This is the mail archive of the
libc-alpha@cygnus.com
mailing list for the glibc project.
atexit problems with egcs...
- To: libc-alpha@cygnus.com
- Subject: atexit problems with egcs...
- From: "Dan A. Dickey" <ddickey@wamnet.com>
- Date: Wed, 15 Apr 1998 11:59:50 -0500
- Organization: Wam!Net, Inc.
- Reply-To: libc-alpha@cygnus.com
Hi,
I've run into a problem with the definition of atexit.
I recently compiled egcs-1.0.2 on a Redhat 5.0 system,
which had libc-5.4.44 and glibc-2.0.7-5 on it. No problems
(except that I didn't do 'make bootstrap'). I installed
egcs into /usr/local.
Then, I compiled and installed glibc-2.0.92 into /usr/local.
Now, I'm trying to rebuild egcs (doing a 'make bootstrap')
with itself (/usr/local/bin/gcc). HOWEVER, the definition
of atexit in /usr/local/include/stdlib.h (from glibc-2.0.92)
does not agree with that in the egcs distribution
(.../gcc/gbl-ctors.h). So, the make bootstrap fails building
libgcc2.a due to a conflicting type error.
Has anyone else run into this? What's the quick fix?
What's the correct fix?
It's times like this that I wish Makefiles were written
with a 'make uninstall' capability (so I could uninstall
glibc2 from /usr/local).
Any and all help appreciated! Thanks. -Dan
P.S. - I'm writing here first because I knew the libc-alpha@cygnus.com
address. And I am aware that the problem may be with egcs.
And I'm also aware that many people here are on both lists.
(So this should be a good first scream for help. :) )
----
Dan A. Dickey
ddickey@wamnet.com