Re: groff 1.23.0 stability (was: using the TQ macro)

[Sam James <sam@gentoo.org>]

Alejandro Colomar <alx@kernel.org> writes:

> [[PGP Signed Part:Undecided]]
> Hi Sam!
>
> On Tue, Oct 31, 2023 at 04:38:13AM +0000, Sam James wrote:
>> "G. Branden Robinson" <g.branden.robinson@gmail.com> writes:
>> > At 2023-10-25T17:08:19+0200, Alejandro Colomar wrote:
>> >> BTW, I just checked and Gentoo still doesn't consider 1.23.0 stable
>> >> enough <https://packages.gentoo.org/packages/sys-apps/groff>.  :|
>> >
>> 
>> Alex, this is based on a misunderstanding of how our process works -- please
>> CC me if you have questions or if something looks off in future, so I
>> can explain/help if required.
>> 
>> > I don't understand that claim.  1.23.x is as stable as it can be; there
>> > have been no point releases.  Its behavior is not changing based on the
>> > calendar.
>> 
>> The standard rule in Gentoo is 30 days after something has been released
>> before it's considered for "stabilisation". We wait longer for critical
>> packages like groff to give more time for any reported bugs in "~arch"
>> (our testing area, which a lot of users participate in). It is generally
>> not a comment on upstream stability at all.
>
> Yep, I understand it's just about your use in combination with other
> packages in your distribution.  What I'm not sure is if by default
> Gentoo installs the stable packages or the testing ones.  If you install
> by default the stable one, I wouldn't want to force a dependency on a
> package that you don't yet install by default.

That's no problem - we regularly have things which require a new
dependency to become stable and it's a nudge if it hasn't happened anyway.

(See below).

>
>> 
>> > I have to assume that there are either changes since 1.22.4
>> > documented in NEWS (and if not, that's probably a bug) that they're
>> > concerned about, or they're worried the broader community hasn't gotten
>> > enough exposure to it yet.  repology.org has been sitting at 64
>> > instances of groff 1.23.0 for weeks now; I think pretty much everyone
>> > who's going to adopt it has done so by now.
>> >
>> 
>> ... in this case, the only blockers were really:
>> * me having https://github.com/Perl/perl5/issues/21239
>>   in the back of my head (wasn't paying full attention, just knew I had
>>   to go back and read any developments/further comments) 
>> 
>> * needing to look into a reported failure
>>   (https://bugs.gentoo.org/910226) - which looks like it should be fixed
>>   when we update our version of openvswitch (or we backport the patch,
>>   or both)
>
> So, if the Linux man-pages forces a dependency of groff-1.23.0, would it
> be problematic for Gentoo before you declare it stable, or would it be
> fine?

Yeah, this is fine - go ahead. The only issue would really be if
groff-1.23.0 was causing many issues which would prevent us from
unleashing newer man-pages any time soon, but that is not the case.

>
>> 
>> > CCing Sam James (the only Gentoo developer I know by name, because he's
>> > been active some of the same places I have been) in case he can throw
>> > some light on this.
>> 
>> Happily! Please feel free to loop me in if you reckon I can give input
>> on things.
>> 
>> So, all in all, none of this is a reflection on upstream, just a mix
>> of: how we do things normally (waiting a bit post-release unless there's
>> some serious regression in our stable version), waiting a bit longer
>> because it's a critical package (sometimes 60 days, sometimes a bit
>> longer), and not getting around to looking at that openvswitch bug yet.
>
> Yeah, the quality of groff-1.23.0 is way better than 1.22.4.  I'm just
> worried that forcing distros to use it too early might be detrimental.
>
> Cheers,
> Alex
>
>> 
>> I promise I would report any problems if I determined they were in any
>> way an upstream issue :)
>> 
>> Thanks for reaching out.
>> 
>> >
>> > Regards,
>> > Branden
>> >
>> 
>> best,
>> sam
>> 

thanks,
sam