Re: Release Cycle to low!

[Murray Cumming <murrayc murrayc com>]

On Wed, 2004-02-18 at 18:00, Ali Akcaagac wrote:
> On Thu, 2004-02-19 at 03:09 +1100, Jeff Waugh wrote:
> > GNOME Development Release 2.5.5
> 
> Sorry for this little compliant but don't you think you are throwing the
> releases like warm cakes now ? The 2.5.4 release is still quite HOT and
> yet 2.5.5 shows up 4 days after 2.5.4 ?
> 
> I think you should give people a break and have them breathe for a
> while. I mean when you go and release GNOME every say 2-3 weeks then
> it's enough time for people to compile the released Tarballs, test them
> and do serious bugreports.

Nonsense. You can't have too many releases. I'd be happy with releases
every half hour if it was possible.

> People do complain now because 'after a few days of painful compiling'
> they now got 2.5.4 compiled and were happy that they are somehow on the
> bleeding edge, they now realize that their NEW GNOME which they spent so
> much time into is old again.

It should be quite obvious that bug reports for week-old releases are
still relevant. If people feel no longer on the "cutting edge" then they
just need to calm down a bit.

> Well one can easily argue that the 'normal user' don't compile
> testreleases (as this was raised some mins in the #gnome-de IRC channel)
> but then the question from my side would be to whom are these released
> Tarballs meant if not the ordinary public ? This is a bit different if
> we update GNOME once per week with CVS and compile a new version out of
> it (what developers tend to do, so they always have the new stuff) or if
> you release Tarballs in really short cycles.
> 
> Assuming this:
> Now people start to test GNOME 2.5.4 find some bugs and report them to
> b.g.o and immediately they get told to update to 2.5.5 because chances
> are that they are fixed already.

But this isn't actually happening.

>  Hard to tell this to someone who just
> spent a couple of days getting 2.5.4 running halfway - who now needs to
> yet go through the pain of compiling 2.5.5 again and once done realizing
> that 2.5.6 is out yet 4 days after that.
> 
> Well I do believe my explaination lacks in some areas but I do believe
> that the one or other got the key message. I would like to ask whether
> the release cycle can be kept in a normal state say all 3-4 weeks
> otherwise it doesn't make sense pressing the developers to release
> Tarballs which no one can really test because of short release cycles.

I think the problems here are
- Mostly imagined.
- Mostly irrelevant.

Sorry for the bluntness, a bit.

-- 
Murray Cumming
www.murrayc.com
murrayc murrayc com