[MPlayer-dev-eng] icc benchmark

[Uoti Urpala]

On Tue, 2007-06-12 at 12:57 +0200, Michael Niedermayer wrote:
> the problem here is that the odds are 1:100 or so that uoti will post a
> patch which breaks gcc 2.95
> so maybe you rather should bet that he posts one which doesnt and hope
> in human error ...

I consider gcc 2.95 support to be useless, and the attempts to maintain
support to be actively harmful when they interfere with useful
development. Because of that I make no attempt to avoid breaking gcc
2.95. But I do not intentionally write code to break it. If gcc 2.95
breaks, it's because it fails to compile what I consider to be the
overall best way to write the code.

Unless compiler issues with more relevant compilers turn up I consider
the way I gave to be the best way to write the asm. I explained the
reasons in the earlier mail. Do you have a reason why another way would
be better, or are you trying to hinder MPlayer development only because
an obsolete compiler cannot compile the best way to write the code?