[Mono-dev] [Ximian-mono-list] 2 Mono releases a year...
sebastien.pouliot at gmail.com
Mon Dec 8 20:31:22 EST 2008
On Mon, 2008-12-08 at 17:27 -0700, Andrew Jorgensen wrote:
> We have customers. When you say "leave some of the QA work to our
> users" you're really saying "leave some of the QA work to our paying
> customers". This is not an option. We want to improve the quality of
> Mono, not diminish it.
That's a very good point. However this does not explain how having less
releases will get us less bugs for users or even customers.
Customers are mainly on SLE which we do not release very often. In this
case investing more time in the release leading to SLE make a lot of
sense (e.g. by skipping a normal release).
Just like having closer Mono releases between SLED releases make sense
too. Why ? because shorter releases means we, as developers, don't have
time to break subtle things without the feedback from either QA or the
community. If Mono is more stable when we start a SLE cycle then we are
doing our customers a great service :)
> Also some of that time is build and installer maintenance, and
> appliance build, etc. etc..
> Also it doesn't make a lot of sense to release more often than the
> major distributions do.
True if all distros were released simultaneously - but they are not.
This means some distros would ship old Mono releases*, which means
recent Mono applications won't work on those recent distros. Neither
users or developers wins in this situation.
* yep I know this already happens but it won't get any better with
> Our point (our meaning the provo team) is that we can't actually move
> forward at the rate we're going with the resources we have. We can't
> even keep up. There are a number of things on our eternal to do lists
> that we likely will never do. And each release it seems that one of
> those items becomes something that we absolutely must do.
I'm not saying the current plan is perfect, at least not with the number
of steps involved. However I suspect less releases will just make the
problem bigger, but seen/felt less often.
I think there are other options that would be less extreme (than the
current plan or 2 releases/year). E.g.
* Longer, full QA (current style)
* One per year (or one per SLE)
* No/less other tasks
* Quicker, less QA, more user involvement (Zoltan's style ;-)
* 2-3 per year
* Some (but never enough) time to clean up todo list
my 2 canuck cents
> >>> "Zoltan Varga" <vargaz at gmail.com> 12/08/08 5:02 PM >>>
> I would suggest the opposite: do releases more frequently, and leave
> some of the QA work
> to our users, as is traditionally done by an OSS project.
> On Tue, Dec 9, 2008 at 12:55 AM, Thomas Wiest <twiest at novell.com>
> > Hey Guys,
> > The build and test team is struggling to keep up with all of our
> > obligations (and our task lists just keep growing). The only way I
> > see being able to get ahead is to reduce the number of Mono releases
> > do a year. Right now we've been trying to do 4 releases a year (it
> > like it's going to be closer to 3 in the short term).
> > I propose that we start doing only 2 Mono releases a year. This
> > allow the build and test team a lot more time in between releases to
> > keep up with the tasks we've been asked to do.
> > So, the question is, if we start doing 2 releases a year, what are
> > negative affects of this?
> > Since the major distros and gnome only have 2 releases a year also,
> > would think the impact would be fairly minimal. Am I wrong?
> > Thomas
> > _______________________________________________
> > Ximian-mono-list mailing list
> > Ximian-mono-list at lists.ximian.com
> > http://lists.ximian.com/mailman/listinfo/ximian-mono-list
> Ximian-mono-list mailing list
> Ximian-mono-list at lists.ximian.com
> Mono-devel-list mailing list
> Mono-devel-list at lists.ximian.com
More information about the Mono-devel-list