Dennis Schridde schreef:
> Am Samstag, 8. März 2008 14:43:46 schrieb Giel van Schijndel:
>>> [EMAIL PROTECTED]:~/devel/debian/games/warzone2100$ cat
>>> debian/patches/fix_quesoglc_0.7.patch diff -Naur
>>> ./lib/ivis_opengl/textdraw.c ./lib/ivis_opengl/textdraw.c ---
>>> ./lib/ivis_opengl/textdraw.c        2008-02-10 15:25:20.000000000 +0000 +++
>>> ./lib/ivis_opengl/textdraw.c        2008-03-07 12:58:03.000000000 +0000 @@
>>> -594,11 +594,7 @@
>>>  {
>>>     GLint matrix_mode = 0;
>>>
>>> -#ifdef WZ_OS_MAC
>>>     pie_SetTexturePage(0);
>>> -#else
>>> -   pie_SetTexturePage(-1);
>>> -#endif
>>>
>>>     glGetIntegerv(GL_MATRIX_MODE, &matrix_mode);
>>>     glMatrixMode(GL_TEXTURE);
>> Actually that patch will result in selecting texture page 0 instead of
>> unbinding whatever textures is currently bound (negative numbers will
>> cause the current texture page to be unbound).
>>
>> So while this may work on _some_ OpenGL implementations, it is not
>> guaranteed to work on all OpenGL implementations.
> Who says it is not guaranteed to work?
> You can always rebind the texture without causing any damage. (You dont have 
> to unbind first.)

Hmm, this actually didn't have a lot to do with OpenGL and a whole lot
more with the inner workings of pie_SetTexturePage(). Anyway, thanks to
Bertrand Coconnier this has been cleared up already.

As I understood it QuesoGLC will rebind the previously bound texture
after it bound its own texture, so invoking pie_SetTexturePage(0) at all
times _should_ be safe.

-- 
Giel

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Warzone-dev mailing list
[email protected]
https://mail.gna.org/listinfo/warzone-dev

Reply via email to