Re: [PATCH] CI: Add a CPAN cache on Windows

From: Jacob Champion <jacob(dot)champion(at)enterprisedb(dot)com>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Nazir Bilal Yavuz <byavuz81(at)gmail(dot)com>
Subject: Re: [PATCH] CI: Add a CPAN cache on Windows
Date: 2026-06-09 17:20:28
Message-ID: CAOYmi+m-gwoQfj93xA8-3u7Z8jesmwsYvQz_M4Hyit86Hv=gQw@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, Jun 5, 2026 at 1:12 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> I'm not particularly concerned about that, tbh. If you suddenly switch which
> perl you use, you'll might get some weird errors.

I'm finding that's easy to do by accident, though.

The interactions between
- Git-for-Windows MSYS,
- our full version(s) of MSYS on either C: or D:,
- the environment variables we set (esp. MSYSTEM), and
- the GITHUB_PATH
are fairly unintuitive to me. The windows-vs job pushes Strawberry
Perl into the PATH -- but I think that only works because there's no
MSYSTEM setting? I haven't found a way to get the windows-mingw job to
do the same thing, because it forces its own "native" Perl to take
precedence.

So if we want the two bash shells to behave similarly, I think some
more fundamental things need to be changed first, probably by someone
who understands MSYS2 better than I do. Maybe that's just a really
small tweak, but I can't see it.

--Jacob

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message surya poondla 2026-06-09 17:49:44 Re: Improving psql autocompletion for SET LOCAL / SET SESSION
Previous Message Tom Lane 2026-06-09 17:03:49 Re: Fix missing semicolon in pl_gram.y for option_value rule