This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: gcc development schedule [Re: sharing libcpp between GDB and GCC]


On Wed, Mar 27, 2002 at 01:17:22PM +0000, Joseph S. Myers wrote:
> On Tue, 26 Mar 2002, Tom Lord wrote:
> 
> > lower cost for all participants?  Certainly put GCC in the position
> > of being better able to make near-instant "emergency releases" to correct
> > defects that escape up-front testing?  Certainly avoid snafus like
> 
> We can make such releases.  The 3.0 branch is or should be in a state of
> perpetual release-readiness, and that's where an emergency release would
> come from.  (If necessary, a new branch could be created off 3.0.4 for a
> 3.0.4.1 release, as was done with the stillborn 2.95.2.1 release.)

Some words from someone outside the active gcc development:

I think there should be much much more releases on the stable/old-stable
branches.  The worst example is gcc 2.95.x:  In the last official
release there are known major bugs that cause well known free software
projects either to not work at all (glibc) or fail in more or less
subtile ways (linux kernel or parts thereof), this leads linux vendors
to ship either snapshots of the 2.95 branch or heavily patches versions
of the official releases.  As long as gcc 3.0 still isn't considered
production ready by many companies it major work is done on this base,
e.g. by the Caldera OpenServer/OpenUnix groups (I'm not part of them).

Having a 2.95.4 or 2.95.3.1 or whatever release would really help to
get this work integrated in a nice way.  And I expect the same to happen
for either 3.1 or 3.2 again, when Linux and other vendors switch to
a stabilized 3.x gcc they want to stay with.

If a real gcc release (i.e. full regression testing) is considered to
heavy-weight for such an ancient branch, patchlevel releases or whatever
would be more than enough for most users.

	Christoph


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]