[MPlayer-cvslog] r26639 - trunk/configure

[Reimar Döffinger]

On Sat, May 03, 2008 at 05:21:09AM +0200, Diego Biurrun wrote:
> On Fri, May 02, 2008 at 03:48:11AM +0200, Michael Niedermayer wrote:
> > 
> > But somehow the "braindead" sounds like the
> > original reasons for this are not understood. And that sounds all to
> > familar, changes being done because the original reasons, original
> > designs or original code is plain not understood instead of anything
> > being wrong with it.
> 
> This assumes that (good) original reasons and designs exist and are
> worth preserving.  For large parts of MPlayer this assumption is not a
> sound one.

Even the worst code quite often had good reasons originally (if we
ignore the cases of bad coders with too little review), but due to icc
support and generally improved compiler quality (even if not as improved
as we would like) the previous code indeed was overkill IMO, but as it
is now on the other hand is really taking it a bit too far, especially
since contrary to what I thought making the Sun C compiler work is
really non-trivial now and needs a bit annoying changes (e.g. commenting
out the "die" is not enough since it still tries overriding with gcc,
even when there is an explicit --cc= option, and even changing that
might still leave cc_vendor set to gnu...

Greetings,
Reimar Döffinger