This is the mail archive of the
glibc-linux@ricardo.ecn.wfu.edu
mailing list for the glibc project.
Re: gcc 2.95 and glibc 2.1.1
>>>>> Allin Cottrell writes:
Allin> I recently posted a suggestion for "fixing" the configure script
Allin> for glibc-2.1.1 so that it will compile using gcc-2.95. For the
Allin> record, I should point out that Andreas Jaeger says this is
Allin> a bad idea. In c.o.l.d.s on August 22 he wrote:
Allin> "I do advise against this. gcc 2.95 comes with new alias
Allin> analysis rules. If you use glibc 2.1.1 together with gcc 2.95,
Allin> you should use -fno-strict-aliasing to compile glibc and all
Allin> programs that use glibc since some files violate the strict
Allin> aliasing rules:-(.
Allin> This will be fixed in glibc 2.1.2."
Allin> Andreas, does this mean one should use -fno-strict-aliasing for
Allin> routine C compilation on a system with gcc 2.95 and glibc 2.1.1,
Allin> even if glibc-2.1.1 itself was compiled using the contemporary
Allin> egcs?
Yes.
The string inline functions (<bits/string*>) in glibc 2.1.x (x < 2) on
ix86 don't follow the strict aliasing rules. Either compile with
-fno-strict-aliasing or disable the use of the string inline
functions. This was discussed on the gcc list some weeks ago. glibc
2.1.2 shouldn't have these problems.
Andreas
--
Andreas Jaeger aj@arthur.rhein-neckar.de jaeger@informatik.uni-kl.de
for pgp-key finger ajaeger@aixd1.rhrk.uni-kl.de