Errors on x86_64 (broken c++ headers?)

[Panu Matilainen]

On Thu, 12 May 2005, Quentin Spencer wrote:

> I resubmitted octave to be built yesterday, with more odd results. Everything 
> looks fine for PPC, but for i386, the build appears to have completed, but 
> there is still a failure.log file that says "Build process terminated by 
> timeout". For x86_64, the build failed. Looking at the log file, there are a 
> whole lot of errors that look something like this:
>
> g++ -c  -fPIC -I. -I../.. -I../../liboctave -I../../src -I../../libcruft/misc 
> -DHAVE_CONFIG_H  -Wall -W -Wshadow -Wcast-align -Wcast-qual -Wpointer-arith 
> -Wwrite-strings -Weffc++ -g -O2 quit.cc -o pic/quit.o
> In file included from quit.cc:30:
> /usr/lib/gcc/x86_64-redhat-linux/4.0.0/../../../../include/c++/4.0.0/iostream:43:28: 
> error: bits/c++config.h: No such file or directory
> In file included from 
> /usr/lib/gcc/x86_64-redhat-linux/4.0.0/../../../../include/c++/4.0.0/ios:43,
>                from 
> /usr/lib/gcc/x86_64-redhat-linux/4.0.0/../../../../include/c++/4.0.0/ostream:44,
>                from 
> /usr/lib/gcc/x86_64-redhat-linux/4.0.0/../../../../include/c++/4.0.0/iostream:44,
>                from quit.cc:30:
>
> So, the errors appear to be caused by the c++ header files. How is it 
> possible that the header files are broken on one arch and not another? Any 
> other insights into what might have caused this and how to fix it?

There's something weird about mach and C++, I've seen this and some other 
weirdness just recently myself. The above can be worked around by setting
a suitable CPPFLAGS=-I/usr/include/c++ ... before %configure.

Another oddity I found is that for Wine to build within mach I had to use
LDFLAGS="-lstdc++" %configure - which doesn't make any sense to me either.

 	- Panu -