Re: [Adonthell-devel] Windows Fullscreen mode again

[Alexandre Courbot]

> I've been browsing the web and our code in the search for a solution
> to the windows fullscreen issue. I stumbled across a few things that
> might be worth trying out.
> 
> First, there's a comment on the SDL ML
> (hhttp://www.libsdl.org/pipermail/sdl/2002-June/046105.html) where the
> guy says "that using SDL's rendering with only the SDL_SWSURFACE and
> SDL_FULLSCREEN flags works correctly, however adding SDL_DOUBLEBUF
> will break it." But we set that flag.
> 
> In the SDL docs (http://sdldoc.csn.ul.ie/sdlsetvideomode.php) it says 
> "SDL_DOUBLEBUF: Enable hardware double buffering; only valid with
> SDL_HWSURFACE"
> 
> But we're not using a HWSURFACE, so it's useless for us anyways (and
> might indeed cause problems under Windows).
> 
> 
> Finally, in our own code, the surface::double_size and
> surface::half_size methods do pixel manipulations of the image without
> locking the surface first. Is that correct? If it is, why? 

Hm, all what you said is perfectly pertinent, and I'd bet it solves the
problem. I've rebuilt the windows binary - on my machine (WineX) it
won't even start in DX mode, but so would the SDL demos, so I guess it's
ok. I've put the binary for you to test:

http://www.lifl.fr/~courbot/win-patch.zip

I'm quite confident on this one. Especially with the lock stuff, if you
remember at Ben's it didn't start at all in 640, while in 320 it started
(but was awfull). I guess this is the reason. Anyway - I'm looking
forward to know if you're successful! :)

Btw: I've also tried to build the latest SDL CVS for Windows, but
unfortunately the compilation seems to be broken. I'm sure it's not
necessary to make Adonthell work in fullscreen anyway!

Good luck! ;)
Alex.
-- 
http://www.gnurou.org