[MPlayer-cvslog] CVS: main configure,1.1197,1.1198

[Ivan Kalvachev]

2006/5/13, Diego Biurrun <diego at biurrun.de>:
> On Sat, May 13, 2006 at 07:39:48PM +0300, Ivan Kalvachev wrote:
> > 2006/5/13, Diego Biurrun <diego at biurrun.de>:
> > >
> > >This discussion ends here for me.  We have already wasted far too much
> > >valuable time over nothing.  My decision is final, I will not accept one
> > >parameter behaving different from the others.
> >
> > It does NOT.
> >
> > What I am trying to explain to you, is that the current semantics is:
> > If option is autodetected then --enable will force it.
> > If option is disabled then  --enable will autodetect it.
>
> This is not the current semantics.  The current semantics are that
> --enable unconditionally enables an option without checking.  This is
> where we differ from autoconf in not trying to second-guess the user
> and accept that compilation may fail in some cases.

Then how is user supposed to request auto detection if option is
disabled by default?

Even if we assume that what you state is current semantics,
then you must elaborate why it is better than the one I explained..

> > That's why all options that have check (they are few indeed) do
> > perform check on "yes" and there is no way to force them (e.g. gui).
>
> This is a false statement, checks are performed on "auto" only.

Can you prove that? gui,xvmc,joystic ?
The other disabled options simply doesn't provide auto-checks.

> I suggest you to reread configure.  You seem to be a little rusty on how
> it works nowadays.

No surprise, when the maintainer doesn't know how configure is
supposed to work ;)