Provided by: shellia_5.3.2_all
shellia - library for interactive shell scripts
shellia is a library that allows to run shell scripts interactive. The word interactive is the most characteristic adjective for shellia and the shortest abbreviation found for interactive is ia. Such a short abbreviation is needed, because most commands brought by shellia start with this short prefix ia. shellia is composed from shell and ia. See also PREFIX in shellia(3).
shellia is a library that allows to run shell scripts silently or interactive and helps to check errors of commands that are combined in steps. This makes it easy to familiarize oneself with a shell script that uses shellia and find unexpected behaviour in the shell script. Also when running the shell script silently shellia can check it step by step. Shell scripts often become large with many subfunctions, calling many commands. This leads to the following problems, where shellia can help: · It is not easy to run only parts of the script, to learn about an existing script, to find an error in a script, to change the order of parts of the script for testing or to rerun only a part of the script that failed. · It is not easy to find out, from which part of the script errors, warnings or messages that appear on stdout or stderr are created. And it is often not clear if output of a shell script is ok and can be ignored. · Often commands are started in a shell script without checking for Errors because this would make the scripts too complex. · Many programs can only enable debug at start time. This can result in much debug output, also if only specific debug output is needed. The shellia library can help with the following features. · run a script or parts of a script interactively · check each command for unexpected output or returncode · jump to interactive mode if an error occurs or exit the script · helps to structure large scripts · in interactive mode shellia can toggled the debug output level
When building a ./script.using.shellia the following hints may be helpful: Do not use a single command as a step It is important to find the right size of a step to be uses in shellia. The idea of a step is not to check the execution of a single command. Instead a single command can be part of a function, that is called by a step. Input and Output of a step Each step should be interactively runable by an end user of the script. This means the end user should be able to handle input and output of a step. This often means a step should not read stdin or send stdout to another step. For example a backup script could have one step to backup /home and another step to backup /usr. But a step that reads data from work disk and another step that writes backup data to a backup disk should better be combined in a single step. Reasonable size of interactive steps If the user of the script can not imagine what an interactive step does, this step may be to detailed and to small. If to many errors may happen in the interactive step, the step may be to large. For example, if you need to mount an external disk, this should be a separate interactive step. Because this can give errors, if the disk is not readable. And this errors can again lead to follow up errors. When to use check option -c and -C Both options will check the output and the exit code of a step. All output has to be allowed with extended regular expressions in check-mode. To be able to control the exit code in the same way, the line exit=<exit code> (with <exit code> replace by the exit code), will be added, if the exit code is nonzero. With option -c, per default every output to stdout and stderr is seen as one output and is displayed as an error. The developer can change this with ia_stdout and ia_ignore. This means with -c every output is checked, also output that would normally go to stdout. This makes it easy to find changed output from a command used, e.g. when upgrading to a new operating system version. With option -C, per default every output to stdout is unchanged and all output to stderr is seen as an error. The devloper can ignore error output with ia_ignore. This means the developer has not to handle stdout, which can be good and bad. If a step calls a programm, that produces stdout and stderr output, and it is most important to retain the order of the output, option -c should be used. With option -c the output of both streams would be combined by shellia to one stream. If it is important to know if output is from stdout or from stderr to decide if it is an error, option -C should be used. Use --quiet, --silent or similar if available If commands are used with check option -c or -C, less ia_stdout or ia_stderr lines have to written, if the commands create less output.
For example imagine a backup script, that writes to an external disk, that should then contain an encrypted and bootable backup. This script may call the following steps, that are itself functions in the script. The main part of the script could look like this: eval "$ia_init" ia_add "opencrypt <-i>" ia_add "mountbkup <-i>" ia_add "synccopy" ia_add "lvcopyvms <-i> ia_add "fixcrypttab" ia_add "fixfstab" ia_add "prepare_chroot" ia_add "fixinitrd" ia_add "run_grub" ia_add "clean_chroot" ia_add "umountbkup <-i>" ia_add "closecrypt <-i>" ia -C