[MPlayer-cvslog] r31100 - trunk/libvo/x11_common.c

[Diego Biurrun]

On Wed, Apr 28, 2010 at 11:55:13AM +0000, Carl Eugen Hoyos wrote:
> Diego Biurrun <diego <at> biurrun.de> writes:
> 
> > > win32dll was broken six weeks ago, I noticed immediately, but did not succeed
> > > in fixing the problem, yes.
> > 
> > libass broke many months ago.
> 
> Probably, but I only need the win32dll function for regression tests, I am not
> very interested in subtitles.
> 
> > Now please answer my question from the other mail.
> 
> Which one?

This one:

  > If you don't change configure to fail for 2.95.3, I will add autodetection for
  > the necessary disable's.

  Fail as in "reject 2.95 completely"?

> Or wasn't I clear when I wrote that I will fix compilation if you don't disable
> 2.95.3 support completely (which only makes sense if you do the same thing in
> FFmpeg, imo).

Not quite.  I don't see a relationship with FFmpeg.

Let me rephrase: Obviously gcc 2.95.x is a vintage compiler that is
hardly ever used by anybody these days.  Witness how long it took to
notice and fix breakage.

gcc 2.95.x has many issues, SSE and declarations vs. statements come to
mind immediately, there probably are others.

I used to care about 2.95.x and used it for years longer than most
people, but I stopped using it long ago and no longer care.  Call me
cold and bitter, but this love affair is through, the old feelings
are gone for good.  I got used to the world of slow compilation newer
versions give me and I'm jaded enough not to complain anymore.

AFAICT you are the only person that still cares about 2.95.x.  I wonder
why, do you really still use it?  Do you really think it's worth the
trouble to keep supporting it?

So yes, I think the time has come to drop support for gcc 2.95.x.

Diego