GCC 3.3 Status

[S. Bosscher]

Mark,

You already have downgraded quite a few PRs, and now you're about to
downgrade a few more.  I think that for 3.3 it is the right decision (heck,
it's got to go out now, can't wait forever), but is it your intention to
upgrade those PRs again once 3.3 is released?

Many PRs against 3.3 have been simply ignored, or were not recognised as
regressions, for some time (until Wolfgang started his Great Work of
analysing all those PRs).  Most of them are still present on the trunk.  And
they _are_ regressions, even if everybody agrees that they should not hold
up 3.3 any longer.

Greetz
Steven





-----Original Message-----
From: Mark Mitchell
To: gcc@gcc.gnu.org
Cc: rearnshaw@arm.com; obrien@FreeBSD.org; geoffk@geoffk.org;
dje@watson.ibm.com; rth@redhat.com; davem@redhat.com; jakub@redhat.com;
joern.rennecke@superh.com; aoliva@redhat.com; per@bothner.com;
apbianco@redhat.com; aph@redhat.com; jh@suse.cz
Sent: 22-4-03 23:40
Subject: GCC 3.3 Status


If you are in the CC list for this email, please read this entire
message.

As of now, we've got 32 high-priority PRs against GCC 3.3.

I'm going to downgrade some of these, including doc issues,
"misleading" (but correct) error messages, etc.  In fact, if we do not
get closure quickly, it is my intention (assuming the SC does not
override my decision) to ship 3.3 with virtually *all* of these PRs
unfixed.  Very few open 3.3 PRs affect primary targets, and of those
that do, most are ice-on-illegal rather than wrong-code, and of those
that are wrong-code, nobody seems to be making any progress on fixing
them.

I intend to personally tackle most of the remaining C++ issues,
including trying to make a little more progress on the compile-time
issues with respect to inlining, probably by doing the appropriate
double-counting of CLEANUP_STMTs.

Below is a list of PRs that are apparently port-specific, together
with the maintainers for that port.  Maintainers, please let me know
that (a) you intend to tackle these PRs in the next few days (by
assigning them to yourself) or at least (b) that you are not going to
try to fix them, by sending mail to me.  Please do either (a) or (b)
so that I know what we have a chance to fix and what we do not.

If you choose (b), GCC 3.3 will probably ship without a fix for the
PR.

If you can, please spend a little time to fix up just one PR; if each
port maintainer below does that, we'll nail almost all of them.

It is my intention to make the first GCC 3.3 prerelease in the early
part of next week.  The branch will close to patches that do not have
my approval at that time.

Here is the list of open PRs:

Geoff Keating, David Edelsohn:

  PR 9745, 10315: PowerPC

Richard Earnshaw:

  PR 6860, PR 10206: ARM

David O'Brien, Richard Henderson, David Miller, Jakub Jelinek:

  PR 10453: SPARC/FreeBSD 

Richard Henderson, David Miller, Jakub Jelinek:

  PR 8300, PR 10160: SPARC

Jeff Law:

  PR 10021, PR 9812: m68k

Joern Rennecke, Alexandre Oliva:

  PR 9594: SH
  
Richard Henderson:

  PR 10308: x86

Per Bothner, Alexandre Petit-Biano, Andrew Haley:

  PR 10353: Java

Jan Hubicka:

  PR 9929: x86 assigned to you
  
--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com