MIME-Version: 1.0
References: 
 <CAApHDvq-4994jpDnspQ+bgSbzo0+=Z4YQbj+YfP7wi97B0e1qg@mail.gmail.com>
 <76DBEF5D-EA47-4C30-8D28-97B3B88E2E2D@greg.burd.me>
 <CAApHDvrvixMmscGEfEyYz0=qbwzM7Y0Up-Uh_AKVmfn8K2sn_A@mail.gmail.com>
 <aNxpxE6a9G3sbfrm@paquier.xyz>
 <CAApHDvoC-W3s2QJdBY6B32BXy2br9iGsNJ3v4uLFOosBgmY2YA@mail.gmail.com>
 <aNyiu0fKRqihdwpV@paquier.xyz>
In-Reply-To: <aNyiu0fKRqihdwpV@paquier.xyz>
From: David Rowley <dgrowleyml@gmail.com>
Date: Wed, 1 Oct 2025 18:00:59 +1300
Message-ID: 
 <CAApHDvrEPw6DAQdZg7u1YknVvN1Gy38kUURQm8bOHSKU2cw+4w@mail.gmail.com>
Subject: Re: [PATCH] Add tests for Bitmapset
To: Michael Paquier <michael@paquier.xyz>
Cc: Greg Burd <greg@burd.me>,
 PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Archived-At: 
 <https://www.postgresql.org/message-id/CAApHDvrEPw6DAQdZg7u1YknVvN1Gy38kUURQm8bOHSKU2cw%2B4w%40mail.gmail.com>
Precedence: bulk

On Wed, 1 Oct 2025 at 16:41, Michael Paquier <michael@paquier.xyz> wrote:
> Okay.  Let's just remove as much code as possible, then.  This is
> roughly what you have sent, minus the hardcoded empty set and the free
> calls.  I'll just go do that in a bit.

Thanks.  Just for reference, you may not need it if you're mostly done
already, but I had to finish this off in my head before moving on.
Feel free to ignore, but there are a couple of typos fixed at the very
least "varrying" should be "varying".

I didn't fully rationalise the tests in test_bitmapset.sql.  There are
still a few which are duplicates due to '(b)' and NULL meaning the
same thing.

git diff --stat
 .../test_bitmapset/expected/test_bitmapset.out     | 400 ++------------
 .../modules/test_bitmapset/sql/test_bitmapset.sql  | 101 +---
 .../modules/test_bitmapset/test_bitmapset--1.0.sql |   2 +-
 src/test/modules/test_bitmapset/test_bitmapset.c   | 605 +++++----------------
 4 files changed, 222 insertions(+), 886 deletions(-)

I didn't look at the coverage report, but on thinking about it,
test_bitmapset.c's "PG_RETURN_NULL(); /* invalid input */" won't be
covered as I stripped out all the tests which purposefully pass
invalid input. You may not want it that way if you're keen to keep the
coverage of the test module high as well as bitmapset.c

David