Subject: The next release & how you can help
To: None <email@example.com, firstname.lastname@example.org>
From: Allen Briggs <email@example.com>
Date: 03/09/2004 02:04:24
NetBSD is moving toward another release. You can help make it the
best release yet by donating some of your time and attention to it.
The list below is a set of fairly 'rough notes' about what things
need to be done and which platforms are affected. We're developing
the exact schedule now, and would like to address as many of these
items as possible before we branch.
How can you help? Try the latest snapshots out. Look through the
documentation. Submit PRs with any changes. If you have submitted
PRs in the past, take a little time to see if they're still applicable
or if they've been "stealth-fixed". Either way, please update the
PRs to make the status current.
In fact, before the branch is made, we're planning on a period of
time where new features will be delayed and developers can concentrate
on fixing bugs and ensuring that things are in good shape for the branch.
Get involved and help us make this the best release yet!
Name: GNATS audit
Issue: Technically, we should have no open PRs in the high
or medium states, going into a release cycle:
"high" -- A solution is needed ASAP
"medium" -- The problem should be solved in the next release
We should at least handle the high ones by solving
or reprioritizing (appropriately). Current numbers from
the excellent summary page:
high: 588 (up from 564)
medium: 1749 (up from 1632)
Impact: Another release with scads of known/reported bugs.
Issue: Not supported yet.
Fixes: Implement it.
vax: Code imported, but does not work yet. ???
Issue: Does not handle the new signal trampoline
Impact: Cannot debug through signal handlers
Fixes: Add a selector function like the svr4 code that can
descern between the old and the new signal trampoline
Issue: Longjmp uses deprecated __sigreturn14.
Impact: Programs that use longjmp will not work unless COMPAT_16
Affected alpha, hppa, m68k, mips, sh5, sparc, sparc64, vax
Fixes: Change all signal trampolines to save a ucontext style
jmp_buf, and then use setcontext instead of __sigreturn14
Issue: The X server is not yet fully cross-buildable
Impact: Releng can not get the X sets built automatically without
this support in place
Affected: atari, sun3
Fixes: Generate the Makefiles
What state are we in now? It looks like the only two servers that
we ship in 1.6.2 that aren't in src/x11/Xserver/Makefile.common are
atari and sun3. Everything else looks OK? What's been tested?
Use NetBSD! http://www.netbsd.org/