From: | Dagfinn Ilmari Mannsåker <ilmari(at)ilmari(dot)org> |
---|---|
To: | Justin Pryzby <pryzby(at)telsasoft(dot)com> |
Cc: | Andres Freund <andres(at)anarazel(dot)de>, pgsql-hackers(at)postgresql(dot)org, Peter Eisentraut <peter(at)eisentraut(dot)org> |
Subject: | Re: meson vs make: missing/inconsistent ENV |
Date: | 2023-02-27 12:30:59 |
Message-ID: | 87a60zjnng.fsf@wibble.ilmari.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Dagfinn Ilmari Mannsåker <ilmari(at)ilmari(dot)org> writes:
> Justin Pryzby <pryzby(at)telsasoft(dot)com> writes:
>
>> On Sun, Feb 26, 2023 at 03:21:04PM -0800, Andres Freund wrote:
>>> > Is there any consideration of promoting these or other warnings to
>>> > fatal?
>>>
>>> You mean the perl warnings?
>>
>> Yes - it'd be nice if the warnings caused an obvious failure to allow
>> addressing the issue. I noticed the icu warning while looking at a bug
>> in 0da243fed, and updating to add ZSTD.
>
> Perl warnings can be made fatal with `use warnings FATAL =>
> <categories>;`, but one should be careful about which categories to
> fatalise, per <https://metacpan.org/pod/warnings#Fatal-Warnings>.
>
> Some categories are inherently unsafe to fatalise, as documented in
> <https://metacpan.org/pod/strictures#CATEGORY-SELECTIONS>.
One disadvantage of making the warnings fatal is that it immediately
aborts the test. Another option would be to to turn warnings into test
failures, à la https://metacpan.org/pod/Test::Warnings or
https://metacpan.org/pod/Test::FailWarnings. Both those modules support
all the Perl versions we do, and have no non-core dependencies, but if
we don't want to add any more dependencies we can incorporate the logic
into one of our own testing modules.
- ilmari
From | Date | Subject | |
---|---|---|---|
Next Message | Andrew Dunstan | 2023-02-27 12:33:33 | Re: meson vs make: missing/inconsistent ENV |
Previous Message | Noel Grandin | 2023-02-27 12:14:10 | Re: how does postgresql handle LOB/CLOB/BLOB column data that dies before the query ends |