the test does not hang
indefinitely if there is a problem that causes the latch to not fire.
Run your test again and it will be green!
Note: CountDownLatches are useful but can make tests slow and
brittle. An easy way to get around them often is to make the
scheduling/threading a dependency of the class you're testing, so
that you can put in "fake" synchronous scheduling within tests.
Since you want to verify that this is a valid test, change the expectation
on your assert to another value, such as 1 and re-run your test.
It fails, which is what we wanted. Now change the value back to 2 and
make it green.
When this ViewModel fetches data it sets values for your view, it also sets
the value of noResultsViewVisibility to INVISIBLE if there are results or
VISIBLE if there are none. Let's add some tests for that. To get started add
the following test:
@Test
fun call_to_searchForCompanions_with_results_sets_the_visibility_of_no_
searchForCompanionViewModel.noResultsViewVisiblity
.observeForever {
}
Assert.assertEquals(INVISIBLE,
searchForCompanionViewModel.noResultsViewVisiblity.value)
}
Since you want to have a failing test first, go to your

SearchForCompanionViewModel and change the following line in your
searchForCompanion function:
to:
noResultsViewVisiblity.postValue(VISIBLE)
Now run your test and it will fail.
Undo the last change in searchForCompanion so that the line is back to:
Run your test again and it will pass.
DRYing up your tests

Tests are code that you need to maintain, so let's write some more tests
for your SearchForCompanionViewModel and DRY (Do not repeat yourself)
them up along the way. To get started, add the following test:
@Test
fun call_to_searchForCompanions_with_no_results_sets_the_visibility_of_
.observeForever {
}
}
Because you want to have a failing test first, your assert is currently not
correct. Run the test and it will fail.
Now change the assert to be correct:

Assert.assertEquals(VISIBLE,
Looking at this test and the previous one you did around visibility they
are very similar:
@Test
.observeForever {
}
}
@Test
.observeForever {
}
}
The only real difference is your companionLocation value and your visibility
assert. Let's refactor this by replacing these two tests with the following:
fun callSearchForCompanionWithALocationAndWaitForVisibilityResult(locat
searchForCompanionViewModel.companionLocation.value = location
.observeForever {
}
return searchForCompanionViewModel
}
@Test
val searchForCompanionViewModel = callSearchForCompanionWithALocation
}
@Test
val searchForCompanionViewModel = callSearchForCompanionWithALocation
}
What you are doing here is using a common function for setting up your
call, CountDownLatch, but keeping your assert in the test. Now, technically,
you could have the assert in your common method and just pass in the
expected value to this common method. This is a matter of style. Since
part of the purpose of unit tests is to provide documentation about how
the code works, in the authors' opinion, not having the assert in the
common method makes it a little bit more readable. That said, if you find
it to be more readable by putting the assert in the common method, that
can be valid as well. The key takeaway is that tests are a form of
documentation and the goal is to structure them to make it easier for a
new person looking at the code base to understand it.
Challenge
Challenge: Test and edge cases
If you didn't finish out your test cases for your ViewCompanionViewModel
to test the other data elements, add tests following a red, green,
refactor pattern.
The tests you did for your SearchForCompanionViewModel missed a lot of

data validation and edge cases. Follow a red, green, refactor pattern
and try to cover all of these cases with very focused assertions.
Key points
Source sets help you to run Espresso tests in either Espresso or
Robolectric.
Not all Espresso tests will run in Robolectric, especially if you are
using Idling resources.
As you get your legacy app under test, start to isolate tests around
Fragments and other components.
ViewModels make it possible to move tests to a unit level.
Be mindful of mocking final classes.
It is possible to unit test Retrofit with MockWebServer.
Strive to practice Red, Green, Refactor.
As your tests get smaller, the number of assertions in each test

should as well.
Strive towards a balanced pyramid, but balance that against the

value that your tests are bringing to the project.
Test code is code to maintain, so don't forget to refactor it as well.
Move slow to go fast.

With this refactoring you have set your project up to go fast. It will help
many homeless companions, and companion-less developers get paired
up. That said, there are other tips and tricks to learn in future chapters.
For example, how do you deal with test data as your suite gets bigger?
How do you handle permissions? Stay tuned as we cover this in later
chapters!
Chapter 16: Strategies for Handling
Test Data
In the previous three chapters, you learned how to move slow to go fast.
Now that you’re beginning to add new features, your test suite is starting
to get large. Lots of homeless coding companions are being placed with
developers. But, as that’s happening, an inevitable problem is presenting
itself. Namely, your test data is starting to get difficult to maintain. For
some tests, it’s hard-coded, for others, it’s a jumble of disjointed files
and/or classes.
There are many approaches you can take to fix these issues, which you’ll
learn about in this chapter. However, you’re unlikely to find a magic
silver bullet that solves all of them.
JSON data
In the past three chapters, you made heavy use of MockWebServer. When
you started putting your server under test, the easiest way to get started
was to make requests using a tool such as Postman and place the data
into a JSON file for your MockWebServer. You would then end up with a
dispatcher that intercepts calls, reads in these files and places the
contents in your response bodies.
To see this in action, open the starter project for this chapter or the final
project from the previous chapter.
Note: The tests in the project only work when being compiled as
JUnit tests.
Look at the CommonTestDataUtil.kt helper class inside the com ▸

raywenderlich ▸ codingcompanionfinder test directory. Looking at your
dispatcher, you’ll see the following:
MockResponse().setResponseCode(200).setBody(
readFile("search_30318.json")
)
}
)
}
else -> {
}
}
}
In this example, you’re doing exact matches on the request path and
specifying files based on the request. Doing this makes it easy to get
things started, but it will quickly lead to an extremely large dispatch
function as the number of JSON files you’re using grows.
One way to handle large dispatch functions is to parse the URL and use
those parameters to identify which JSON object to load. For example,
with the Coding Companion Finder app, you might have a dispatch
method that looks like this:

val fileNameAndPath = getFileNameAndPath(request.path)
return if (!fileNameAndPath.isEmpty()) {
readFile(fileNameAndPath)
)
else -> {
}
}
}
getFileNameAndPath(path: String) converts the URL into a file name by

replacing characters.
For example, a path of /animals?limit=20&location=30318 becomes

animals_limit_20_location_30318.json, and /animals?
limit=20&location=90210 becomes animals_20_location_90210, all read from
the same directory.
Alternatively, your method could use a more sophisticated directory

structure. /big/animals?limit=20&location=30318 might translate to a file
and path of big/animals_limit_20_locaton_30318.json, and small/animals?
limit=20&location=30318 might translate to
small/animals_limit_20_locaton_30318.json.
Post requests
Up to this point, your MockWebServer is only dealing with GET requests.
It’s time to look at how to handle POST requests.
The PetFinder API used in your app requires OAuth credentials. POST
requests are used to support this OAuth workflow during certain steps
that — up to this point — you’ve short-circuited with your
MockWebServer. It's time to address this gap in your test coverage.
Since you need to understand how OAuth works to test it, let’s do a quick
review of the OAuth flow as you’re currently using it:
Open AuthorizationInterceptor.kt inside your apps retrofit package, and

