Re: Bug#796345: Status report on perl 5.22 transition readiness (30th Sept)

[Emilio Pozuelo Monfort <pochu@debian.org>]

On 03/12/15 13:15, Dominic Hargreaves wrote:
> On Wed, Dec 02, 2015 at 07:08:55PM +0100, Emilio Pozuelo Monfort wrote:
>> On 02/12/15 17:48, Dominic Hargreaves wrote:
>>> On Wed, Dec 02, 2015 at 11:09:09AM +0000, Dominic Hargreaves wrote:
>>>> On Wed, Dec 02, 2015 at 01:24:29AM +0100, Axel Beckert wrote:
>>>>> Hi,
>>>>>
>>>>> Emilio Pozuelo Monfort wrote:
>>>>>> On 30/10/15 14:34, Emilio Pozuelo Monfort wrote:
>>>>>>> That'd only leave us with the apache bug.
>>>>>>
>>>>>> There's a patch available for that now, right?
>>>>>
>>>>> Yes. It has been included in the upload to experimental 1.5 days ago:
>>>>> https://packages.qa.debian.org/liba/libapache2-mod-perl2/news/20151130T194855Z.html
>>>>
>>>> I will run some test builds with perl 5.22, that package, and the packages
>>>> build-depending on libapache2-mod-perl2 over the next day or so. Then
>>>> hopefully we can really get this transition under way!
>>>
>>> I've tested the packages which were blocked by libapache2-mod-perl2
>>> today, and filed two new bug reports, against libembperl-perl[1] and
>>> libapache-gallery-perl[2]. The former unfortunately has a history of
>>> breaking with new perl releases and fixes may not be forthcoming;
>>> it also has a low and diminishing popcon, so I think at this stage we
>>> should not let it block our transition.
>>>
>>> The latter is a trivial fix (and does not block the transition as
>>> it's an arch:all package); I expect it will be fixed either by the
>>> maintainer, or by NMU, soon.
>>>
>>> I will try some real world testing with the new libapache2-mod-perl2
>>> package in sid/perl 5.20 later this week, and then I think we can plan
>>> to go ahead with the transition after that - as soon as this weekend
>>> if other ongoing transitions allow?
>>
>> Yeah, that's probably fine. Let us know how your tests go.
> 
> Niko reminded me that 5.22.1 is due out as soon as the weekend, and
> it would make sense to transition with that rather than have to 
> build a mini-transition in later. So we'll work to integrate that
> into experimental with suitable QA before the transition, if that's
> okay with you. I think that should only delay things by a couple of
> days.

Sure, that sounds better indeed. Let's do that.

Cheers,
Emilio