Re: New redhat-rpm-config for ac3

[Mike Barnes <mike@xxxxxxxxxxxxx>]

On 14/03/2007, at 8:13 PM, Oliver Falk wrote:
> Maybe you can add some statement on how to fix it in core-cvs  
> without conflicting with the primary archs; I saw you just removed  
> it from macros...

Yeah, I just patched it out for the moment, pending a better solution.

I'm not entirely sure that dropping it from macros is actually the  
right approach, since in theory it shouldn't make any difference if  
it's there or not. The actual problem I was seeing was that in some  
autoconf scripts, the test for a working -fPIC was failing - this was  
what was causing build failures in a lot of packages.

I'll try to follow through the configure tests and narrow it down a  
little further. If it's the case that the -fstack-protector flag  
should do _nothing_ on an unsupported architecture, and it's doing  
_something_, then it really strikes me as a GCC bug, not something to  
fix in the RPM macros, which ideally should be consistent across  
architectures.