you’ll see where these steps take place:
class AuthorizationInterceptor : Interceptor, KoinComponent {
private var token = Token()
// 1
override fun intercept(chain: Interceptor.Chain): Response {
var mainResponse = chain.proceed(chain.request())
val mainRequest = chain.request()
// 2
if ((mainResponse.code() == 401 ||
mainResponse.code() == 403) &&
!mainResponse.request().url().url()
.toString().contains("oauth2/token")) {
// 3
val tokenResponse = tokenRequest.execute()
if (tokenResponse.isSuccessful) {
// 4
tokenResponse.body()?.let {
token = it
// 5
val builder =
mainRequest.newBuilder().header("Authorization",
"Bearer " + it.accessToken)
.method(mainRequest.method(), mainRequest.body())
mainResponse = chain.proceed(builder.build())
} }
}
// 6
return mainResponse
}
}
Here’s how this code works:
1. First, a GET request is made to the server to retrieve some data on

the /animals endpoint. An accessToken is passed into it. When it’s the
first time a call has been made after opening the app, this
accessToken will be blank.
1. If the token is either blank, invalid or expired, the server responds
back with a 401, indicating that the current credentials are
unauthorized.
2. This 401 is caught by your apps AuthorizationInterceptor, which

posts an /oauth2/token request with your API Key and API Secret.
3. A token is returned to your AuthorizationInterceptor.
4. Your AuthorizationInterceptor then retries the original request to

/animals.
5. The response with animal records is passed to the original caller of
#1.
You’re going to add some non-focused test coverage. This non-focused

test coverage will cause your tests to fail if you break something.
Go back to CommonTestDataUtil.kt and replace dispatch(request:

RecordedRequest) with the following:
val headers = request.headers
// 1
if(request.method.equals("POST")){
if(request.path.equals("/oauth2/token")){
return MockResponse().setResponseCode(200).setBody(
"{\"access_token\":\"valid_token\"}")
}
}
// 2
val authorization = headers.values("Authorization")
if (!authorization.isEmpty() &&
authorization.get(0).equals("Bearer valid_token")) {
CommonTestDataUtil.readFile("search_30318.json")
)
}
}
else -> {
}
}
} else {
// 3
return MockResponse().setResponseCode(401).setBody("{}")
}
}
This code adds the following things to dispatch():
1. Logic to look for a post request to /oauth2/token and to return a

token value of valid_token.
2. Logic to check if a valid token is on your GET request before

processing it.
3. A return statement to return a 401 if the token is not valid_token.
Now, create a run configuration to execute all of your tests in

Robolectric. You first need to select Edit configurations... from the Run
configurations dropdown.
Then, click on the + button and select Android JUnit.
On the righthand side of the window, you’ll need to make some changes.
Unlike the previous chapter, you’ll run all of the tests in the test package,
so select All in package.
For your package choose, com.raywenderlich.codingcompanionfinder.
Finally, there’s a small red error in the bottom part of the window; it
demands that you select a module, so select app.
After you press OK, your new configuration is ready to go. Give it a spin
and execute all of your tests in Robolectric.
Oh, no! Most of the tests are running except for those in
SearchForCompanionViewModelTest. Before you changed your dispatcher,
they were green. If you trace into the stack trace for these, you’ll see the
following error:
lateinit property koinContext has not been initialized
This ends up tracing back to the following line in

AuthorizationInterceptor.kt (the one you looked at earlier):

If you look at where petFinderService is declared, around the beginning of

the class, you’ll see the following:
This is using Koin to inject a PetFinderService into this class.
Open SearchForCompanionViewModelTest.kt. You might remember from

the last chapter that you didn’t use Koin in this test. There are two
different ways you can fix this. One way is to configure this test to use
Koin. The other is to create a second dispatcher that doesn’t hit the
authorization interceptor in this test. Since the first option was exercised
in FindCompanionsInstrumentedTest.kt, you’re going to use option two.
Go to CommonTestDataUtil.kt and add the following function:
fun nonInterceptedDispatch(
): MockResponse? {
val headers = request.headers
)
}
}
else -> {
}
}
}
This code adds a version of your dispatcher that does not end up causing
your AuthorizationInterceptor in your app to be called.
Go back to SearchForCompanionViewModelTest.kt, and replace the

dispatcher declaration with the following:

): MockResponse {
return CommonTestDataUtil.nonInterceptedDispatch(request) ?:
}
}
This code calls your new dispatcher.
Run all of your tests again, and everything is green.

Using JSON data has the following benefits:
It makes it easier to quickly get data into your app to test.
It exercises a larger portion of your web request stack.
By taking snapshots from requests you make from tools like

Postman, you can get data that most closely resembles production
data.
Some drawbacks to using JSON include:
If your requests return large objects, the files can be cumbersome.
Modifying data in the files usually has to be done by hand or requires

a non-trivial effort to write code to read and modify it.
As your dataset grows, managing the files and the corresponding

logic can get complex.
Asserts generally need to be hard-coded values since you don’t have

a reference value from an object to perform an assert against.
Hard-coded data
Another approach you’ve used is hard-coded data. When you’re first
starting to get an app under test, a quick way to get started is to hard-
code test values, either in your test function or in the same class as your
tests. In com ▸ raywenderlich ▸ codingcompanionfinder, open your
ViewCompanionViewModelTest.kt and you’ll see the following:
// 1
22,
Contact(
phone = "404-867-5309",
address = Address(
"",
"",
"Atlanta",
"GA",
"30303",
"USA"
) ),
"5",
"small",
arrayListOf(),
"Spike",
"male",
)
@Test
fun populateFromAnimal_sets_the_animals_name_to_the_view_model(){
// 2
// 3
assert(viewCompanionViewModel.name.equals("Spike"))
}
}
This is a classic example of hard-coded test data. Here’s how it works:
1. Create an Animal using a series of hard-coded values.
2. Pass the object into populateFromAnimal(animal: Animal).

3. Perform an assertion based on the outcome of the function call.
Advantages to using hard-coded data include:
Test data is contained within the same class — sometimes the same
functions as your tests — making it easier to see what the values
should be.
It‘s faster to get started writing tests.
Disadvantages include:
The initialization calls need hard-coded values that create larger

test files, which reduces the readability of a test class as it gets
larger.
A lot of effort is duplicated across test files creating the data.
Large data sets become too verbose.
Test object libraries

One way to address the hard-coded data issues is to create a test object
library. This is best explained with code.
To get started, create a new package inside the com ▸ raywenderlich ▸

codingcompanionfinder test package, and name it data.
Inside the newly created data package, create a file named

AnimalData.kt, and add the following content to it:
object AnimalData {
val atlantaShihTzuNamedSpike = Animal(
22,
atlantaCodingShelter,
"5",
"small",
arrayListOf(),
shihTzu,
"Spike",
"male",
)
}
object AddressData {
val atlantaAddress = Address(
"",
"",
"Atlanta",
"GA",
"30303",
"USA"
)
}
object BreedsData {
val shihTzu = Breeds("shih tzu", "", false, false)
}
object ContactsData {
val atlantaCodingShelter = Contact(
phone = "404-867-5309",
address = atlantaAddress
)
}
When you import the required classes, make sure you import the correct
one for Address — the one in your project, not the one from the
framework packages. This creates test objects with the same data you
hard-coded in ViewCompanionViewModelTest.kt.
Next, open ViewCompanionViewModelTest.kt and replace its contents

with the following:
@Test
fun populateFromAnimal_sets_the_animals_name_to_the_view_model() {
viewCompanionViewModel
.populateFromAnimal(atlantaShihTzuNamedSpike)
assert(viewCompanionViewModel.name
.equals(atlantaShihTzuNamedSpike.name))
}
}
This code completely removes the hard-coded object creation, allowing

you to focus on the test and specific data it’s testing. As you test other
functions that need an Animal, you’ll have much more concise code and
shared objects that you can reuse.
Run your test, and you’ll notice it’s still green.
Green is good; green is wonderful in TDD, so take a moment to enjoy this

view!
Pros of using test object libraries include:
Test files that are more readable with shorter data set up code.
Test data that can easily be reused across test classes.
Cons of using them include:
It takes a little extra effort to set them up initially.
You have to look in another file to see details about the test data.
Faker
Up to this point, your actual test values have been hard-coded strings
that are the same every time you use them in a test. Over time, you may
run out of ideas for names. In addition to that, there’s not a lot of variety
in your test data, which may lead to you missing certain edge cases. But
don’t worry, there’s a tool to help you address this: Faker. You first saw
this in Chapter 10, "Testing the Network Layer."
Faker has data generators for various kinds of data, such as names and
addresses. To see the full power of it, you need to add it to your app. To
get started, open the app level build.gradle, and add the following to the
testImplementation 'com.github.javafaker:javafaker:0.18'
androidTestImplementation 'com.github.javafaker:javafaker:0.18'
This is adding the dependencies to your project. Next, open up your

AnimalData.kt file from above and add the following at the top level of
the file:
val faker = Faker()
This code gives you an instance of a Faker.
Next, add the following val to AnimalData:
val fakerAnimal = Animal(

faker.number().digits(3).toInt(),
fakerShelter,
faker.number().digit(),
faker.commerce().productName(),
arrayListOf(),
fakerBreed,
faker.name().name(),
faker.dog().gender(),
faker.chuckNorris().fact()
)
This code adds another animal object with Faker-generated data for
things like the companion's name and gender. (Yes, we added the Chuck
Norris fact for the description for fun.)
Next, add the following to AddressData:

val fakerAddress = Address(
faker.address().streetAddress(),
faker.address().secondaryAddress(),
faker.address().city(),
faker.address().state(),
faker.address().zipCode(),
faker.address().country()
)
This code adds Faker elements for the address.
Add this to BreedsData:
val fakerBreed = Breeds(faker.cat().breed(),

faker.dog().breed(), faker.bool().bool(), faker.bool().bool())
This code fakes out the breed.
Add this to ContactsData:
val fakerShelter = Contact(

faker.phoneNumber().cellPhone(),
faker.internet().emailAddress(),
fakerAddress
)
This code fakes out the contact information.
Now, go back to ViewCompanionViewModelTest.kt and add the following

test:
@Test
fun populateFromAnimal_sets_the_animals_description_to_the_view_model()
System.out.println(fakerAnimal.toString())
viewCompanionViewModel.populateFromAnimal(fakerAnimal)
assertEquals("faker", viewCompanionViewModel.description)
}
This code creates a test that does an assert on the description. It also
prints the contents of your object so that you can see what kind of data
Faker creates.
Run this test and you’ll see a failed test:

Your object output is on one line; however, breaking it down, so it’s more
readable, you’ll end up with something that looks like this:
Animal(
id=798,
contact=Contact(
phone=1-256-143-0873,
[email protected],
address=Address(
address1=09548 Wayne Dale,
address2=Suite 523,
city=Charitybury,
state=West Virginia,
postcode=30725-9938,
country=Northern Mariana Islands
)
),
age=0,
size=Synergistic Wool Bottle,
photos=[],
breeds=Breeds(
primary=Khao Manee,
secondary=Sealyham Terrier,
mixed=false,
unknown=false),
name=Miss Linnea Hills,
gender=female,
description=For Chuck Norris, NP-Hard = O(1).)
Of course, since you’re now using Faker, the information will be different
each time. At the moment, your assertion is not the right value, which is
reflected in the error message.
To fix this incorrect value, change the assert in your test on the last line
to:
assertEquals(fakerAnimal.description,
viewCompanionViewModel.description)
You’re now making sure that your assertion tests using the data
generated by Faker. You want to ensure that viewCompanionViewModel
contains the same description as the fakeAnimal, regardless of what Faker
generated there.
Run your tests again, and notice they’re all green.
The pros of using Faker include:
Random data that’s more representative of actual app data.
You don’t need to come up with test data yourself.
The cons of using Faker include:
While Faker has test data many scenarios, it doesn’t include all
possibilities.
Your assertions become assertions against values in Faker objects

rather than text you can read in your tests.
Sometimes, the data generated by Faker is outside of the scope of

your business logic, and you need to limit it.
Locally persisted data

If you have an app that locally persists data, you may have some tests
that need to have the datastore in a certain state before you run your
test.
If you’ve had experience doing server-side TDD, your first thought might
be to:
1. Get your datastore into the state you want it to be.
2. Dump it to a file.
3. Load up that data before running your tests.
On Android, at the time of this writing, the only way to do something

like this is for your app to have root access or to use adb, which is not
practical for Robolectric tests. This leaves you with only one option: To
load test data using code.
To load test data using code, you could, for example, use Faker with a
series of test objects you load using a helper function before running the
tests.
Because datastores tend to be slower, especially if you need to test with a

significant number of test records, you should consider executing that set
up using a @BeforeClass annotated function.
Key points
There are no magic silver bullets with test data.
JSON data is great for quickly getting started with end-to-end tests
but can become difficult to maintain as your test suites get larger.
Hard-coded data works well when your test suite is small, but it lacks
variety in test data as your test suite grows.
Faker makes it easier to generate a variety of test data for object

libraries.
Tests that need to get data stores into a certain state can be
expensive because you need to insert data into the data store
programmatically.

