Is boost 1.46.1 in rawhide for real?

[Nicolas Chauvet]

2011/4/6 Petr Machata <pmachata at redhat.com>:
> Bruno Wolff III <bruno at wolff.to> writes:
>
>> I don't remember seeing a soname bump announcement for boost and since
>> for branched it went from 1.46.0 to 1.46.1 and then back to 1.46.0, I don't
>> want to start rebuilding stuff if this is going to happen in rawhide too.
>
> It's not our plan te revert this in rawhide, but we plan to rebase again
> later in the cycle, probably for 1.48.  I see that nothing has been
> rebuilt against 1.46.1 yet, so what I could do is keep the patchlevel
> and just drop the SONAME back to 1.46.0.  The changes between 1.46.0 and
> 1.46.1 _should_ be safe--not quite safe enough for pushing to F15, in my
> opinion, but rawhide has seen worse.  That way boost users shouldn't
> have to rebuild twice.

I personnally have a bad experience about reverting ABI version name
against upstream, specially in the case of DirectFB
https://bugzilla.redhat.com/show_bug.cgi?id=673842

The best is to continue to try to educate upstream about this problem,
specially if this update 'sounds like' a bugfix update. I would be
more in favour to have a dedicated dist-tag used for the rebuilt of
dependent packages (done preliminary in rawhide) and not to create a
SONAME deviating from upstream.

Nicolas (kwizart)