Re: [freesci-develop] Releases 0.3.4, 0.3.4a, 0.3.4b, 0.6.1 removed due to licencing issues

[Christoph Reichenbach]

Matt,

> I saw the roll back submission today. Will there be any further changes 
> to the fmopl code in the near future or will it stay as currently
> is? Since ScummVM uses the same version as we have right now I would not 
> expect much quality difference, or were there FreeSCI specific changes 
> included in the fmopl code? 

I'm not entirely sure, as the patch was Walter's work, but I believe
that it's essentially the same (except for some hacks to increase the
number of voices available to twelve). I can't think of anything we'd
want to fix right away; legality should have been restored at least
for the stable branch.

> I am finishing up the Mac wrapper and was able to create FreeSCI with 
> SDL statically linked. The only way I was able to do this was by 
> changing the configure file myself and replace '--libs' by 
> '--static-libs' (around line 7636 'SDL_LIBS=`$SDL_CONFIG $sdlconf_args 
> --static-libs`'). 

The 'configure' script is auto-generated. I would recommend trying to
find ways to modify 'configure.in' or 'aclocal.in' instead to achieve
the same result.

> Am I missing some kind of option to set this in another way or has SDL 
> always be linked dynamically with FreeSCI?

On the UNIX side, SDL has probably always been linked dynamically
(from what I've seen).

> The pro for linking SDL statically is that it reduced the number of 
> dependencies to just one: X11. And X11 is standard shipped with Panther 
> and freely downloadable from the Apple website. 
> It increased the size somewhat: 3,5 MB instead of 2,5 MB. 

Personally, I'm fine with that.

> First reason to keep SDL inside FreeSCI for Mac is because the unix 
> sound server does not work currently (so no sound if SDL is not 
> included). 

Yes, this has been pointed out to me. The issue seems to be that
fork()ed processes do not inherit certain properties that allow them
to do sound output. Whether this will affect glutton depends on the
timing mechanism used there (right now, there is none that would
exhibit the same problem, though).

> Second reason is because (believe it or not) X11&FreeSCI had problems 
> starting on at least one PowerMac (of a friend). 
> It looked like X11 was trying to get some color support and then just 
> gave up (it hung and after a few moments FreeSCI crashed). 

Interesting. If you have any further information that would indicate a
failure in FreeSCI, could you submit a bug report
(http://www-plan.cs.colorado.edu/reichenb/bugzilla), please?

> SDL for video worked fine on that machine. 

The SDL driver has performance issues, though. But you've probably
noticed that...

> I think have found a Panther specific linking issue with FreeSCI but I 
> want to check something before I send the summary on it.

OK. Good luck, and thanks for checking!


-- Christoph