I've confirmed with Mental, Bulia, Kees, Simarilius, and Ted.
We are now officially in Hard Freeze. We will create the 0.40 tagged
branch tomorrow evening.
Between now and then, doublecheck the packaging, proofread the
ReleaseNotes, and update the NEWS and AUTHORS are up to date.
Also, be certain to use each of the recently added added features (such
as layers), and report any major remaining glitches. It may be too late
to fix them, but the more we know the better.
On Sat, 27 Nov 2004, Bryce Harrington wrote:
We've held off on going into hard freeze for a few days, due to
troublesome bugs. However, I'd like to propose that we move ahead with
the hard freeze.
It's impossible to release a bug-free product. It's understandable and
commendable that we hold such a high set of standards for bug-free-ness,
but the fact is, we'll have more releases in the future, and we'll
_always_ have "one more thing to do..." If for some people, 0.40 won't
run, they can stick with 0.39, or wait for 0.41.
We've done a large amount of research into the bugs that exist; we
can provide a good, detailed 'known issue' list. Heck, maybe this'll
give some others inspiration and a headstart on figuring those problems
out in the coming weeks.
0.40 has been one of the most energy-intensive releases that Inkscape's
done so far. We've integrated a number of new dependencies, added
scores of new features, and fixed hundreds of bugs. I would bet folks
are getting worn out! With Christmas coming up, we ought to get the
release behind us, so folks can enjoy the time off without worrying
about Inkscape issues. This is a great release and I think the
community's going to really enjoy it.
As far as the remaining bugs go, here are some thoughts.
The infamous Windows open file and freeze error is the worst remaining
bug, however the current working hypothesis is that it's an issue with
libgc. It is entirely conceivable that the fix would go against libgc,
not inkscape, so may be entirely orthogonal to our release. However, if
indeed it is an inkscape problem, it is an intermittent bug, and doesn't
occur on every Windows user's box. As well, the use case for it is on
file open, rather than save or during work, so the chance of data loss
is small. I think we could simply document the problem and ensure users
are aware of it, and provide workarounds/solutions later as we find
The 'windows gets stuck in loop' error is a similarly annoying problem,
and does involve data loss. However, it sounds like it is even less
reproduceable than the previous one.
The other level 9 bugs do not sound very serious to me. Most appear to
be issues for a particular user running on a specific configuration
(e.g., with Japanese language encodings, or installing with the
autopackage snapshot, or when compiled with a particular version of gcc,
etc.) All of these need fixing, but I don't think they're critical
enough to warrant delaying the release.
I suggest we write the hypotheses for each level-9 bug, list them as
'known issues' in the release notes, and move forward with the hard
Bulia, Mental, Kees, Peter, Jon Cruz, Ishmal, Simarilius, and Ted, how
do you feel about moving forward with the release at this time?