Re: [PATCH v3 1/2] t: add a test helper for getting hostname

[Jeff King <peff@peff.net>]

On Tue, Mar 19, 2024 at 07:37:21PM +0100, Ignacio Encinas wrote:

> diff --git a/t/t6500-gc.sh b/t/t6500-gc.sh
> index 18fe1c25e6a0..613c766e2bb4 100755
> --- a/t/t6500-gc.sh
> +++ b/t/t6500-gc.sh
> @@ -395,7 +395,6 @@ test_expect_success 'background auto gc respects lock for all operations' '
>  
>  	# now fake a concurrent gc that holds the lock; we can use our
>  	# shell pid so that it looks valid.
> -	hostname=$(hostname || echo unknown) &&
>  	shell_pid=$$ &&
>  	if test_have_prereq MINGW && test -f /proc/$shell_pid/winpid
>  	then
> @@ -404,7 +403,7 @@ test_expect_success 'background auto gc respects lock for all operations' '
>  		# the Windows PID in this case.
>  		shell_pid=$(cat /proc/$shell_pid/winpid)
>  	fi &&
> -	printf "%d %s" "$shell_pid" "$hostname" >.git/gc.pid &&
> +	printf "%d %s" "$shell_pid" "$(test-tool xgethostname)" >.git/gc.pid &&

Hmm. Before, we were compensating for a failure to run "hostname" by
putting in the string "unknown". But I wonder if there were platforms
where gethostname() simply fails (e.g., because the hostname is too
long). In that case your test-tool returns the empty string, but git-gc
internally will use the string "unknown".

I think it may be OK, though. The failing exit code from test-tool will
be ignored, and we'll end up with a file containing "123 " or similar.
Normally we'd use kill() to check that the pid is valid, but with a
mis-matched hostname we'll just assume the process is still around, and
the outcome is the same.

-Peff