This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Anyone got irix5? I have questions about your ctype implementation
- From: Alexandre Oliva <aoliva at redhat dot com>
- To: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>, mark at codesourcery dot com
- Cc: cgf at redhat dot com, gcc at gcc dot gnu dot org, ro at TechFak dot Uni-Bielefeld dot DE, gcc-patches at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org, java-patches at gcc dot gnu dot org
- Date: 07 May 2002 20:16:15 -0300
- Subject: Re: Anyone got irix5? I have questions about your ctype implementation
- Organization: GCC Team, Red Hat
- References: <200205020414.AAA15565@caip.rutgers.edu><20020503013353.GA2985@redhat.com><200205031423.KAA13608@caip.rutgers.edu><yddg019w2mf.fsf@xayide.TechFak.Uni-Bielefeld.DE><200205032010.QAA10493@caip.rutgers.edu><or8z6zrxh2.fsf@livre.redhat.lsd.ic.unicamp.br><200205060400.AAA21303@caip.rutgers.edu><ory9explnm.fsf@free.redhat.lsd.ic.unicamp.br><200205061211.IAA02448@caip.rutgers.edu><orbsbr50h0.fsf@free.redhat.lsd.ic.unicamp.br><or8z6v3ebk.fsf@free.redhat.lsd.ic.unicamp.br>
[apologies if you get this twice; for some reason the mailing list
processor rejected the message for gcc-patches and java-patches
because they were not in To: or Cc: (they were in the second
occurrence of Cc:)]
On May 7, 2002, Alexandre Oliva <aoliva@redhat.com> wrote:
> Never mind. I've just duplicated it with a full bootstrap. Something
> is polluting mabi=64/config.cache with an incorrect setting of LD.
> Either libf2c or libstdc++-v3. I'm looking into it.
Ok, any libtool library would do it, due to an ill effect of (some)
shells.
My understanding is that, if an exported variable is modified in a
shell script, the modified value is not supposed to be automatically
re-exported. It should be kept private to the script.
It appears that IRIX's /bin/sh doesn't agree with my understanding, so
this work around was necessary. Ideally, libtool shouldn't mess up
with LD, but I'd rather not touch this at this point.
This patch sucessfully bootstrapped the gcc 3.1 branch on
mips-sgi-irix6.5, and lt_cv_path_LD in both multilibs' config.cache
contained exactly the original value of LD, not the one modified by
libtool.m4.
Mark, ok for 3.1?
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist Professional serial bug killer