We currently convert all incoming textures to 32bpp RGBA for
passing to OpenGL - most OpenGL implementations will now
support the packed pixel extension, so we should be able to pass
on 16 or 24 bit pixel data directly rather than copying it.
This could also simplify the texture swapping code in IRUpdate.c,
which is quite complex and a bit misleading (e.g., 16bpp texture
data is not really swapped, it's permuted, and yet the variable is
called 'swap').
Logged In: YES
user_id=433183
Originator: NO
Texture loading now uses 24-bit RGB when there is no alpha channel. Since Quesa pixmaps are top to bottom and OpenGL textures are bottom to top, we need to row-order conversion on the texture data anyway. So I'm not sure if there's really much point to this.
Logged In: YES
user_id=433183
Originator: NO
Texture loading now uses 24-bit RGB when there is no alpha channel. Since Quesa pixmaps are top to bottom and OpenGL textures are bottom to top, we need to row-order conversion on the texture data anyway. So I'm not sure if there's really much point to this.
Logged In: YES
user_id=433183
Originator: NO
Texture loading now uses 24-bit RGB when there is no alpha channel. Since Quesa pixmaps are top to bottom and OpenGL textures are bottom to top, we need to row-order conversion on the texture data anyway. So I'm not sure if there's really much point to this.
Logged In: YES
user_id=433183
Originator: NO
Texture loading now uses 24-bit RGB when there is no alpha channel. Since Quesa pixmaps are top to bottom and OpenGL textures are bottom to top, we need to row-order conversion on the texture data anyway. So I'm not sure if there's really much point to this.