Re: How to determine the urgency of an upload
Goswin von Brederlow <brederlo@informatik.uni-tuebingen.de> wrote:
> joshk@triplehelix.org (Joshua Kwan) writes:
>> On Mon, Jan 12, 2004 at 05:19:50PM +0100, Tobias Toedter wrote:
>>> I think it's clear to set the urgency to low for almost every purpose --
>>> except for security updates, I guess that should be urgency=high.
[...]
> Most of the time with non trivial Build-Depends chains you have to
> wait for the slower archs to compile the package longer than the "low"
> or "medium" period anyway. Currently I don't think you can get a
> package autobuild within 2 days at all.
[...]
>> it with urgency medium or high so it gets into testing within 5 or 2
>> days respectively. Same reasoning if you have just fixed a particularly
>> nasty bug.
>> Of course, urgency != low should be used sparingly to allow the unstable
>> users to test the new packages and to not throw off the whole transition-
>> to-sarge system.
> Personally I feel most comfortable with:
> high: security update
> medium: RC bugs (especially adding a Build-Depends and similar trivial fixes)
> low: everything else
> But thats just me.
[...]
I do think it would also be perfectly sane to upload something with
only trivial changes with urgency high, if there is a nice window for
testing propagation currently, bacause the chances for introducing new
breakage by trivial changes are small.
E.g. if I had a version that was already waiting 9 days and was not
inhibited by dependencies currently I would not think it to be
unreasonable to upload a version that _only_ added three updated
debconf-translations with urgency high.
cu andreas
--
Hey, da ist ein Ballonautomat auf der Toilette!
Unofficial _Debian-packages_ of latest unstable _tin_
http://www.logic.univie.ac.at/~ametzler/debian/tin-snapshot/
Reply to: