This is the mail archive of the
libffi-discuss@sourceware.org
mailing list for the libffi project.
Re: libffi on OpenBSD 4.3
- From: "Sebastian Reitenbach" <sebastia at l00-bugdead-prods dot de>
- To: aph at redhat dot com
- Cc: libffi-discuss at sourceware dot org
- Date: Tue, 27 May 2008 15:44:47 +0200
- Subject: Re: libffi on OpenBSD 4.3
- Reply-to: "Sebastian Reitenbach" <sebastia at l00-bugdead-prods dot de>
Andrew Haley <aph@redhat.com> wrote:
> Sebastian Reitenbach wrote:
> > Andrew Haley <aph@redhat.com> wrote:
> >> Sebastian Reitenbach wrote:
> > OK, found the logfile, here is an excerpt of the errors:
> >
>
> >
> >
Running /usr/local/src/libffi-3.0.5/testsuite/libffi.special/special.exp ...
> > set_ld_library_path_env_vars:
> > ld_library_path=.::/usr/local/src/libffi-3.0.5/testsuite/../.libs
> > Executing on host:
> > gcc /usr/local/src/libffi-3.0.5/testsuite/libffi.special/unwindtest.cc
> > -shared-libgcc -lstdc++ -O0 -W -Wall
> > -I/usr/local/src/libffi-3.0.5/testsuite/../include
> > -I/usr/local/src/libffi-3.0.5/testsuite/../include
> > -I/usr/local/src/libffi-3.0.5/testsuite/../include/..
> > -L/usr/local/src/libffi-3.0.5/testsuite/../.libs -lffi -lm -o
> > ./unwindtest.exe
> > (timeout = 300)
> > /usr/local/src/libffi-3.0.5/testsuite/libffi.special/unwindtest.cc:11:
> > error: syntax^M
> > error before `__attribute__'^M
> > compiler exited with status 1
> > output is:
> > /usr/local/src/libffi-3.0.5/testsuite/libffi.special/unwindtest.cc:11:
> > error: syntax^M
> > error before `__attribute__'^M
> >
> > /usr/local/src/libffi-3.0.5/testsuite/libffi.special/unwindtest.cc:11:
> > error: syntax
> > error before `__attribute__'
>
> I don't understand this. It's supposed to be gcc 4.2 but it looks like
> it's not correctly parsing the __attribute__ syntax. This should have
> been fixed with gcc 3.4.0.
The system compiler is gcc 3.3.5.
When running "make check", then this compiler is always used, regardless of
CC CXX or CPP environment variables.
I had to move the system gcc, g++, and cpp out of the way, and then create
symbolic links to egcc and so on. To be able to use the gcc-4.2 in the
tests. maybe I forgot sth. from the 3.3. to move out of the way.
Sebastian