[Freeciv-Dev] Re: request for --disable-sound

[Davide Pagnin <nightmare@xxxxxxxxxx>]

On Wed, 2002-05-29 at 11:13, Raimar Falke wrote:
> On Wed, May 29, 2002 at 12:38:19AM +0200, Davide Pagnin wrote:
> >     Hi all!
> > 
> > I've read of some people having trouble with sound and 
> > I hope that there is general consensus that a --disable-sound
> > option has to be added to the configuration script.
> > 
> > I've managed to write a macro and made the few changes needed
> > to have the patch work with configure.in (hope that the 
> > changes to configure.ac works in the way I hope!).
> > 
> > The result is quite good (for my point of view!).
> > 
> > This was mine environment :
> > 
> > A NFS server without sound libraries installed
> > An exported /usr/local directory to 5 client machine
> > 1 client machine with sound support 
> > 4 client machine without sound support
> > the esound related libraries installed in the
> > share /usr/local/ directory via a rpm package
> > 
> > This give me some headache, because configure
> > finds esound stuff and want to use that, afterwords
> > it finds out that the library related libmme.so
> > isn't present in the client system (unless the
> > machine which has the sound support)
> > 
> > With may patch, the client works even for the clients
> > without sound support.
> > 
> > The patch is rather rough, and obviously need improvement,
> > the main in related to clean all the messages that the
> > client spit out saying the lack of stdsound.spec, even
> > in the case of sound disabled via the --disable-sound
> > option.
> > 
> > Please, let me know what you think.
> 
> I would prefer a solution where the checks for ESD and SDL are made
> more strict and so these options are removed on your systems.

A run-time test of both the libraries, could do the trick, for my
systems. Well, mostly, because in the actual setup, if I compile on the
machine with sound support and install as usual on /usr/local, there is
nothing wrong to have sound on, but if I try to run on the other
no-sound machine, the execution halt with an error message like
"mapping of libmee.so impossible". Obviously if I compile on a no-sound
machine, and the run-time checks works well, the sound will be disabled.
But, why not having the opportunity to disable sound? If someone for
whatever reason don't want sound enabled, couldn't be a desirable  
option?

At the moment I've added a little patch to the one already submitted,
that get rid of the annoying error message that the audio.c file
spit in the case of no sound plugin loaded (this is obviously the case
with disable sound!)

Per, Have you tried the disable sound patch? What do you think of the
m4 macro wording? 

> 
> I don't know the exact details but it should be possible to do a test
> compile in the esd.m4 and sdl.m4 and link with the appropriate
> libraries to check for their existence.
> 
>       Raimar
> 
> -- 
>  email: rf13@xxxxxxxxxxxxxxxxx
>  "We just typed make..."
>     -- Stephen Lambrigh, Director of Server Product Marketing at Informix,
>                          about porting their Database to Linux
>

audio.c.patch
Description: Text document