Wrangling your test data is another way that you’re able to move slow to
go fast. In your case, going fast means more companions in homes and
programmers with higher quality code. To learn more about Faker, check
out the project page at https://github.com/DiUS/java-faker.
Chapter 17: Continuous
Integration & Other
Related Tools
Before continuous integration became a
common practice, many software projects
followed a development flow where a number of
developers would work on their own feature for
a long period of time in a silo. Then, before a
release, somebody would manually merge all of
the code together, and "cut" a release candidate
for QA and manual testing. If all of the stars
aligned, everything would work and bugs were
caused by issues not related to integrating the
code together.
What actually happened on a lot of projects was

much uglier. Often there would be issues
merging the code together. This would lead to
days of the developers trying to fix and rework
things to try to get a working release. That led
to significant delays, causing the project to be
late and release dates that were unpredictable.
Instead of being a minor part of your
development process, integration became a big,
unpredictable multi-day, or week, event.
Up to this point you have learned a lot of

techniques for testing your application and have
been running those tests on your local
development machine. While doing that is much
more effective than not having tests, only doing
this has many drawbacks including:
Situations where the tests run on one

developers machine but not another.
Test suites that may take a long time to run

which, in effect, can block a developer from
working on another task.
No common build status that all developers

have visibility into.
Continuous integration helps you to address

this and prevent integration becoming large
event.
Continuous integration
fundamentals
Continuous integration or CI refers to the
practice of having a machine or environment
that is dedicated to the task of building the code
base. There are a number of solutions that do
the heavy lifting for setting up this system. You
will learn more about that later on in this
chapter. In a very simple CI setup you may have
a workflow with the following steps:
1. When a user pushes up a new some changes

to a Git repo, the CI server kicks off a new
build.
2. If the build is successful, an APK with a

unique build number is generated and
placed up on a common server for the team
to test, and a status page is updated with a
success status.
3. If the build is not successful a status page is

updated with a failed status and a message
is sent to the team notifying them of the
broken build.
There are many variations of this, but the

general idea is that whenever something new is
pushed up to some central repo, a clean build is
started to make sure that everything builds and
is working correctly. Many development teams
with no automated tests will use this approach.
On the next page is an example of a Jenkins

server status page that shows the status of
various project's builds.
The green check marks indicate the last build

was successful, the ! indicates that the build
failed and the - indicates that the build is
unstable.
When you have a suite of unit, integration and
end-to-end tests an extra step is added so that it
looks like this:
1. A build is kicked off on a commit to the

repo.
2. If the build is successful, all unit,

integration and end-to-end tests are run.
3. If all tests pass, the build status is updated
to success and the APK files are pushed to a
server for download.
4. If the tests or build fail, the status is

updated to failed and a message is sent to
team members notifying them of a failed
build.
By using CI and TDD together, every time a

commit is pushed, if the build ends up green you
can have a reasonable level of confidence that
this commit leaves the app in a state where it
could be shipped. Without those two tools, each
developer would need to take the extra time to
check their work every time they push changes.
If they neglect to do that, at some point the

project will run into a state where one developer
pushes something up that breaks the build that
is not caught for hours or days while subsequent
pushes are done. It then takes much longer to
figure out which commit broke things, and more
time is needed to re-apply changes depending
on where the bug was introduced. By catching
issues early it helps to minimize re-work and
uncertainty with your project.
Branches and CI
Many teams end up using a branching strategy
with their source control repository. One that is
commonly used is called Git Flow
https://nvie.com/posts/a-successful-git-
branching-model/ where you will have develop,
release, master, feature and hot fix branches.
If your entire test suite does not take a lot of

time to run, a best practice is to have CI run
whenever a new push is made to a branch.
But, if you have a large test suite that takes

several minutes to run you may need to take a
more sophisticated approach. To better
understand this consider a test suite that has
the following attributes:
Fifty percent of tests are unit tests that run

quickly on the JVM. The entire suite of unit
tests take one minute to execute.
Thirty five percent are integration tests that
can run in either Robolectric or Espresso.
The suite of integration tests take two
minutes to execute in Robolectric and
twenty minutes in Espresso.
Fifteen percent of the tests are end-to-end

that mostly only run in Robolectric. This
suite takes eight minutes to execute in
Espresso.
Running tests in Espresso versus Robolectric

give the most realistic test results, but they take
twenty nine minutes to execute. In this
scenario, running this test suite using Espresso
on your local development machine before
pushing up changes would not be a practical
thing to do. Most development teams facing this
will:
Run tests related to a feature they are

working on along with all unit and
Robolectric tests locally.
Push it up to source control and let CI run

the full test suite with Espresso.
With the CI test runs you have another issue.
When using a branching strategy, a good
practice is to have short running feature
branches. Ideally that branch is only worked on
for few hours, reviewed and merged them into
the development branch. Waiting for a half an
hour before merging feature branches would
slow down integration. Especially on a larger
team. There are multiple ways you could
address that, but like most things in
programming, there's trade-offs for each one.
Possible solutions include:
Only running Robolectric and unit tests via

CI on feature branches.
Using a less frequent schedule, IE: once or

twice a day, for running the full Espresso
suite from your development or feature
branches.
CI tools
CI solutions need to do a lot of things including:
Monitoring your source code repository for
changes.
Managing and orchestrating build agents

(computers or VMs that actually perform
builds).
Offering some sort of dashboard or

visualization of the build status.
When looking at CI tools there are two main

categories you will run across:
1. Self hosted solutions.
2. Cloud based solutions.
Self hosted CI
A large number of organizations, arguably a
majority of Android teams, have historically
used this approach. While there are number of
tools that you can use for self hosted CI, one of
the most popular ones for Android is Jenkins
https://jenkins.io/. Jenkins installations have
two main components:
1. A Build Server.
2. One, or multiple build executors.
The build server is what handles things such as:
The project configuration.
Build pipeline through checkout, build,

test, deploy, etc.
Coordinating build executors.
Build executors are the machines/VM's where

builds take place. Each executor that you use
needs to have the proper tools installed to be
able to compile and run your project. A build
executor can only run one pipeline at a time. So
let's say that you have one developer working on
feature branch A and another that is working on
feature branch B. Developer A pushes up a
commit that starts a CI pipeline. Before that
build is completed a commit is pushed up to
feature branch B. If your Jenkins system only
has one build executor, it will not start that
pipeline until the build executor is done
working an the pipeline for branch A. If the test
suite takes a while to run, or there are a lot of
developers on the project this could cause a CI
backlog. To address this you can add additional
build executors which will allow multiple branch
pipelines to run at the same time.
Advantages of these self hosted solutions

include:
More control over your build environment.
Less expensive than using a cloud based

solution, especially when running a lot of
builds.
The ability to add your own physical devices

to perform automated tests on.
Better security when working in a high

security environment or behind a corporate
firewall.
Significantly more work involved in

managing server instances.
Build executors need to be updated when
updating versions of Android tools.
Headless servers can provide extra

