On 01/24/2020 09:28 AM, Joe Orton wrote:
> I'd like to catch empty APLOGNO() use in CI, I think this should be an
> error and IIRC it's tripped us up in 2.4 releases in the past. Two ways
> to do this are:
>
> 1. For #ifdef AP_DEBUG, or for some other special-case, catch and fail
> at compile-time for empty APLOGNO(). I think this should be possible
> with some preprocessor tricks.
>
> 2. Have a special CI job which runs "make update-log-msg-tags", treats
> duplicates as errors and fails if there any newly-filled APLOGNO tags
>
> I think my preference is (2). During development it's easier to leave
> APLOGNO() empty and fill it in at the last minute, so having this caught
> in Travis is sufficient and should prevent errors getting as far as
> 2.4.x backports.
>
> Any other opinions, or an alternative method?
+1 on 2.
Regards
Rüdiger
