Re: [reSIProcate] ClientSubscriptionHandler::onNewSubscription regression
Hi Francis,
I already cherry-picked that fix into the 1.8 branch before - is there
anything else I should have changed or should it be OK as-is?
Regards,
Daniel
On 05/08/13 22:25, Francis Joanis wrote:
> Hi Daniel,
>
> The proper fix will most likely have to wait for later (e.g. September) so we
> could aim to do another 1.8.x release tarball tomorrow. I will integrate my
> fix onto the resiprocate-1.8 branch later today and it will be ready for the
> "respin" tomorrow.
>
> I'll let you know as soon as it's done and thank you for your help with this.
>
> Thanks,
> Francis
>
> On 2013-08-05, at 12:23 PM, Daniel Pocock <daniel@xxxxxxxxxxxxx> wrote:
>
>> On 03/08/13 03:23, Francis Joanis wrote:
>>> Hi,
>>>
>>> A regression was found (thanks to Jeremy Geras for reporting it) under
>>> which ClientSubscriptionHandler::onNewSubscription isn't called for
>>> the normal SUBSCRIBE/202/NOTIFY use case.
>>>
>>> The bug was introduced in revision 10128 as an attempt to fix another,
>>> less common, bug that prevented onNewSubscription from being called
>>> when failures happened. More details can be obtained here:
>>> http://list.resiprocate.org/archive/resiprocate-devel/thrd83.html#08362
>>>
>>> The code changed in revision 10128 has been reverted on the mainline
>>> until a proper fix for the latter bug can be identified.
>>
>> Thanks for clarifying this - I was going to make up another release
>> tarball today or tomorrow, do you believe the revised fix will be
>> available shortly, should I defer the release to allow more time for
>> this, or is it something that can wait until a future release (e.g. in
>> September)?
>>
>