[MPlayer-dev-eng] [PATCH] faad and gcc-3.2.3

[Magnus Damm]

Hi there,

> instead of messing with version numbers, s1 should separate some sample
> source which reproduce the bug, and try to compile that sample in
> ./configure, with the selected compiler optim flags.

Sure, but that someone is not me, because I'm able to compile and use faad,
until someone disabled it for me without any option to turn it on...

And that leads us to how --enable-*/--disable-* options are handled in
configure. How are they supposed to be work?

This is what I believe is sane and "standard":

1. First do some kind of autodetection. 
2. Let --enable-*/--disable-* override the autodetected value.
3. Check if it's possible to use the feature.
4. If --enable-* is used and the feature can't be used, break and
   show the error, otherwise disable the feature and continue.

Or is some other way preferred?

The question is, should faad be auto-detected as off if gcc-3.2.x
is used but --enable-faad should override and let the user try.
Or someone could implement a test, that would be the best.

/ magnus