[DRE-maint] Bug#1027339: metadata

[Santiago Vila]

El 15/12/24 a las 6:24, Paul Gevers escribió:
> Hi Santiago,
> 
> On Sun, 24 Nov 2024 00:21:59 +0100 Santiago Vila <sanvila at debian.org> wrote:
>> # the bug is "not having a version high enough" so it's present in trixie and sid
>> found 1027339 2.2.7-1.1
> 
> Care to elaborate? The package builds fine on r-b infrastructure, so it isn't obvious what you mean here.
> 
> Did you maybe mean that ruby-rackup (which this bug is no longer filed against) is still buggy?

The bug was archived prematurely, and this was my attempt at avoiding that to happen again.

The bug was initially filed against ruby-rackup, which did not build from source.

Later, it was determined that the blame for that problem belongs to ruby-rack, for
it having a version which was not high enough for ruby-rackup to be buildable, and
for that reason the bug was reassigned to ruby-rack.

The bug was declared fixed in version 3.0.0-1 of ruby-rack, which was only
present in experimental, and because the bug was also closed, it was archived
after some time.

But in my opinion such archive (which is a consequence of the closing) does not help because it hides problems that we currently face in unstable, namely, that we still can't build ruby-rackup in unstable because the version of ruby-rack is not high enough.

So, if we agree that the real bug is in ruby-rackup, then IMO we should also agree that the present version in unstable has the bug, as it causes ruby-rackup not to be buildable in unstable.

Note: I'm also Cc:ing Andreas, to whom I sometimes ask for advice/help about this sort of things regarding version tracking.

If not this "found" command I issued, what else could I have done to avoid this premature archive to happen again?

I personally would have reopened the bug (keeping the fact that it's fixed in 3.0.0-1) but some people are so attached to version tracking ("if it's fixed it should be closed!") that they would surely complain.

I can live with FTBFS bugs in stable being closed because they were already fixed in trixie/sid. I still don't like that because a closed bug creates a sense that "there is nothing more to do", which I don't think it's appropriate for FTBFS bugs in stable.

But closing bugs in experimental when they are still present in unstable seems actively harmful to me.

Thanks.