Re: TAP tests for psql \g piped into program

On 02.01.23 22:32, Daniel Verite wrote:
> This is a follow-up to commit d2a44904 from the 2022-11 CF [1]
> The TAP tests were left out with the suggestion to use Perl instead of
> cat (Unix) / findstr (Windows) as the program to pipe into.
>
> PFA a patch implementing that suggestion.

The perl binary refactoring in this patch caught my attention, since I
ran into this issue in another patch as well. I'm always happy to
consider a refactoring, but I think in this case I wouldn't do it.

If you grep for PostgreSQL::Test::Utils::windows_os, you'll find quite a
few pieces of code that somehow fix up paths for Windows. By hiding the
Perl stuff in a function, we give the illusion that you don't have to
worry about it and it's all taken care of in the test library. But you
have to worry about it in the very next line in
025_stuck_on_old_timeline.pl! We should handle this all on the same
level: either in the test code or in the test library. It would be
useful to work toward a general "prepare path for shell" routine. But
until we have that, I don't think this is sufficient progress.

So for your patch, I would just do the path adjustment ad hoc in-line.
It's just one additional line.