configuration challenges for Espresso based
tests.
Cloud based CI
Cloud based CI solutions move everything that
Jenkins provides to the cloud. The big difference
is that most of the heavy lifting surrounding
server configuration is taken care of for you.
Some of the more popular services include:
CircleCI
Bitrise
Travis CI
Each service has its own nuances, but in general

they have have different plans that go up in
price based on the concurrency and total system
build/run time that you need. All of these
services also have free plans for open source
software projects.
Advantages of these solutions include:
Significantly less effort is needed to

configure and manage server instances.
Updates for build tools are often taken care

of by the service.
Scaling for more concurrency can be done

by changing your plan.
Free plans for open source projects.
Generally, less control over your build

agents.
Limited support for running Espresso suites

on your own test devices.
No support for high security environments

behind firewalls.
A system needing a lot of concurrency and

build time can get expensive.
Device farms
When running Espresso tests, most CI solutions
only provide direct support for running your
tests on an emulator. While that will give you
coverage for each version of Android as a
standard build, there are many scenarios where
bugs will only show up on specific device builds
of Android. Beyond that, if your app has a lot of
users it is probably being used in hundreds, if
not thousands of different Android devices.
Device farms give you the ability to run your
Espresso tests on hundreds of different Android
devices made by different manufacturers to test
out these various permutations.
Depending on your needs, you can even opt to

run all of your CI Espresso tests in a device farm.
There are a number of device farm providers.
Two popular ones are:
1. Firebase Test Lab
2. AWS Device Farm

As is the case with cloud based CI offerings, the
more device time you use, the more the farms
will charge you.
CI strategy guidelines
Depending on the size of your project, user
base, code quality needs and size/type of your
test suite your CI strategy will change. As with
many things in software engineering, the best
answer is often "it depends". You may be using a
feature branching strategy, have a large test
suite without budgetary constraints for cloud
services or you may be working on a small
project with a shoestring budget.
Wait time
The most important consideration with the test
suites you run in CI is time. Ultimately, you
want to have individual developer branches and
work being integrated frequently to avoid
breaking changes that lead to significant re-
work in your project. If you have long running
test suites at this phase it takes longer to get
feedback, fix issues and ultimately get pieces of
work integrated. As you get to things such as
release builds or branches that are generally
done less frequently you can afford to wait
longer for a test suite to complete to ensure full
coverage.
Here are some guidelines based on branch

commit frequency:
One or more commits per hour - test suites

that run in less than five minutes.
One to two commits per day - test suites

that run in less than thirty minutes.
Releases or less than once per day - all test

suites.
Test service cost

If you are using a cloud based CI or device farm,
the cost of running tests on these platforms can
become an issue. Developer time is generally
your largest project cost. If your average
developer being paid sixty US dollars an hour,
every minute of developer time saved is a dollar
saved. That said, there may be instances where
y
you can reduce build times by throwing more
build agents as a problem, but not getting a lot
of value out of those extra resources.
An example of this might be where you have a

large, modularized application with a long
running Espresso test suite. One approach that
some teams have taken to reduce test execution
time is to "shard" (divide up) the test runs so
that they can be executed in parallel. Doing this
could allow you to run a test suite that takes 30
minutes in less than five minutes by breaking
the work up among multiple build agents as part
a complex build pipeline. That will reduce your
developer time, but if you are providing enough
resources to allow all of these tests to run for
feature branches it will also significantly
increase you CI service cost. This will give you
better coverage by executing more tests
frequently, but is that extra frequency providing
enough value to offset the cost?
Ultimately, you want to have all of your tests

run in CI before cutting QA and Release versions
of your app, but there may be times where the
cost to run some of the more expensive test
frequently may not be offset by the benefits.
Device coverage
If you are using a device farm that has hundreds
of different devices to execute your tests, how
do you determine which ones to execute your
tests on? In an ideal world, the answer would be,
all devices all of the time. Unfortunately, in
reality that would be a very expensive
proposition.
For example, lets take a look at the cost, at the

time of this book writing, to run an Espresso
test suite in Firebase Test Lab. This suite takes
six minutes to run on a device, and you want to
run it on one hundred different Android devices:
100 devices * 6 minutes = 600 device

minutes or 6 device hours
1 Test lab physical device hour costs $4.20

USD
6 device hours * $4.20 = $25.20 USD

If you average twenty work days in a month and
ran these tests once a day on work days your
device farm cost would be $504 USD. For a small
project that cost may not justify the benefit. If
your project could justify that, but you ran these
this for all daily commits it would become cost
prohibitive.
There are no magic bullets on how to address

this but here are some ideas to figure out the
best strategy:
Look at your applications analytics and only

choose the devices that have the most
users.
Run tests more frequently on devices that

have surfaced more device specific bugs.
Run frequently executed test runs on a

randomly selected device from a list.
Only run your suite on a large number of

devices infrequently, i.e. before cutting a
release candidate.
Key points
CI helps to ensure that you are frequently
integrating all developer's code to avoid
rework.
You should run CI on all project branches,

but may need to limit test suites run on
branches with frequent pushes to reduce
test execution time.
Many organizations use self hosted CI

solutions like Jenkins, but the cloud based
ones are usually easier to set up.
When scaling your CI you may need to

balance the cost of scaling against less
frequent test suite execution for expensive
Espresso test suites.
Device farms allow you to test on real

devices but can get very expensive.
There are no magic bullets when figuring

out the best CI strategy.
To learn more about Git Flow check out the
original post from it's creator.
https://nvie.com/posts/a-successful-git-
branching-model/
The Jenkins page has a lot of resources for

setting up your own Jenkins CI
environment. https://jenkins.io/
To learn how to set up CircleCI for an

Android project have a look at their
Android developer documents.
https://circleci.com/docs/2.0/language-
android/
To learn how to set up Bitrise for an

Android project have a look at their
Android developer documents.
https://devcenter.bitrise.io/getting-
started/getting-started-with-android-apps/
For Travis CI on Android, check out their

