[debian-mysql] Bug#416145: Replication failure with auto-increment and on duplicate key update

[Alan Tam]

Christian Hammers wrote:
> So if I understand correctly, the version in etch is not buggy in a sense
> that it causes trouble when people only use etch<->etch replication and
> patching this minor bug would require some tricks?
>
> In this case I would rather stay with the recommendation to only use the very
> same version in replication and if people need a newer version they should
> use backports.org on both sides.
>
> I leave this bug open for reference though.
>   
I agree with this resolution since now etch is released. Otherwise we
may create even more confusions.

For the record, fixing the bug only needs to apply a two-liner diff. The
other portions of the upstream diff is used for validating (and warning)
that the master and slave are having compatible versions, i.e. none of
them falls into the trap-period that the etch version is in.

-- 
Regards,
Alan