Re: [PATCH] t7500: fix tests with absolute path following ":(optional)" on Windows

From: Junio C Hamano <gitster@pobox.com>
Date: Oct 20, 2025, 18:06 UTC
Message-ID: <xmqqtsztzbvo.fsf@gitster.g>
In-Reply-To: <5d780103-285b-4e6c-9b26-2a87609837cf@kdbg.org>

Johannes Sixt <j6t@kdbg.org> writes:

Show 11 quoted lines

> Am 20.10.25 um 15:43 schrieb Ben Knoble:
>> Going forward I will probably stick with using pwd, given the
>> difference in platform behavior.
> $(pwd) is usually safe, but not always. If we have to look at every
> instance anyway, we can use $PWD for efficiency if it does not matter,
> and $(pwd) only when it is necessary.
>
>> Is there a doc or test lint for that? If not, might be useful.
>
> If this were documented somewhere, would you have found it and obeyed
> the recommendations?

I myself forget about it every time, even after getting bitten at least 3 times in the past, maybe more.

t/README has this.

 - When a test checks for an absolute path that a git command generated,
   construct the expected value using $(pwd) rather than $PWD,
   $TEST_DIRECTORY, or $TRASH_DIRECTORY. It makes a difference on
   Windows, where the shell (MSYS bash) mangles absolute path names.
   For details, see the commit message of 4114156ae9.

It is mentioned in t/README, I know it is mentioned in t/README, and I did re-read the part of t/README, every time I needed to decide between $PWD and $(pwd), but I still got it wrong 50% of the time X-<.