Android guide. https://docs.travis-
ci.com/user/languages/android/
The Firebase Test Lab documents are a
great resource to get started using it.
https://firebase.google.com/docs/test-lab
You can find out more about getting started

with the AWS device farm on their product
page. https://aws.amazon.com/device-farm/
Chapter 18: Testing Around Other
Components
Up to this point you have focused on testing functionality that is part of
your application and that you and your team have written. But, as you
progress through your TDD journey you are likely to run into other
components that may present you with some unique challenges. These
generally fall into one of three categories:
1. Testable: These are components that can be easily tested and/or

verified, so no problems here.
2. Mockable: Mockable components expose a boundary between your

application and the component. Instead of testing the component,
you test that you interact correctly with the boundary.
3. Untestable: Sometimes you will run across components that are

exceedingly difficult or impossible to test.
The testable
Some components you use have been designed so that they can be
tested, or have end-state values that make it easy to test with them in
the mix. For example, in Chapter 9, "Testing the Persistence Layer," you
learned about how to test the persistence layer in your application tests.
Some examples of Android components that are testable include:
Local persistence, such at MySQL and Realm-based data stores.
Libraries that manipulate the UI or other states of the system in a

repeatable manner, like DiffUtil or Redux-based mechanisms.
Libraries that have testing hooks built-in.
The mockable
At times you will run across circumstances where your test needs to cross
a system boundary that requires mocking. For example, in Chapter 10,
"Testing the Network Layer," the MockWebServer you are using is mocking
out your okhttp calls that are made to a server. In this case, the mocking
was taken care of for you. In other instances you will need to mock out
system boundaries manually.
Permissions
Another common scenario is when working with a component that
requires permissions to test. As a quick review, there are three types of
Android permissions:
1. Normal Permissions: These only require a declaration in the

manifest to grant the user permission.
2. Signature Permissions: The system grants these permissions at

install time, but only when the app that is using it is signed by the
same app that grants the permission.
3. Dangerous Permissions: These are permissions that are accessing

things such private user data, etc. In order to request these you need
to have the permission specified in the app manifest and prompt the
user for it at run time.
If you are testing functionality that requires normal permissions, there is

nothing you will need to do other than include them in your app's
manifest. Signature permissions follow a similar pattern. Things get a
little bit more involved when working with dangerous permissions. To
understand this, let's look at an example that uses the CALL_PHONE
permission.
To get started import the starter Coding Companion project. Once it is

imported, add your device keys you generated in Chapter 13, "High-Level
Testing With Espresso," and run the app.
Tap on Find Companion, enter a location, tap on Find followed by
selecting a companion.
Now, tap on the phone number for the companion, select ALLOW to
allow the permission request and then a call will begin to the shelter.
Open ViewCompanionFragment.kt and look at the
setupPhoneNumberOnClick() function:
private fun setupPhonNumberOnClick(){

fragmentViewCompanionBinding.telephone.setOnClickListener {
// 1
Dexter.withActivity(activity)
.withPermission(Manifest.permission.CALL_PHONE)
.withListener(
object : PermissionListener {
// 3
override fun onPermissionGranted(
response: PermissionGrantedResponse
) {/* ... */
val intent = Intent(Intent.ACTION_CALL, Uri.parse(
"tel:" + viewCompanionViewModel.telephone))
startActivity(intent)
}
override fun onPermissionDenied(

response: PermissionDeniedResponse
) {/* ... */}
// 2
override fun onPermissionRationaleShouldBeShown(
permission: PermissionRequest,
token: PermissionToken
) {/* ... */
token.continuePermissionRequest()
}
}
)
.onSameThread()
.check()
}
}
This is doing three things:
1. It uses the Dexter library to check for the CALL_PHONE permission.
2. If permission has not been granted it shows a dialog asking for it.
3. If permission is granted it starts a ACTION_CALL intent to call the

number you tapped on.
Note: To learn more about the Dexter library you can visit
https://github.com/Karumi/Dexter.
To test this scenario you are going to need to do two things:
1. Grant CALL_PHONE permissions in your test.
2. Do an assert on the intent.
To get started go to FindCompanionsInstrumentedTest.kt and add the

following to the beginning of your class:
@get:Rule
val grantPermissionRule: GrantPermissionRule =
GrantPermissionRule.grant(
android.Manifest.permission.CALL_PHONE)
This is granting the CALL_PHONE permission for all of the tests in your
class. Next add the following to your app level build.gradle in your
androidTestImplementation 'androidx.test.espresso:espresso-intents:3.2.
This adds in support for Espresso Intents, which allows you to assert that
an intent is sent. You can learn more about Espresso Intents at
https://developer.android.com/training/testing/espresso/intents.
Now add the following to the beginning of your

FindCompanionsInstrumentedTest:
@get:Rule
val intentsTestRule = IntentsTestRule(MainActivity::class.java)
This adds a rule to use Espresso Intents in your test. Now add the
following test:
@Test
fun verify_that_tapping_on_phone_number_dials_phone() {
// 1
val intent = Intent()
val result =
Instrumentation.ActivityResult(Activity.RESULT_OK, intent)
intending(
allOf(
hasAction(Intent.ACTION_CALL)
)
).respondWith(result)
// 2
onView(withText("(706) 236-4537")).perform(click())
// 3
intended(allOf(hasAction(Intent.ACTION_CALL),
hasData("tel:(706) 236-4537")))
}
1. Creates a mock Intent that will be called when the ACTION_CALL intent
is fired in your app.
2. Navigates to a search result and clicks on the phone number.
3. Asserts that the ACTION_CALL intent has been sent.
Run your test and it will now be green:
But, if you try running all of your tests you will have a failure:
If you dig into this, the root cause is that your IntentsTestRule is causing
your MockWebServer to crash. Luckily your ViewCompanionTest does not
depend on MockWebServer, and this test should really be there anyway.
To fix things, remove all of the changes you just made to

FindCompanionsInstrumentedTest. Next, open your ViewCompanionTest and
add the following to the class:
@get:Rule
val grantPermissionRule: GrantPermissionRule =
GrantPermissionRule.grant(
android.Manifest.permission.CALL_PHONE)
@get:Rule
val intentsTestRule = IntentsTestRule(MainActivity::class.java)
This is adding in the GrantPermissionRule and IntentsTestRule that you

previously added to FindCompanionsInstrumentedTest. Next add the
following test to ViewCompanionTest:
@Test
fun verify_that_tapping_on_phone_number_dials_phone() {
// 1
val intent = Intent()
val result =
Instrumentation.ActivityResult(Activity.RESULT_OK, intent)
Intents.intending(
CoreMatchers.allOf(
IntentMatchers.hasAction(Intent.ACTION_CALL)
)
).respondWith(result)
// 2
.perform(ViewActions.click())
// 3
Intents.intended(
CoreMatchers.allOf(
IntentMatchers.hasAction(Intent.ACTION_CALL),
IntentMatchers.hasData("tel:(706) 236-4537")
)
)
}
This is adding the same test you had before without the call to navigate
to this fragment since you are testing just the fragment in isolation. The
tests in this file also do not depend on MockWebServer. Run all of the tests
in it and they will be green.
Other mockable components

There are many classes of external components where your best testing
strategy will be to mock out your interaction with them. Some good
candidates for this include:
Glide
Picasso
Video Players
Intents to other applications etc.
Network calls
Interactions with sensors and hardware
The untestable
There are some components where the best TDD option is to not test it.
Determining that the component is untestable can be tricky. TDD is hard.
On one hand you don't want to give up too soon on testing something.
On the other hand you don't want to spend too much time trying to test
the untestable.
Some traits that may make a component untestable include:
It draws things on a graphical Canvas.
There is not a testable end state after interacting with it.

No test extensions or mechanisms are provided by the library
author.
The component is implemented primarily though the NDK.
The library is popular, but a Google search doesn't turn up any

instructions on testing it.
You are making system calls to get device traits.
Let's look at some examples to better understand the thought process

that goes into this. The names have been changed to protect the
innocent, but these are actual scenarios that the author encountered in
projects.
Google maps Android SDK

Google Maps Android SDK
https://developers.google.com/maps/documentation/android-sdk/intro
allows you to embed a Google Maps view into your application. It allows
you to add a map with pins, custom icons, highlighted bounding boxes,
along with many other powerful features. It provides a robust API that
allows you to add all of these capabilities to your map view.
If you dive in to a view generated by it, the elements on the map are not
in a structure where you can verify that a component is in a specific
location because it a basically a canvas view. If you do a Google search
the only solutions you will find for testing this are using a UI automator
to click on a specific position on a pre-determined map. That test will
work fine when you have a specific screen size, but if you are supporting
multiple screen sizes and DPIs, the coordinates you use for one screen
size may not work for another.
Your best bet when testing with this SDK is to:
Test callbacks made by your map — i.e. when you click on a point on
the map.
Test the data and items being used to create things on the map.
Rely on manual testing for the map functionality.
These will also work well for similiar libraries which are hard to test.
System setting API calls

Imagine you have an app that needs to get the values for some system
attributes on your device such as the device's IP address, SIM card
provider and current GPS location. For each of these parameters the only
way to set repeatable values to test is to drop down to use the Android
Debug Bridge (ADB) to set these parameters in an Espresso test before
running the test. While this can be done in unit tests, it can be a
problematic solution for multiple reasons including:
In order to give the system time to apply the setting you will need to
add Thread.sleep() calls which will cause much longer test execution
times, and also might result in unreliable tests.
Depending on the version of Android, some settings may not be

available via ADB or be very difficult to get at.
Some settings may only be testable on real devices, not emulators.
With enough work you might be able to create reliable, repeatable tests
for these. In many cases it might be better to abstract that code into a
simple library (sometimes called a shim) that is manually tested. In your
unit tests you could then use Dependency Injection to mock this shim
you created to test the code that depends on it.
Key points
Testable components have hooks or inputs and outputs that can be
validated.
Many components are best tested by mocking your interaction with

them.
It is possible to test dangerous permissions.
There are some components that are untestable.

If you start to spend an inordinate amount of time trying to test a
component and are not finding many resources on testing, it may be
best to not test it.

For a tutorial on how to get started with the Google Maps SDK go to
https://www.raywenderlich.com/230-introduction-to-google-maps-
api-for-android-with-kotlin
To learn more about the GrantPermissionRule go to

https://developer.android.com/reference/android/support/test/rule/
GrantPermissionRule
To learn more about Espresso Intents check out

https://developer.android.com/training/testing/espresso/intents
If your curious about permissions check out

https://developer.android.com/guide/topics/permissions/overview
Appendix: Other Related
Techniques
TDD focuses on writing tests for classes and methods before writing code
to support the app’s features. While this approach might be useful for the
developers, some argue that it doesn’t consider the end user since they
tend to focus on the expected behavior of the app or its use cases.
Writing tests for functionality may even be easier to validate with the
users.
After exploring TDD and reading this chapter, you’ll discover that other
techniques and frameworks exist to enhance or complement TDD.
Acceptance test-driven development

In Acceptance test-driven development (ATDD) the entire team first
defines what the acceptance criteria of a feature is before development
on that feature starts.
Here’s how it works: The team collaborates with the end users. Based on
that collaboration, the team writes one or more acceptance tests for each
requirement. After that, the development team writes the code needed
for the requirement to pass those acceptance tests.
ATDD focuses on the acceptance criteria agreed upon with the users. If
these tests pass, you can assume the users will be happy with your app.
This process gives you an idea of the feature development progress and
also avoids gold-plating, which means to over-develop features or add
functionality that end users don’t need or want.
With ATDD, you can take each requirement and write it in the user story
format. Once you have that, you can write the acceptance tests for each
one. That’s why, sometimes, this technique is also called Story Test
Driven Development (STDD).
Here’s an example template you might use to write a user story:

As a <role> I can <capability>, so that <receive benefit>.
Once you have that, you can turn it into this:
As a seller, I can publish an item so that I can sell it.

As a buyer, I can search for an item so that I can purchase it.
You can write user acceptance tests (UATs), also known as Story Tests, in
a plain document or even a spreadsheet. Because these tests are not
technical, any non-technical person can read and modify them. Once the
team is satisfied, developers can use these tests as input to translate into
test code. For example, in an app that has a login feature:
1. Open the application.
2. Log in with invalid <username> and <password>.
3. Verify an error message is shown: “Invalid <username> and/or

<password>!”
Or you may have a spreadsheet with inputs and expected output, for
example, in a calculator app:
Number Operator Number Expected

1 + 2 3
2 * 0 0
These story tests provide the following advantages:
You can use them to communicate between all of the roles of the
team.
They’re easier to agree with the final user.
The examples are concrete.
Analysts, developers, testers and the end users have a more active
role. They’re more involved and they all work together in creating
possible scenarios. They understand the stories, the examples and
the acceptance criteria.
Behavior-driven development
Behavior Driven Development (BDD) is an extension of TDD. You might
think of BDD as an enhanced version of TDD.
BDD explains that behaviors should drive the development of an app’s

features. The behaviors are explained with examples in a simple
language that’s understandable by all of the people involved in the app.
That means that developers, testers and analysts can all collaborate in
BDD. Usually, BDD is used as a tool that allows creating tests in plain
English and allows collaboration between technical and non-technical
members of the team.
The main focus of TDD is to write tests first and to think about designing
the features. BDD not only does that, but also uses a natural language to
describe the tests. To begin with, it pays attention to tests method
naming. For example, the name of the methods could start with should,
or use the given/when/then format instead of the traditional naming,
which starts with test. This method naming was selected to focus and
better describe a behavior. The idea is to break down a test scenario into
three parts:
given: Describes the state of the world before you begin the behavior
you’re specifying in this scenario. These are the pre-conditions to
the test.
when: You specify the behavior.
then: Here you describe the expected changes due to the behavior.
Instead of writing a test class per class of a feature, BDD suggests writing
a class per requirement. It also enforces writing code that uses the same
business language, so there’s a ubiquitous language. This means that
analysts and developers use the same language. For example, if you’re
writing a game where the analyst states that cars are racing, you should
have a class named Car and not a class called Automobile.
BDD doesn’t care about how the app is architected or implemented; its
main focus is on the behavior.
In BDD, you can write the acceptance criteria as follows:
Given [a context]
When [an event happens]
Then [assert something]
Cucumber is a tool that supports BDD using this style for the acceptance
criteria. It reads executable specifications that you provide written in
plain English and it'll validate that your app does what those
specifications state. The specifications consists of multiple examples, or
scenarios.
For example, suppose you’re writing a calculator application for Android.

