Re: [PATCH] random_normal function

Hello old thread.

On Jan 19, 2023, at 01:01, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> I think "ignore:" was a kluge we put in twenty-plus years ago when our
> testing standards were a lot lower, and it's way past time we got
> rid of it.

Today I discovered a bad regex in the pgTAP Makefile that matched 18.0 as an 8.0 release and used it `ignore:` a test requiring RLS. I found it after I saw this test failure on Postgres 18:

# syntax error in schedule file "test/build/run.sch" line 3: ignore: policy

I fixed the regex (removed any detection of pre-9.0 matching, in fact), so all good there. But I wanted to point out that `ignore:` is documented in the wiki[1], so other extensions may use it. If no one else notices perhaps it’s not a big deal, but I want to call out that pg_regress features, “documented” or not, might be used outside the core.

Maybe the proper way to address this issue is to add formal docs for pg_regress. I always have the damnedest time finding its docs and usually end up on the Wiki. If there was official documentation for it like there is for `psql`, etc., it might help to prevent such issues in the future. Minor as they may be, admittedly.

Thoughts?

Best,

David

[1]: https://wiki.postgresql.org/wiki/Regression_test_authoring