What's keeping us from getting 2.8 out?
Hi there,
I guess I've been a monumental pain to some in the past weeks with all
my nagging, but somehow it feels as if we're afraid of getting 2.8 out
of the door.
I'm not sure how much Enselic's list of remaining tasks matches reality
(not sure how it gets decided what tasks get added there and when
they're considered done -- I guess it feeds off the 2.8 milestone in
Bugzilla?), but if it's not completely off track (I don't think it is),
development (or at least "ticking off tasks") has been stagnant since
the end of 2011.
To get out of this (at least perceived) state of gridlock, we should
have some criteria with which to decide what must be done for 2.8 to
happen, so we have a list which we can simply tick off one after
another.
These are the criteria I would use, with a bit of guessing where we are
regarding them interspersed:
- Finish release notes and related docs:
- Document changed and new API -- anyone other than Mitch who
can do this? I guess every bit counts but somebody knowledgeable
enough needs to say "yes, this should be about all".
- http://developer.gimp.org/api/2.0/libgimp/libgimp-index-new-in-2-8.html
doesn't exist yet, ditto.
- Some screenshots are still missing.
- Anything else?
- Get out 2.7.5. Only I would call it 2.8-RC, if only to keep us on our
toes :-).
- Fix milestone 2.8 bugs which are either critical or would need
incompatible API/ABI changes for being fixed, or push the latter to a
later milestone if the bug is something with which we can live for the
duration of 2.8. This needs a bit of triaging I suppose -- I would do
that, but would probably err on the side of pushing to 2.10/3.0 :-).
- Once we're done code-wise:
- Tag 2.8.0
- Roll a tarball, publish it.
- Get binary builds for Windows/Mac done? That wouldn't need to
hold up the release IMO however.
- Post stories how we've beat Duke Nukem Forever by years :-D to
Freshmeat or however that's called today, libregraphicsworld.org
and the other usual suspects.
So that's my take on it. Comments?
Other than that, I have some (possibly heretical) ideas of how to get
future versions out quicker, but that's stuff for another post (and
should probably be discussed after 2.8 is out -- we don't need any more
distractions right now).
Nils
