On Tue, 20 Mar 2012, Gil Pedersen wrote:
On 20/03/2012, at 01.37, Luca Barbato wrote:
On 19/03/12 16:43, Måns Rullgård wrote:
From my point of view, OSX is fringe. Besides, we have plain C code too.
The speed difference is not staggering.
If it isn't we might try to simplify it by checking that construct is
supported and leave unsupported compilers out.
I guess we could make 100% sure brew, gentoo-prefix and every other
offset-distro provide a compiler working for the target and just add a
big warning and point on the safe compiler they could use if they need
this additional bit of speed.
From the discussion this sounds like a llvm specific bug. On OS X the default
gcc compiler uses a llvm backend for the compilation, as can be seen from
--version
$ gcc --version
i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1 (Based on Apple Inc. build 5658)
(LLVM build 2336.9.00)
However we still have a pure gcc compiler in gcc-4.2:
$ gcc-4.2 --version
i686-apple-darwin11-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3)
I expect compiles will still work using the latter.
On OS X 10.6, the latter is still the default:
$ gcc --version
i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5666) (dot 3)
With this compiler and the patch in
http://lists.libav.org/pipermail/libav-devel/2012-February/023084.html, I
get the same compile errors as Ronald showed in that mail.
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
