[MPlayer-dev-eng] [PATCH] clean up tests in configure

[Diego Biurrun]

On Tue, Jul 05, 2005 at 12:41:23PM -0400, Rich Felker wrote:
> On Tue, Jul 05, 2005 at 02:42:15AM +0200, Diego Biurrun wrote:
> > -a/-o vs &&/||:
> > ===============
> > 
> > You say that
> > 
> > 1) -a/-o is less readable (even "obfuscated"),
> > 2) the possibility for ambiguity with -a/-o is reason enough to reject
> >    the patch.
> > 
> > I disagree on both counts since
> > 
> > 1a) -a/-o is shorter (and IMO this aids readability on long lines),
> 
> then use [ instead of test.

That's orthogonal to using -a/-o (and yes, I'm all for [ instead of test).

> > 2) I don't think we run the risk of ever hitting that ambiguity and
> >    1a-c) might be a real gain in practise.
> 
> essentially zero gain. 90% or more of configure time is spent in gcc
> anyway, except on broken archs like windows where fork takes forever.

I'll try to come up with some numbers during the next days, but yes,
Cygwin would be a nice test case.

Diego