mod-to-sh sub-command

Produce shell code of environment changes made by loading designated modulefiles and send this content to the report message channel. User views the produced shell code as output.

mod-to-sh could be send as the opposite of sh-to-mod. The former transforms a modulefile into shell code whereas the latter transforms shell code into a modulefile.



Sub-command properties

General properties:

  • Shortcut name: none

  • Accepted option: --force, --icase, --auto, --no-auto

  • Expected number of argument: 2 to N

  • Accept boolean variant specification: yes

  • Parse module version specification: yes

  • Fully read modulefile when checking validity: yes

  • Sub-command only called from top level: yes

  • Lead to modulefile evaluation: yes (load mode)

Arguments expected:

  • First argument should be a supported shell name

    • like bash or fish

    • An error is produced if an unsupported shell is specified

    • It defines the language that sub-command will use to produce shell code

  • After the first argument, comes the modulefile specification

    • Multiple modulefiles can be specified

    • With advanced version specifiers and variants

    • It defines the modulefiles to load

mod-to-sh loads designated modulefile(s) and instead of rendering the corresponding environment changes to the stdout channel, the shell code is sent to the message report channel (stderr).

Target shell is changed, from the one specified as argument to modulecmd.tcl script to the one specified as argument to mod-to-sh sub-command.

Once shell code is produced and sent to the message channel (stderr), target shell defined for modulecmd.tcl is restored to get on the regular code channel (stdout) the evaluation status. For instance for sh shell: test 0 if successful, test 0 = 1 otherwise.

Shell code produced by mod-to-sh is only sent to message channel. It is not sent to the code channel.

modtosh_real_shell state is introduced and set by mod-to-sh sub-command. When defined, shell code rendering procedure knows if code should be sent to the message channel and then cleared, not to be rendered on code channel. This state contains the target shell of modulecmd.tcl. So shell rendering procedure can restore target shell then produce the final shell code to render (evaluation status).

Modulefile evaluation is made like a regular module load. Thus if specified modulefile is already loaded, its corresponding environment changes will not appear on the mod-to-sh result.

Shell code produced for mod-to-sh is filtered to only contain environment changes of designated modulefiles and not the Modules-specific environment context (like LOADEDMODULES or _LMFILES_ environment variables).

mod-to-sh sub-command automatically adjusts the verbosity configuration option to the silent mode. As produced shell code is sent to the message channel, it should not mix with report messages. Setting verbosity to silent disables loading messages or reportWarning, reportError and puts stderr messages defined in modulefiles. verbosity is not changed if set to trace mode or any higher mode (debugging).

module-info command

As mod-to-sh leads to modulefile evaluation, the module-info command modulefile command returns the mod-to-sh string when used in a modulefile evaluated through this sub-command. It can also be queried against this command name string to return Boolean value depending if mod-to-sh is the sub-command currently running.