This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: GCC 3.0 Release Criteria
- To: gcc at gcc dot gnu dot org
- Subject: Re: GCC 3.0 Release Criteria
- From: David Ronis <ronis at montroll dot chem dot mcgill dot ca>
- Date: Thu, 27 Apr 2000 22:56:06 -0400
- Reply-to: ronis at onsager dot chem dot mcgill dot ca
Perhaps time for a reality check?
More than 50% of the recent snapshots fail to bootstrap with
aggressive optimization turned on on ix86's. Worse, when they build,
the resulting stdc++ library is broken and leads to hundreds of C++
testsuite failures (BTW, this problem is already present in the 2.95.2
release version). Finally, there are the stack alignment problems
pointed out months ago by Steve Johnson which are performance killers
on i686's (these may have been fixed, but I've not tried his test program
in some time, given the other 2 problems I just mentioned).
If full optimization cannot be used in building the compiler and its
included libraries, then it's simply not ready for release. [IMHO, this
should be added to the release criteria document].
David Ronis