Regressions: the enemy of 1.0

[Dan Kegel]

There's been lots of progress in Wine lately; many
important applications work better than ever.
But the inevitable cost of such progress is that
some applications are now broken and/or
wine has odd problems.

For instance, as of the last two versions of Wine,
StarUML can't be unminimized
( http://bugs.winehq.org/show_bug.cgi?id=12697 ),
and as of 0.9.60, the IME support seems to be
making wine apps run dog slow on at least one
of my machines
( http://www.winehq.org/pipermail/wine-devel/2008-April/064999.html ).
These seem like regressions that ought to be fixed
before 1.0.

So... I guess my point is: the code freeze is starting soon,
which is good; that will help coax developers into looking
for and fixing recently introduced bugs rather than old ones.
i.e. fight to hang on to what we've got (or had recently)
rather than pushing to conquer new ground.

Those who absolutely have to work on risky new stuff
in the meantime will increasingly have to keep their
work outside the main tree for now.
Once 1.0 is out, those patches can land, and
the fight for new ground can resume.