[arch-general] [testing] protocol - {WAS: [arch-dev-public] WARNING - glibc-2.13-2}

[Seblu]

On Mon, Feb 7, 2011 at 11:37 AM, Allan McRae <allan at archlinux.org> wrote:
> On 07/02/11 20:26, Seblu wrote:
>>
>> 2011/2/6 Allan McRae<allan at archlinux.org>:
>>>
>>> On 06/02/11 20:16, Ng Oon-Ee wrote:
>>>>
>>>> On Sun, 2011-02-06 at 20:10 +1000, Allan McRae wrote:
>>>>>
>>>>> I have removed the glibc-2.13-2 package until a fix is committed to
>>>>> upstream git.  All users should downgrade to the version in [core].
>>>>>
>>>>> Allan
>>>>
>>>> While I understand why Allan did this, shouldn't this be unnecessary for
>>>> the [testing] repo? Those who use it know what to do, after all....
>>>
>>> Probably.  And if the breakage was more minor, I would have left it there
>>> (e.g. what has been done with the make package).   But given I still do
>>> not
>>> 100% understand the bug and the there is no real schedule for an upstream
>>> fix, I thought it best to just make everybody's life easier and remove
>>> the
>>> package for the time being.
>>
>> Why not just remove the patch increment pkgrel and let pacman -Su do the
>> rest?
>>
>
> Why?  User of [testing] should see the message on arch-dev-public and be
> competent enough to downgrade.
Why loop. Hum. I wondered what was the best thing to do in this case
and I would not have done it right the first time.

I naively pushed back an earlier packet with a new release version and
report my error.
Because, this allow every already updated users to smoothly upgrade
without questioning himself about my regression. But causing all
others users to upgrade to same package at next upgrade. This is
transparent for everybody, save dev focusing, and save some time.

So, I was wondering if I ignore some rules or if I do not see
something that we should not do this.

-- 
Sébastien Luttringer
www.seblu.net