Re: [AD] Docs and set_gfx_mode

[Chris <kcat@xxxxxxxxxx>]

Evert Glebbeek wrote:
> What we should have, and this is probably a good thing to have for the 4.2
> version, is a vram_single_surface function. This could either report the
> status on a driver (before a mode is set) or on the current screen/driver
> settings (after the mode is set). In the latter case, we could also add a
> flag to gfx_capabilities. Actually, we can probably do that anyway.
> The only real question is wether or not this is actually useful: you won't
> know if you need a virtual screen until after you've called set_gfx_mode().
> Sure, you can call it again, but that's not very nice IMO...

Right. You shouldn't need to set a mode twice just to get something you 
want. Plus, as I brought up before, you can't really use a function to 
check when dealing with GFX_AUTODETECT[_FULLSCREEN|_WINDOWED] since you 
can't know which driver you get until it's first been detected, and then 
inited. If either fail, it tries another with the same parameters (it 
does this for sound drivers it seems.. not sure about gfx).

This is a case point that Allegro should just allocate as much VRAM as 
possible, and just use the driver method for creating video bitmaps. Of 
course, there's two problems with this... DOS (maybe? not sure) and X. 
Doing that would mean you couldn't do any fake VRAM surfacing, like the 
X11 drivers do.. emulating more VRAM by offsetting drawing operations 
with an enlarged memory bitmap. Their might be some ways around it, but 
IMO it's not at all worth it and we're better off just waiting for the 
new gfx API (how is that going to handle virtual screens, btw?).

- Kitty Cat