This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc(0614 cvs) compile error on solaris 8
- To: gcc at gcc dot gnu dot org
- Subject: Re: gcc(0614 cvs) compile error on solaris 8
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- Date: Thu, 14 Jun 2001 21:03:20 -0500 (CDT)
- Cc: imyaman at account dot kouf dot org
- Organization: Networks and Infrastructure Lab (IL02/2240), Motorola Labs
In article <Pine.GSO.4.05.10106142316290.8553-100000@account.kouf.org>
you write:
> i configured with following options...
> ../configure --prefix=/export/home/imyaman/local --enable-shared
> --enable-threads=pthread --enable-languages=java,objc --enable-libgcj
> --enable-nls
Luckily, your problem is not a result of recent patches submitted to
improve threading in libstdc++-v3, but it does mean that I almost
fully understand all configuration related to threading for many
platforms. ;-)
Please use:
--enable-threads=posix
instead of:
--enable-threads=pthread
In light of your report, I am currently attempting to figure out how
to extend the documentation to best help people in this area for the
pending gcc 3.0 release. I do not yet know why gcc/config.gcc in
conjunction with gcc/configure isn't catching an unknown thread
configuration...
In gcc/config.gcc, *many* platforms explicitly map
--enable-threads=pthread to posix but some do not. Solaris is one
such port that does not do such mapping...
In gcc/configure, if config.gcc didn't set thread_file to anything
then by default thread_file is set to whatever you listed after
`--enable-threads='. Since the solaris configuration doesn't map
pthread to posix, you get a reference to gthr-pthread.h which doesn't
exist in the $srcdir/gcc.
Regards,
Loren
--
Loren J. Rittle
Senior Staff Software Engineer, Distributed Object Technology Lab
Networks and Infrastructure Research Lab (IL02/2240), Motorola Labs
rittle@rsch.comm.mot.com, KeyID: 2048/ADCE34A5, FDC0292446937F2A240BC07D42763672