[FFmpeg-devel] [PATCH] Move libm replacements to new header libm.h

[Måns Rullgård]

Aurelien Jacobs <aurel at gnuage.org> writes:

> On Fri, Mar 12, 2010 at 02:18:58PM +0000, M?ns Rullg?rd wrote:
>> Ramiro Polla <ramiro.polla at gmail.com> writes:
>> 
>> > 2010/3/9 M?ns Rullg?rd <mans at mansr.com>:
>> >> Michael Kostylev <michael.kostylev at gmail.com> writes:
>> >>
>> >>> On Tue Mar ?9 13:12:41 2010
>> >>> M?ns Rullg?rd wrote:
>> >>>
>> >>>>>> ffmpeg.c uses lrintf(), which is missing on some systems.
>> >>>>>
>> >>>>> On which ones? Djgpp is almost dead, so I'd prefer to fix it locally.
>> >>>>> Dietlibc? It is much more incomplete and no one cares.
>> >>>>
>> >>>> The DOS builds are failing because of this. We added these hacks to
>> >>>> work around deficiencies like this.
>> >>>
>> >>> I remember.
>> >>>
>> >>>> ?Are you suggesting we drop them?
>> >>>
>> >>> More precisely, *lrint*(), round*() and truncf().
>> >>
>> >> Do we want to drop support for these systems? ?Will they be fixed
>> >> upstream? ?If the answers to both of those questions are "no", we need
>> >> to keep the workarounds.
>> >
>> > On a cygwin 1.7 config.h I get:
>> > #define HAVE_LLRINT 1
>> > #define HAVE_LRINT 1
>> > #define HAVE_LRINTF 1
>> > #define HAVE_ROUND 1
>> > #define HAVE_ROUNDF 1
>> > #define HAVE_TRUNCF 1
>> >
>> > On a cygwin 1.5 config.h I get:
>> > #define HAVE_LLRINT 0
>> > #define HAVE_LRINT 1
>> > #define HAVE_LRINTF 1
>> > #define HAVE_ROUND 1
>> > #define HAVE_ROUNDF 1
>> > #define HAVE_TRUNCF 1
>> 
>> Thanks for that.  Nice to see some progress on the cygwin side.
>> 
>> Now that all FATE configs are working again (ia64 has other issues), I
>> prefer to leave these workarounds as long as they don't cause any
>> trouble.  Some of the BSDs and uclibc still need a few of them too.
>
> FYI, latest uClibc release now includes log2f() and exp2f() which were
> previously missing.

That doesn't imply all the silly chip vendors have updated their patch
sets.

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