[DRAFT] Proposed release schedule for 1.9.x

Christian Stimming stimming at tuhh.de
Tue Jan 17 05:40:27 EST 2006


Hi all,

Derek Atkins wrote:
>> Hmm... I have a different concern.  I think we could basically start
>> rolling tarballs as soon as distcheck works, but I don't know if the
>> 2.0.0 date is reasonable or not.  I just don't have a feeling for what
>> the state of g2 is right now.  Probably we won't know until we start
>> getting wider testing.
> 
> I agree that the 2.0.0 release date needs to be flexible based on
> how well the 1.9.x releases are received.

I agree, too. The 2.0.0 release date as well as the number of 1.9.x 
releases of course will be adapted to the progress and to the number of 
outstanding issues.

>> So, I think we should release something soon.  I don't think we need
>> to advertise a rigid release schedule.  I think we should aim for at
>> least every three weeks, but we should feel no hesitation to release
>> *before* three weeks if we feel we need to.
> 
> Well, Jan 29 is "soon", is it not?  I agree we should release something
> soon, but we should also keep track of all the known issues..

My proposal basically gives the "time-based release" idea priority over 
"feature-based release". We could identify a small number of issues that 
*have* to be fixed before the first 1.9.0 (a.k.a. 1.9.0-blockers), but 
IMHO that must be *very* *small* number of issues.

http://bugzilla.gnome.org/show_bug.cgi?id=326578 is one of them.
The existence of the Orphan-USD account, OTOH, IMHO is not one of those 
1.9.0-blockers, although it's bothering me, too. Maybe a 2.0.0-blocker, 
but IMHO definitely not a 1.9.0-blocker.

What we do observe is that people start to sell SVN-snapshots as 
releases (like that gentoo thing), and this is a very bad development. 
The users of those snapshots will complain when they encounter temporary 
build system glitches, and we will complain if people complain about 
that. After all, SVN is SVN and should stay that way. I think the only 
way to discourage those medium-level-testers from using too much 
SVN-snapshots is to provide actual testing tarballs, and this is why I 
proposed a rather aggressive first 1.9.0 release and quick and numerous 
additional 1.9.x releases.

So, what are the actual blockers for 1.9.0? If we mark those with the 
"milestone=1.9.0" in bugzilla, then a query for all issues is
http://bugzilla.gnome.org/buglist.cgi?product=GnuCash&target_milestone=1.9.0
(added link to wiki page http://wiki.gnucash.org/wiki/Release_Schedule )

Developers, can you please add outstanding issues to bugzilla and mark 
them with milestone=1.9.0 if you think they need to be resolved before 
1.9.0? Thanks a lot.

Christian



More information about the gnucash-devel mailing list