[FFmpeg-devel] [RFC] replace some static with asm_visibility or so

Uoti Urpala uoti.urpala
Sun Jan 27 23:26:06 CET 2008

On Sun, 2008-01-27 at 22:37 +0100, Michael Niedermayer wrote:
> On Sun, Jan 27, 2008 at 10:54:43PM +0200, Uoti Urpala wrote:
> > On Sun, 2008-01-27 at 21:22 +0100, Michael Niedermayer wrote:
> > > On Sun, Jan 27, 2008 at 09:18:57PM +0200, Uoti Urpala wrote:
> > > > I see no evidence for any of those claims ("months of work", "more
> > > > fragile" or "harder to maintain").
> > 
> > ... and you still gave no reason to believe that even one of your claims
> > would have been justified.
> to proof the time, it needs to implement it, one would have to implement it.
> As noone will implement your idea (not even you) one can only speculate
> how long it would have taken.

It doesn't take much speculation to see your claim was wrong.

> the harder to maintain and more fragile is obvious from the need of
> additional code which would not be needed otherwies (special cases for
> gcc 2.95/gcc 3

Disabling asm for some compiler versions shouldn't be hard to maintain.

> , vissibility attributes, different PIC/no PIC flags for 
> some files, ...)

If you compile files with asm without -fPIC then visibility attributes
do not matter for those files, so listing these both as additional
requirements makes no sense.

Using different flags only matters if you really want to generate x86
binaries where some parts are PIC and some are not. How many actually
benefit from that? Do the people who currently add -fPIC to flags on x86
understand that they're not really getting PIC binaries because of

> you have posted some patches which "work" for your definition of "work"
> but not for the definition used by anyone else

I think my definition of "work" is shared by at least as many others as

> other people like portable code which works with more than latest gcc

Yes many do, but most of them do not mean software from a decade ago
when they say "more than latest".

More information about the ffmpeg-devel mailing list