[FFmpeg-devel] Darwin gnu99

[Måns Rullgård]

Michael Witten <mfwitten at MIT.EDU> writes:

> On 19 Oct 2008, at 5:27 AM, M?ns Rullg?rd wrote:
>
>> Michael Witten <mfwitten at MIT.EDU> writes:
>>
>>> Hello,
>>>
>>> I had to make the following change to the configure script
>>> in order to build on Mac OS X 10.5:
>>>
>>> --- configure	(revision 15641)
>>> +++ configure	(working copy)
>>> @@ -1149,7 +1149,7 @@
>>> TMPS="${TMPDIR1}/ffmpeg-conf-${RANDOM}-$$-${RANDOM}.S"
>>> TMPSH="${TMPDIR1}/ffmpeg-conf-${RANDOM}-$$-${RANDOM}.sh"
>>>
>>> -check_cflags -std=c99
>>> +check_cflags -std=gnu99
>>
>> Rejected.  Please explain what breaks without any patches.
>
> That language is a bit overly strong.

It's pretty standard around here.

> It's quite clear from my message that I never intended that
> diff output to be considered a patch:
>
>> I had to make the following change to the configure script
>> in order to build on Mac OS X 10.5:

That could be interpreted as meaning that no other fix is possible.

> In any case, it would seem conrad was clever enough to pick
> up on the problem:
>
>> $ svn log libswscale/yuv2rgb_altivec.c
>> ------------------------------------------------------------------------
>> r27805 | conrad | 2008-10-19 13:56:35 -0500 (Sun, 19 Oct 2008) | 1
>> line
>>
>> Convert typeof keyword into __typeof__; this is a gcc extension and
>> the former is not accepted in C99 without GNU extensions enabled
>> (e.g. via -fasm).  This fixes compilation on PPC.
>
> However, I wonder if I influenced this change, though I was
> admittedly quiet on the subject; it would be nice to include
> references to people that may have helped find the problem,
> because it makes people want to find and report more problems.

Had you bothered to explain what went wrong without your change, maybe
someone would have fixed it sooner.  As for attributions, you get
credited if you send a patch that gets committed.

-- 
M?ns Rullg?rd
mans at mansr.com