[Pkg-clamav-devel] Fwd: Bug in init script

[Scott Kitterman]

On Wednesday, July 02, 2014 20:57:04 Andreas Cadhalpun wrote:
> Hi Scott,
> 
> On 02.07.2014 19:52, Scott Kitterman wrote:
> > On July 2, 2014 12:30:06 PM EDT, Andreas Cadhalpun 
<andreas.cadhalpun at googlemail.com> wrote:
> >> On 02.07.2014 13:23, Sebastian Andrzej Siewior wrote:
> >>> On 2014-07-01 22:37:33 [+0200], Andreas Cadhalpun wrote:
> >>>> Assuming they would allow it, this would probably not make it into
> >> 
> >> 7.6,
> >> 
> >>>> which is coming during the next weekend.
> >>>> First we have to fix things in unstable and then wait for them to
> >> 
> >> transition
> >> 
> >>>> to testing.
> >>>> And then we would have to coordinate stable updates of all affected
> >>>> reverse-dependencies together with clamav...
> >>> 
> >>> Puh. The current clamav version shouldn't make it into stable as-is.
> >> 
> >> Indeed. As that API change is a serious bug, I think we should
> >> cherry-pick the SONAME bump and upload to experimental as soon as
> >> possible, so that the new package libclamav7 can go through binary-NEW.
> > 
> > Only if we're sure there won't be other ABI breaking changes in the next
> > release. We might be better off just to use breaks to manage the
> > transition.
> We can never be sure that there won't be more API changes, but I don't
> think any are planned. Anyway the API changed incompatibly and so the
> soversion should be bumped.
> If we use breaks now, we'll have to make another transition for the next
> upstream release, although probably the ABI won't change.
> And if there really is another change, we could also handle that
> transition with breaks.
> 
> So I don't see what would be better, if we postpone the soversion bump.

Fortunately this is probably OBE due to the change they just made.  I think we 
can fix this and get back to the old API/ABI, so no concerns for 0.98.4.

Scott K