Re: Clarification of NMU policy
On Mon, Oct 31, 2005 at 07:21:48PM +0100, Peter Eisentraut wrote:
> I have been in a discussion with a fellow developer about the exact
> meaning of the "0-day NMU policy" that is currently in effect.
Ok, I've tried to answer the questions as good as possible.
I think this aren't the important questions though. The important
- Did it fix the bug(s) it claimed to fix?
- Where these things really bugs?
- Did it not introduce any other bugs?
- Did the NMUer provide a full patch?
If the answer to all these questions is "yes", who cares wether the
NMU was done after 6 or 60 days after the bug was filed, or wether
the NMUer did send the patch 30 minutes before or after the upload,
or if bug had the severity serious or important (e.g. if it was a
FTBFS on i386 or kfreebsd-i386), etc...?
AFAICT discussions about NMUs have always destroyed way more productivity
than NMUs themself ;)
Of course all rules only apply in absent of an opinion of the
> 1. Does the 0-day policy only apply to RC bugs or to all bugs?
Strictly speaking only RC bugs, though there are of course classes
of bugs you can (and should) fix at the same time. E.g. things like
"FTBFS on (BSD/HURD/...)".
All other bugs can be NMU-worthy in certain cases but they most
probably don't suffer from a few days more delay.
> 2. Does the "0 day" apply to the delay after the upload or to the delay
> between the submission of the bug and the upload?
Delay after the upload. For myself I have a 7-day rule between the
bug filing and the upload. I think that was an usual time frame during
the sarge release process, however I have no written references about
it right now.
There are of course reasons to break such rules but I do that on a
case-by-case basis only.
> 3. Does any rule, policy, or BSP excuse people from following the NMU
> protocol in the developer's reference, for example, the requirement to
> submit a bug before the upload?
A BSP excuses you from not announcing the NMU to the maintainer
prior to the upload (the BSP announcement serves as a surrogate).
The usual NMU delay can also be reduced during a BSP (but reducing
a 0 delay is obviously difficult ;)).
A BSP certainly doesn't excuse doing NMUs for bugs that are not filed
or not sending a patch to the BTS. The question wether an NMU for a bug
that was open for some hours is ok depends on the answer to question 2.
There can be made special rules for NMUs for other reasons, see the
announcement of the CXX transition for an example.
Frank Lichtenheld <email@example.com>