You could write the following feature file for Cucumber:
Feature: Calculate a result

Perform an arithmetic operation on two numbers using a mathematical o
Scenario Outline: Enter a digit, an operator and another digit

Given I have a CalculatorActivity
When I press <num1>
And I press <op>
And I press <num2>
And I press =
Then I should see "<result>" on the display
Examples:
| num1 | num2 | op | result |
| 9 | 8 | + | 17.0 |
| 7 | 6 | – | 1.0 |
| 5 | 4 | x | 20.0 |
| 3 | 2 | / | 1.5 |
| 1 | 0 | / | Infinity |
Here’s how this works:
You specify the name of the feature, give a description and write the
scenario using the given/when/then format, in plain English.
You also have to translate those steps into Java or Kotlin. These steps are
known as step definitions, For example:
@When("I press {operator}")
fun i_press_op(op: Char) {
when (op) {
'+' -> onView(withId(R.id.btn_op_add)).perform(click())
'–' -> onView(withId(R.id.btn_op_subtract)).perform(click())
'x' -> onView(withId(R.id.btn_op_multiply)).perform(click())
'/' -> onView(withId(R.id.btn_op_divide)).perform(click())
'=' -> onView(withId(R.id.btn_op_equals)).perform(click())
}
}
The step I press {operator} is parameterized with an operator.

Therefore, you create a function that handles this and translates each of
the operators into an action within the UI. In this case, it’s pressing the
corresponding button. This type of code is often called glue-code
because it maps the English sentences into a lower-level language such
as Kotlin.
By having enough step definitions mapped, you can let a non-technical

person write various scenarios.
ATDD vs. BDD

You may find that some people consider both techniques the same thing
since they focus on the end user requirements, understanding behaviors
and generating similar outcomes.
Here’s a closer look at each technique you've seen so far:
TDD enforces good class design, ensuring that those classes work
correctly in units and collaboration with others. ATDD and BDD
enforce building the appropriate app for the user.
ATDD focuses on acceptance tests written in collaboration with the

stakeholders. Sometimes these tests are written using spreadsheets
and are understandable by non-technical people. However, at some
point, the technical team needs to translate these tests to an
automatizable test in code.
In BDD, you try to bring the business team and the technical team
closer by using a common language that’s understandable by both
parties. Using the given/when/then style, even non-technical people
can write tests that can directly run in automated environments.
ATDD and BDD both have an objective of tackling down any

misunderstood requirements inside of the team. All team members
should interpret them the same way.
No matter which technique you choose or if you take parts of each one,
they're all great practices to explore on your team.
Key points
In TDD, you need to write tests before adding or modifying features.
You write tests in a technical language such as Java or Kotlin.
The people involved in TDD need to have a technical background.

But in BDD or ATDD, both technical and non-technical people on
the team can get involved.
TDD focuses on the implementation of the features. BDD focuses on

the behavior. ATDD focuses on capturing the requirements and the
acceptance criteria.
Each of the techniques outlined in this chapter enforces creating

tests before adding or modifying features. This is in contrast to
traditional development.

If you want to learn more about other TDD techniques, look at the
following:
"TDD: State of the art and trends" by Carlos Fontela:

http://web.fi.uba.ar/~cfontela/Fontela_EstadoDelArteTDD_UNLP_EIS
.pdf
To write tests with the given/when/then style using Cucumber, check

its documentation: https://cucumber.io/docs
Conclusion
Congratulations on completing this book! You
should have a firm understanding of Android
test-driven development. Throughout this book,
you learned how to write a variety of different
tests for many different apps in various states of
their development life cycle.
We hope this book provides you an awesome

reference as you undertake making high quality
apps built with TDD.
If you have any questions or comments as you

work through this book, please stop by our
forums at http://forums.raywenderlich.com and
look for the particular forum category for this
book.
Thank you again for purchasing this book. Your

continued support is what makes the tutorials,
books, videos, conferences and other things we
do at raywenderlich.com possible, and we truly
appreciate it!
Wishing you all the best in your continued
Android adventures,
Victoria, Fernando, Lance, Alex, Matei, Tammy,

Manda and Jonathan
The Android Test-Driven Development by

Tutorials team

Android Test-Driven Development by Tutorials

Uploaded by

Copyright:

Available Formats

Android Test-Driven Development by Tutorials

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Android Test-Driven Development by Tutorials

Uploaded by

Copyright:

Available Formats

What are the different techniques discussed in the book?

What are the different techniques discussed in the book?

What is the difference between TDD, ATDD and BDD?

What is the difference between TDD, ATDD and BDD?

Android

Copyright ©2019 Razeware LLC.

"To my girlfriend Romina, who will soon become

Victoria Gonda is an author of this book.

Tammy Coron is an editor of this book. She is an

Alex Curran is a technical editor of this book. He

Matei Suica is a technical editor of this book. He

Vicki Wenderlich is the designer and artist of

Kotlin 1.3: This book uses Kotlin 1.3

Android Studio 3.5 or later. Android Studio

If you haven't installed the latest version of

You are allowed to use and/or modify the

You are allowed to use and/or modify all

The source code included in Android Test-

This book is for your personal use only. You

All materials provided with this book are

In no event shall the authors or copyright

All trademarks and registered trademarks

If you bought the print version

Buying the digital edition version of the book

Visit our Android Test-Driven Development

And if you purchased the print version of this

The best way to get update notiﬁcations is to

Ravens are extremely adaptable and will test

With Android Test-Driven Development, you’ll

You’ll then get your feet wet by writing your ﬁrst

Speciﬁcally, this section includes the following

Chapter 1: Introduction: If you're new to testing

Chapter 3: What Is TDD? Now that you know

In this book, you’ll learn about test-driven

Whether you’re a testing beginner, or ready to

Although it’s not required, it’s helpful to have

This book covers everything from the bare

How to read this book

Section 1: This is where to start if you’re

Section 2: If you know the fundamentals of

Section 3: When everything in the previous

With that, it’s time for a sneak peek of each

You’re here now. You’re learning what this book

Chapter 2: What Is a Test?

Continue to this chapter to learn about

Chapter 3: What Is TDD?

This chapter introduces a core concept of this

As you might imagine, there are multiple ways

Chapter 5: Unit Tests

This is the chapter where you’ll learn about unit

Chapter 6: Architecting for Testing

The architecture of your project can make or

Revisiting the Cocktail app, you’ll take your unit

This chapter looks at testing how different

Chapter 9: Testing the Persistence Layer

Continuing with the Wishlist app, this chapter

Chapter 10: Testing the Network Layer

While HTTP requests can be unpredictable,

Chapter 11: User Interface

Finishing out this section is all about testing