On 8/6/26 2:57 PM, Eric Covener wrote:
> On Thu, Aug 6, 2026 at 8:39 AM Branko Čibej <[email protected]> wrote:
>>
>> On 6. 8. 2026 14:28, Eric Covener wrote:
>>
>> On Thu, Aug 6, 2026 at 7:00 AM Ruediger Pluem <[email protected]> wrote:
>>
>> Please hold a sec. I guess we broke the Oracle DBD driver. Details to follow.
>>
>> Release is published but not announced (stuck on release automation issues).
>>
>> I'm not sure what's best at this stage, but I think the version number
>> is burnt so we should do a followup release (if needed). The oracle
>> changes were just lifted from elsewhere which was indicated in the
>> original report.
>>
>> I am inclined to continue announcing rather than leaving in limbo.
>>
>>
>> If the Oracle DBD driver is broken and this is verifies, then this is a
>> regression. It's fine to retire a release before it has been announced, fix
>> the bug, roll a new one and update the release notes and announcement to
>> state that the apr-util-1.6.4 was retired. We can remove the artefacts from
>> the servers and the archive to prevent users from inadvertently downloading
>> it.
>>
>> I don't think we should release a version with a known regression.
>
> Sorry, too late. Small saving grace, it's a build-time failure (when
> this feature's dependencies are available)
I meanwhile tested the patch r1936912 against Oracle Instant Client 19.31 and
added the two tweaks
r1936917 and r1936918. All is backported down to 1.6.x and hence I guess
releasing 1.6.5 should be possible quickly.
It just misses a CHANGES entry over 1.6.4 that the brokenness of the Oracle DBD
driver was fixed.
Regards
Rüdiger
