Re: [AD] 4.2.0 Beta 3 woes

[Elias Pschernig <elias@xxxxxxxxxx>]

On Fri, 2005-06-03 at 00:45 +0200, Evert Glebbeek wrote:

> > The GFX_HAS_VGA is the intended fix though right now, right? So it
> > should be applied for test.c and that example..
> 
> Yes. In essence, it makes sure that the #ifdef MODEX construction you could
> use earlier can still be used, but with a different symbol name. Someone
> should check all the sources to make sure we have *really* found the last
> occurence of this problem now.
> By the way, the symbol modex_split_screen (or whatever) is now supposed to
> be always defined, but just do nothing on platforms where it makes no
> sense. If it isn't, then I think that's a bug.
> 

Apparently it isn't in Shawn's case.

src/misc/modex.c has this near the top:

#if (!defined ALLEGRO_LINUX) || ((defined ALLEGRO_LINUX_VGA) && ((!defined ALLEGRO_WITH_MODULES) || (defined ALLEGRO_MODULE)))

So, when ALLEGRO_LINUX is defined, but ALLEGRO_LINUX_VGA is not (heh, so
my advice to disable VGA was quite useless indeed :P), then it will not
be defined. Just removing the ALLEGRO_LINUX_VGA check will always define
it.

So, should we do that instead? Anyone knows what was the reason to not
compile modex.c when ALLEGRO_LINUX_VGA is not defined?

-- 
Elias Pschernig