cyrus-imapd 2.1.13 built with gcc 3.3 on linux (redhat)?

Will Partain will.partain at verilab.com
Mon Jun 2 12:39:56 EDT 2003


Henrique de Moraes Holschuh <hmh at debian.org> writes:

> On Thu, 29 May 2003, Will Partain wrote:
> > Folks, has anyone built cyrus-imapd 2.1.13 with gcc 3.3 on
> > linux (redhat)?  I am getting odd compile errors, using a
> 
> Debian, gcc 3.3.  Works well for Cyrus.  Won't define STDC with your little
> hello.c proggie, though.

Thanks for that.  I've dug around further, and am not much
the wiser.  The specific symptom is:

   % cd <build-dir>
   % cd et
   % make com_err.o
   /our/bin/gcc -c -I/our/.-ark-deploy/cyrus-sasl--2.1.13/include -I/our/.-ark-deploy/openssl--0.9.7b/include -I/our/.-ark-deploy/cyrus-sasl--2.1.13/include -I/our/.-ark-deploy/openssl--0.9.7b/include -I/usr/local/.-ark-deploy/tcp-wrappers--7.6-ipv6.1/include -I/our/.-ark-deploy/cyrus-imapd--2.1.13/include -I/our/.-ark-deploy/openssl--0.9.7b/include -I/usr/local/.-ark-deploy/tcp-wrappers--7.6-ipv6.1/include -DHAVE_CONFIG_H -I. -I./../lib -I. -I.. -Wall -O2 -march=i686 -I/our/.-ark-deploy/cyrus-sasl--2.1.13/include -I/our/.-ark-deploy/openssl--0.9.7b/include -I/our/.-ark-deploy/cyrus-sasl--2.1.13/include -I/our/.-ark-deploy/openssl--0.9.7b/include -I/usr/local/.-ark-deploy/tcp-wrappers--7.6-ipv6.1/include -I/our/.-ark-deploy/cyrus-imapd--2.1.13/include com_err.c
   In file included from com_err.c:54:
   /our/.-ark-deploy/gcc--3.3/lib/gcc-lib/i686-pc-linux-gnu/3.3/include/varargs.h:4:2: #error "GCC no longer implements <varargs.h>."
   /our/.-ark-deploy/gcc--3.3/lib/gcc-lib/i686-pc-linux-gnu/3.3/include/varargs.h:5:2: #error "Revise your code to use <stdarg.h>."
   [snip -- everything goes to pieces after that]

Dropping in a gcc-3.2.2 for /our/bin/gcc (which is 3.3)
makes it go through fine.

Experiments (replacing '-c' with '-E -dD') suggest that the
gcc-oid varargs.h [e.g. the one cited above] really did
change from a did-something version in 3.2.2 to a
don't-use-me version in 3.3.  With both versions, the
varargs.h is getting sucked on (__STDC__ notwithstanding).

So, the question is: is your varargs.h different (e.g. still
a do-something version), or are you avoiding the new one
(e.g. by the HAVE_STDARG_H thing really work)?

I will gratefully receive (private email?) someone's
equivalent 'gcc -E -dD ... com_err.c' from a working gcc-3.3
build, for further rumination.  Thanks for any help,

Will





More information about the Info-cyrus mailing list