[arch-dev-public] Signoffs on simple version bumps (was: [signoff] man-pages 3.08-1)
belanger at ASTRO.UMontreal.CA
Wed Aug 27 15:27:49 EDT 2008
On Wed, 27 Aug 2008, Aaron Griffin wrote:
> On Wed, Aug 27, 2008 at 1:26 PM, Dan McGee <dpmcgee at gmail.com> wrote:
>> On Wed, Aug 27, 2008 at 1:22 PM, Aaron Griffin <aaronmgriffin at gmail.com> wrote:
>>> On Wed, Aug 27, 2008 at 1:03 PM, Aaron Griffin <aaronmgriffin at gmail.com> wrote:
>>>> On Wed, Aug 27, 2008 at 11:58 AM, Dan McGee <dpmcgee at gmail.com> wrote:
>>>>> On Wed, Aug 27, 2008 at 11:25 AM, Andreas Radke <a.radke at arcor.de> wrote:
>>>>>> too slow. now please signoff 3.08-1
>>>>> New proposal here. For packages that meet the following criteria (this
>>>>> was on-the-fly, I really don't think it needs to be set in stone):
>>>>> 1. Frequent releases (anything 2 weeks or quicker)
>>>>> 2. Little to no system impact if broken (initscripts would not fit the
>>>>> bill, for instance, but man-pages are not critical)
>>>>> 3. Can be easily verified by the maintainer to be working
>>>>> Can we skip the signoff procedure? The maintainer is of course free to
>>>>> still ask for it, but it would keep our pipeline from getting
>>>>> logjammed by stuff that really isn't in need of serious testing.
>>>>> Maybe just have list of packages in core we agree can be moved without
>>>>> signoffs, such as man-pages, tzdata, etc.
>>>> Yeah, I thought about this in the beginning, but it's easier to make a
>>>> blanket rule than to start adding exceptions.
>>>> I'm interested in hearing opinions on this.
>>> Dusty: Do you think it would be possible to add a DB table for
>>> packages excluded from signoffs? If we can do this, and integrate it
>>> with the web interface for signoffs, I can manage the actual contents
>>> of the table myself (or, well, anyone with django admin access can).
>> Wouldn't this be overkill? Sometimes a package would want signoffs,
>> other times you may not. And the signoff links don't show up for
>> packages not in testing anyway. And what happens when packages in
>> extra get thrown into testing? I guess it looks like a can of worms
>> not worth opening to me.
> That's a good point. You're talking about skipping testing altogether.
> Hmm, I'd have to think on this. Even though man-pages is not system
> critical, I like the concept of preserving the integrity of [core].
I wonder how many packages would fit in the 3 conditions suggested by Dan.
If only a few packages would be in the exception list, then it won't
really solve the logjam problem. So we might decide to still signoff all
core packages just to be consistent. I believe the logjam is mostly due
to packages that very few of us use (or know how to test) so the number of
potential testers is limited.
> I think, if we all start using the web-based signoffs a little more,
> this might become a quicker process. I was able to do a scan down the
> list and signoff on a few very quickly.
Is the web-based signoff testing completed? Then it should be used for
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
More information about the arch-dev-public