Processes

Lecture 4

Michael J. Wise
 Processes
• A process is program in
execution
• Each call to bash or opening a
new window creates a separate
processes.
• Windows can be created by the
window manager (via menu)
• The process that controls the
window is called the foreground
process. That is, it can take
input from the keyboard (if the
window is active) and write to
the window.

L4 Processes - 2
 Processes
• You can run more than one process in a window by
making the other processes background processes.
(simply append & to end of command).
• Note: Background processes can neither accept
keyboard input nor write to a window, so if either is
required it must come from/go to a file.
E.g. % gzip large_file.txt > trace 2> errs &
Which then allows you to get on with other things will
the gzip file compression is happening the background.

BTW. Unix gzip compresses files (gunzip extracts the

L4 Processes - 3
 Some Process Oriented Commands
ps <options>
• ps prints out information about processes you have
running. If no options are specified, only the
processes for the currently active window are
reported.
• -l Provides a long, listing containing more
information.
• –f provides more information about the command
(which options were used)
• The most important pieces of information are the
names of the processes and their process IDs or
simply PIDs, e.g.
3638 ttys002 0:00.06 python
8808 ttys003 0:00.03 bash
L4 Processes - 4
 Some Process Oriented Commands
• top
Lists a range of statistics about all the processes, and
then lists the processes one by one in order of their
CPU use. Useful for finding that process that has gone
off the rails.
– lots of processes are asleep
• kill <options> <PIDs>
A signal is sent to each of the listed processes. The
options - generally not required - specify which signal
is to be sent; kill -9 is stronger medicine

L4 Processes - 5
 Interlude: More Useful Commands - cut
• There is a HUGE number of commands/analyses
available for Unix system. Lots to choose from. This is
selection from those I most frequently use.
cut [ -d <delim>] -f <list> <files>
Given an ASCII file structured as a number of fields
separated by some delimiter, e.g. <TAB>, cut returns
the fields indicated.
-f The comma separated list following –f specifies the
fields that are to remain
-d Specifies an alternate delimiter
For example:
cut –d , -f 1,5-7 datafile.csv

L4 Processes - 6
 Interlude: More Useful Commands - paste
paste [ -d <sep> ] <files>
Steps through one or more files in parallel. The output is
a line containing all the first lines (separated by <sep>,
say the Tab char), followed by a line containing all the
second lines, and so on.
% cat file1 % cat file2
Twinkle Twinkle little bat
How I wonder what you're at!
Up above the world you fly
Like a tea tray in the sky

What does paste -d" " file1 file2 do?

L4 Processes - 7
 Interlude: More Useful Commands - tr
tr <options> <string1> <string2>
• Translate characters in the first string (say lower
case letters) to characters in the second string (UC
letters). Input from stdin.
• Most useful options are:
-C Complement characters in the first string
-s Squeeze multiple contiguous occurrences of string1
characters
tr ’[A-Z]’ ’[a-z]’ < textfile
tr –s ’ ’ ’#’ < textfile
tr ’\015’ ’\012’ < windows_textfile
The mapping of characters to codes: man ascii
L4 Processes - 8
Interlude: More Useful Commands - comm
comm [<output options>] <file1> <file2>
The two sorted files are compared, and three column
output is produced. Column 1 contains those lines only
found in <file1>, Column 2 contains lines only in
<file2>, Column 3 contains lines common to both files.
The output options suppress the named columns,e.g.
-1 Suppress column unique to <file1>
-23 Suppress column unique to <file2> and column of
common entries

L4 Processes - 9
 Interlude: More Useful Commands - uniq
uniq <file>
Compresses adjacent lines that are repeated in a file to
a single copy (i.e. duplicates removed). The
-c option outputs the single lines together with the
counts of the number of copies of those line.

L4 Processes - 10
 Interlude: More Useful Commands - sort
sort [options] <files>
This powerful sort utility takes the named files (or
stdin) and sends the sorted output to stdout. It has
MANY options. Some are:
-u Removes duplicated entries following the sort (i.e. as
if uniq had been applied)
-t The character following the -t flag is used as the field
separator (the default separator is <TAB>.

L4 Processes - 11
 Sort
-k <start>[<type>][,<end>[<type>]
Instead of sorting from the first field, sort from the
<start> field. If <end> is specified, only the fields from
<start> to <end> are sorted.
For example:
ps| sort -k 4 # list processes alphabetically

L4 Processes - 12
 Demo L0 (again)
• In this revisiting of the L0 demo, I want to be able to
have reported the list of the top N words (in order of
descending order of occurrence). In other words, the
user specifies both the file and the number of top
hits to be reported.

L4 Processes - 13
 Demo
• Real example. In a bit of a version control problem, I
found I had two directories with over-lapping sets of
files; some with the same name, some different. I
needed to write a small script, dir_diffs, which,
given two directories, lists the files that are unique
to the first directory and the files that are unique to
the second directory.

L4 Processes - 14