exercises 2

2. Basic Tasks

Everybody should be able to complete the following tasks.

2.1. Parameters/Variables

The following tasks concern primarily the use of parameters/variables.

2.1.1. Counting

The Linux utility seq are useful for generating sequences of numbers. Here, you will write your own approximation to these utilities. Write a pure shell script seq.sh for counting.

Your script must accept and require exactly two integer arguments:

the value at which to start counting

the value at which to stop counting

Output must be one number per line.

You may assume:

that exactly two arguments are indeed provided,

that the arguments provided are indeed integers, and

that the stop value is greater than or equal to the start value.

Here are some examples:

$ sh seq.sh 1 4

1

2

3

4

$

$ sh seq.sh 98 101

98

99

100

101

Tip

You will need to use:

a while loop

the -lt operator of the test command (man test)

arithmetic expansion

2.1.2.Checking Arguments

Extend your script seq.sh such that it verifies that exactly and only two arguments have in fact been provided.

Examples:

$ sh seq.sh

usage: seq.sh START STOP

$ sh seq.sh 100

usage: seq.sh START STOP

$ sh seq.sh 98 100

98

99

100

$ sh seq.sh 98 100 4

usage: seq.sh START STOP

Tip

The special variable $# contains a count of the number of arguments provided.

Read the manual page for test to see how you can compare values.

Your script should yield an appropriate exit code.

2.1.3. Default Arguments

Extend your script seq.sh such that:

if no second argument is provided, then a default start value of 1 is used

Example:

$ sh seq.sh 4

1

2

3

4

Tip

The form ${parameter:-word} can be used to use default values. The form ${parameter:=word} can be used to assign default values.

2.1.4. More Default Arguments

Extend your script seq.sh such that:

if no arguments are provided, then it counts from 1 upwards forever (or, at least until stopped by some other means).

Example:

$ sh seq.sh | head -6

1

2

3

4

6

Tip

Look carefully at and extend the condition of your while loop.

2.2. Command Substitution

The following tasks concern primarily the use of command substitution.

2.2.1. Preliminaries

The du -s command outputs a summary of the disk space occupied by a file or directory (du stands for disk usage, s stands for summary).

Examples:

$ du -s ~/.zshenv

4       /home/charlie/.zshenv

$ du -s ~/local/sbenv

65288   /home/charlie/local/sbenv

The cut utility selects individual or groups of fields from its input.

Examples:

$ du -s ~/.zshenv | cut -f 1

4

$ du -s ~/local/sbenv | cut -f 2

/home/charlie/local/sbenv

Tip

For the details of what these utilities do, read their manual pages.

Many utilities accept flags which adjust their behaviour. If you need a particular behaviour from a utility, it's likely that others have needed that behaviour too, so its likely that a relevant flag exists.

2.2.2. File/Directory Sizes, From Oldest to Newest

Write a script sizes.sh which outputs one line for each file or directory in the current working directory:

each line must contain

the size of the file/directory contents in kilobytes

the name of the file or directory

lines must be ordered by their last-modified date, from oldest to youngest

Example:

$ sh sizes.sh

4       Makefile

12      labsheet-01-intro.ascii

36      labsheet-01-intro.html

20      01-intro.slide

168     01-intro.html

192     02-more.html

28      02-more.slide

40      labsheet-02-more.html

12      labsheet-02-more.ascii

(These files are in order of last modified date).



Tip

You may choose to use command substitution.

The command ls -tr outputs a list of files/directories in the required order.

Does your solution require a while loop? A solution does not requires a while loop.

2.2.3. File Size Report

Write a scipt file_size_report.sh which:

accepts some number of file names as arguments, and

for each file, outputs a message:

The size of FILENAME is SIZE kilobytes

(where FILENAME and SIZE are replaced appropriately.

Example:

$ sh file_size_report.sh ~/.zshenv ~/local/sbenv

The size of /home/charlie/.zshenv is 4 kilobytes

The size of /home/charlie/local/sbenv is 65288 kilobytes

Tip

You may choose to use command substitution, and may have to use a pipe within the command being substituted.

