[Xquartz-dev] 2.3.2_rc4

[Harald Hanche-Olsen]

+ Jeremy Huddleston <jeremyhu at apple.com>:

> So this thread has pretty much settled down, and I'm still not 100% 
> sure where the chips landed.

The chips have landed? Umm, okay, in a manner of speaking I guess 
they have.

> Did I miss a criticism of this approach (most replies seemed to be 
> questions/comments rather than problems).

Not really, except Cameron Simpson would like to replace the final 
fall through clause in the case statement to call his own generic 
script instead. Personally, I have no really strong opinions about 
that, except I wonder if it adds a lot of flexibility to the 
approach. For example, I pretty much doubt that it would work with 
scsh. But then, after browsing through scsh.net for a while I get the 
distinct impression that there is no such thing as a scsh login 
shell, nor do I think there is a default startup file being run by 
scsh; in short, I think we should not worry about scsh. If anybody 
out there is masochist enough to run scsh as their login shell, we 
can also reasonably expect them to find out about and use the .x11run 
mechanism.

I have only two small changes to the script as you posted:
Quote "${SHELL}" in the basename call, and we might as well use exec 
-l on es and rc to make this case look more like the others, thus 
highlighting the essential differences between the shells and 
ignoring the trivial ones. This works for me at least:

#!/bin/bash --login

set "$(dirname "$0")"/X11.bin "${@}"

if [ -x ~/.x11run ]; then
	exec ~/.x11run "${@}"
fi

case $(basename "${SHELL}") in
	ksh|sh|zsh)    exec -l "${SHELL}" -c 'exec "${@}"' - "${@}" ;;
	csh|tcsh)      exec -l "${SHELL}" -c 'exec $argv:q' "${@}" ;;
	es|rc)         exec -l "${SHELL}" -c 'exec $*' "${@}" ;;
	*)             exec    "${@}"
esac

- Harald