gcc-2.96 (was Re: [GnomeMeeting-list] RPMS again)



Chris Tooley wrote:
> > > > gcc 2.95.4 (Debian) works fine, or replacing -mcpu=i686 by -mcpu=i386
> > > > with gcc 2.96 works fine.
> > >
> > > You have officially been bitten by the RedHat gcc bug that everyone was
> > > bitching about. :)
> >
> > Which one? Bugzilla number?
> 
> Correction, Not a "bug" just a "pre-release" version of the software
> that has been heavily modified. The fact that gcc-2.95 works and 2.96
> doesn't (at least in this instance) shows that 2.96 isn't providing
> expected behaviour.

It is. This problem (and the vast majority of reported problems against
gcc-2.96) have turned out to be problems with the code, not the compiler
(for a good collection of invalid code reported as "bugs" against
gcc-2.96, see http://www.bero.org/gcc296.html). The problem is that most
people who like to criticize gcc-2.96 for being "buggy" often
conveniently "forget" to mention that, which usually just makes such
discussions a giant trollfest and unappropriate for anything but
Slashdot.


> It started with RedHat 7.0 and has gotten significantly better, but is
> still an "unsupported" version by the gcc team.

The main issue from the gcc developers is that it uses the "gcc" name
and a version number that could be mistaken as a release from the gcc
team, which probably was a big mistake. That doesn't stop gcc-2.96 from
being one of the best free compilers today, though.


> An open letter on the issue by Bob Young explains their standing on the issue
> as of October 12, 2000.  But it's still not a release, no matter how many
> patches you apply, that the gcc team would like people using, 2.95 is.
> http://slashdot.org/article.pl?sid=00/10/12/163218&mode=thread

Yeah. But it's still free software and a damn good compiler, maybe even
the best there is right now, which may be why also Mandrake uses it.
Anyway, this is offtopic for gnomemeeting-list.


Christian



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]