Re: [PATCH] t7507-*.sh: Fix test #8 (could not run '"$FAKE_EDITOR"')

From: Junio C Hamano <gitster@pobox.com>
Date: Nov 20, 2013, 18:33 UTC
Message-ID: <xmqqhab67wjr.fsf@gitster.dls.corp.google.com>
In-Reply-To: <20131120174226.GA16453@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 9 quoted lines

> The test_set_editor helper does some magic to help with quoting, but
> that should not be an issue in this case (since we are using "cat"). We
> are using test_set_editor elsewhere in the script, which would have set
> EDITOR previously. But I would think that GIT_EDITOR, which we are using
> here, would supersede that. However, the error message he shows
> indicates that git is using EDITOR (as FAKE_EDITOR is part of that quote
> magic).
>
> Am I misremembering the issues with one-shot variables and functions?

I think there are two problems involved.

The first is that we are not really using GIT_EDITOR under some shells; to wit:

------------------------ >8 ------------------------

$ cat >/var/tmp/dashtest.sh <<\EOF #!/bin/sh

test_must_fail () {
	(
		env | sed -n -e '/EDITOR/s/^/>> /p'
	)
}

EDITOR=dog export EDITOR

GIT_EDITOR=cat test_must_fail foo EOF

$ dash /var/tmp/dashtest.sh

>> EDITOR=dog

$ bash /var/tmp/dashtest.sh

>> GIT_EDITOR=cat
>> EDITOR=dog

------------------------ 8< ------------------------

So it appears that GIT_EDITOR that was never exported in the script fails to get exported with the "VAR=VAL cmd" syntax under dash, when cmd is not a command but is a shell function. The "git commit" in question ends up using $EDITOR.

Another is that EDITOR="$FAKE_EDITOR" that is set up earlier in the is having trouble launching (I have a feeling that it never was actually used because everybody uses "commit -F <file>").