Re: RFC: Debian Quality Assurance Group
On 23 Apr, tho@thomsen.isdn.cs.tu-berlin.de wrote:
> In message <[🔎] 19990422220636.C16449@cibalia.gkvk.hr>, Josip Rodin writes:
>>
>> In the normal development period (when only unstable and stable
>> distributions exist), delays are as follows:
>> * fix for a critical/grave bug: 2 days
>> * fix for any kind of security or important bug: 7 days
>> * fix for a bug of normal severity: 30 days
>> * cosmetic fix or a wish implementation: 45 days
>> Note: the second two types may only be uploaded to unstable.
>>
>> The above delays are reduced by a factor of 2 in the month preceding a
>> freeze. During a freeze (when there is a stable, frozen, and unstable
>> distribution), delays are as follows:
>> * fix for a security bug: 1 day
>> * any other bug fix: 7 days
>
> Well, well, but how are those (ambitious) periods to be archieved? Do
> you want to chain QA-group member to their worstations? Are they fired,
> if they fail? Without some means to guarantee such periods, they are
> just hot air and not very serious. IMHO there is no possibility to give
> deadlines for fixing bugs; they might be arbitrarily complex. Another
> possibility would be, to provide a work-around, not fix, for example, to
> fall back to an old, proven package, if the current can't be fixed in
> a given period.
I interpreted the passage as a set of minimum times to wait before
doing an NMU, not as an absolute requirement for bug fixes. It might
take a month for QA to get to a bug, but once the bug is fixed by a QA
member, (s)he must wait for X days before trumping the regular
maintainer and doing an NMU.
The idea is to provide deadlines for the maintainers, not for QA, and
for allowing QA to do its job without pissing off the rest of the
maintainers.
Reply to: