[MPlayer-dev-eng] Re: Problem compiling mplayer for PowerPC + AltiVec

[Alexander Strasser]

Hi,

Marcin 'Morgoth' Kurek wrote:
> On 2005-07-09, you wrote:
> 
> > Could you try to pinpoint the date when it broke?  Just try some CVS
> > snapshots and find out on which date it stopped working.
> 
> First sorry because the delay, I was away this weekend. Anyway I was able to
> do a small investigation and I think locate the source of my problem.
> 
> First I visit http://www1.mplayerhq.hu/MPlayer/cvs/ (I can't use cvs because
> dead slow inet connection) and locate the place when it starts to fail.
> 
> The first broken version is dated to 20050531, I can normaly compile 20050530
> version using gcc 3.4.4.
> 
> It seems the problem is not related to any change in the source code but to a
> compiler flag. The problematic flag is -std=gnu99 added somewhere in 20050531
> snapshot. When I switched to default gnu89 I was able to compile cvs version of mplayer without any problem using both gcc 3.4.4 and 4.0.1.
> 
> I wonder is there a good idea to use this flag ? Acording to the GCC manuals
> this standard is not fully implemented and the default gnu89 still allows to use many 
> of the 99 features.

  I might be totally wrong, but maybe it was introduced because of
some lrintf issues. If it was, then it might be very important.
  Not important other side effect of this change: it's causing some
warnings in conjunction with g++ invocation, but the compiler ignores
it afaict so i don't care.

  Alex (beastd)