Re: Multi threaded server : Segementation fault on ia64 Linux

[James E Wilson <wilson at specifixinc dot com>]

Gerald Pfeifer wrote:
> None of these shipped with GCC 2.96, which was not even an official GCC 
> release, but a Red Hat-specific version.  (For a statement by the GCC 
> developers, see <http://gcc.gnu.org/gcc-2.96.html>.)

This is a misquoted reference.  It does not apply to the compiler in 
question.

The GCC SC did object to Red Hat shipping an x86 OS based on a gcc 
snapshot that was called gcc-2.96.

However, Red Hat also shipped an Itanium compiler based on a gcc 
snapshot that was called gcc-2.96, and this compiler was disclosed to 
the GCC SC, and was blessed by the GCC SC.  Note that in this case there 
was really no other choice, as the Itanium work came too late to be 
included in gcc-2.95.  If we didn't have permission to use a snapshot, 
it would have been impossible to make any Itanium Linux releases until 
sometime after the gcc-3.0 release.  That could have set back Itanium 
Linux work by several years, so clearly we had to use a snapshot.

SuSE must have also made an Itanium Linux release using a gcc-2.96 
compiler, otherwise they would not have been able to have one of the 
earliest itanium linux releases, which came out before the gcc-3.0 
release.  However, I don't know if they ever used the "2.96 20000731 
(Red Hat Linux 7.1 2.96-101)" release.  It is more likely that they used 
one of the "Cygnus" gcc-2.96 releases that I made.  But it is possible 
that they used the later "Red Hat" gcc-2.96 release.
--
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com