[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: pkgsrc-2013Q4 freeze started
On Fri, Dec 20, 2013 at 01:13:16AM +0900, Izumi Tsutsui wrote:
> > 2-3 days is nowhere near long enough to perform even an update bulk
> > build on a slow to medium platform, get the results and be able to act
> > upon them before the freeze.
> I'm asking about update of leaf packages.
> Is it so serious on bulk build?
No; that's the point of talking about leaf packages at all.
But before we go around the bend trying to make complicated rules to
use in place of thinking:
1. Part of the reason for having a freeze is to allow running bulk
builds, and then rerunning them as fixes go in to check those fixes,
without spending days waiting for the resultsn of someone touching
e.g. perl or libpng.
2. The other part of the reason for having a freeze is to sort out
problems so that the stable branch doesn't have those problems.
Any commit that impedes either of these goals should wait until after
the freeze, regardless of what the rules might specifically say.
Likewise, any commit that advances these goals is ok.
Commits that trade off one goal for the other or that risk impeding
either goal should be run by pkgsrc-pmc for a decision. This is not
hard to do and doesn't usually incur a large response latency; when in
As I recall, the argument last time was caused by updating firefox
without notice; firefox, though a leaf package, is one of a handful of
highly visible important packages that would embarrass the project if
it were broken on a stable branch; thus updating it risks goal #2 and
should be discussed beforehand. As I recall the logic that prompted
updating it was perfectly reasonable; it would just have been better
to have that discussion before than after.
David A. Holland
Main Index |
Thread Index |