2.3. Shell Functions

The following tasks concern primarily the use of shell functions.

2.3.1. Web Cat

Write a shell function wcat which accepts some number of URLs as arguments and downloads each, outputting the downloaded content to standard output.

Example:

$ wcat http://www.google.com/ http://www.facebook.com/

.

<copious output>

Tip

Use the wget utility. You'll have to read its manual page. There is a wget option to direct output to standard output.

2.3.2. Repeat

Write a shell function called repeat which accepts a command and (optionally) some arguments and executes the command with its arguments (if any) once per second.

$ repeat date

Thu 10 Nov 2011 10:34:31 GMT

Thu 10 Nov 2011 10:34:32 GMT

Thu 10 Nov 2011 10:34:33 GMT

Thu 10 Nov 2011 10:34:34 GMT

...

$ repeat du -s ~

968238  /home/charlie

968129  /home/charlie

967846  /home/charlie

967732  /home/charlie

...

Tip

The command sleep can be used to sleep for some number of seconds.

2.3.3. Find a Usable Executable

Write a shell function usable which accepts some number of arguments and outputs the full path of the executable of the first for which the command is in fact installed on the system at hand.

Example:

$ usable firefox opera uzbl

/usr/local/bin/opera

$ usable urxvt rxvt xterm

/usr/local/bin/xterm

Here, no command named firefox is installed, but opera is. urxvt and rxvt are not installed, but xterm is.

Tip

The external command 'which' should be used to locate executables. Use command 'which XXX' to be sure that you are indeed using the external which command.

Your function should succeed if a suitable executable is found, and fail otherwise.

3. Advanced Tasks

Some of these tasks are significantly more difficult, and may involve material that has not been covered in the course.

3.1. Capture Output

The rationale for this exercise will become apparent in the context of the [chronix] task, below.

Write a script capture:

capture command [ arguments ... ]

capture

When called with a command and (optional) arguments, capture:

calls the command with its arguments, capturing the command's standard output and standard error in a temporary file, and

outputs to standard output only the name of the temporary file.

Its exit code should be the same as that of the command.

For example:

#!/bin/sh

capture ls /etc           # outputs only "/tmp/_capture_charlie_37176.Hs6dqi"

                          # $? is '0' here

capture ls /non-existent  # outputs only "/tmp/_capture_charlie_37269.k2s8AC"

                          # $? is '1' here

In this case, /tmp/_capture_charlie_37176.Hs6dqi contains the output of the first command, and /tmp/_capture_charlie_37269.k2s8AC contains the output of the second command.

The temporary file should be created in the directory indicated by the environment variable TMPDIR, if it exists and is not null, or in the directory /tmp, otherwise.

When called with no arguments, capture should output a (possibly empty) list of previously-created temporary files that have not yet been removed.

For example:

#!/bin/sh

capture | xargs rm

could be used to remove any remaining temporary files.

Tip

The mktemp utility may prove useful (man mktemp).

3.2. Chronix

The cron daemon can be used to run jobs at scheduled intervals.

If a cron job produces any output (on either standard output or standard error), then that output is emailed to the owner of the job.

This creates a problem.

Frequently, cron jobs do produce output, although that output is of little interest.

Frustrated by the number of unnecessary emails they receive, users add flags to commands or redirect their outputs to suppress these unnecessary emails.

This, of course, is then a problem when from time to time the output of a cron jobs is actually needed.

Write a script chronix:

chronix command [ arguments ... ]

chronix calls the given command with its arguments, if any, and behaves as follows:

If the command's exit code is 0, then chronix discards all of the commands outputs and exits itself with an exit code of 0.

If the command's exit code is non-zero, then chronix:

outputs to its own standard output all of the output which the command produced, and

exits with the same exit code as that of the command.

For example, chronix might be used in a crontab file as follows:

@daily chronix sh backup-script.sh

Here, the backup script may or may not produce output. If it succeeds, any output will be discarded. If it fails, its output will be emailed to the user.

Tip

The [capture] script described above may prove useful.

chronix should remove any temporary files it creates.

Note

This exercise is based on the chronic command which is part of the moreutils suite of utilities.