Re: [AD] midi driver detection screws up digi driver detection

[George Foot <gfoot@xxxxxxxxxx>]

On Thu, May 30, 2002 at 04:57:26PM +0200, Martijn Versteegh wrote:
> If I have time I'll try to patch up install sound a bit. I propose:
> 1) install sound should fail if you explicitly specify a
> driver which isn't detected. Right now it'll do autodetection if it fails
> on the explicitly specified driver, giving you another driver than you 
> expect. Maybe it should switch to digi/midi_none and never fail?
> 2) if something goes wrong in installing one of digi_driver/midi_driver
> this should not screw up the installation of the other one. This means
> that at 1) it should probably switch to the *_none driver and never fail.

I agree with (1), if the user asks for a particular driver we
should use only that driver.  But for (2) I only agree in the
case that the failed driver was set to autodetect (or none).  If
the user asks for MIDI_ALSA and it fails, the function should
return failure, and if the function returns failure, the digi
driver should not be installed either.

It would be interesting to see in which cases the digi and midi
drivers *do* depend on one another.  But presumably this has
already beet sorted out for Allegro 5. :)

George