Re: [AD] OpenGL and upside down

[Peter Wang <novalazy@xxxxxxxxxx>]

On 2008-03-18, Peter Hull <peterhull90@xxxxxxxxxx> wrote:
> On Tue, Mar 18, 2008 at 2:43 AM, Peter Wang <novalazy@xxxxxxxxxx> wrote:
> > Correct me if I'm wrong, but OpenGL shouldn't be aware of any "negative
> > pitch" at all.  It's just any Allegro and user code which locks a
> > rectangle needs to keep in mind that ptr += rect.pitch might actually
> > decrement ptr.
> That's what I was thinking; Allegro would set itself to suit OpenGL,
> rather than the other way around.

I was curious so had a quick test :-)  The examples mainly work except for text
rendering.

Peter

--- src/opengl/ogl_bitmap.c     (revision 11437)
+++ src/opengl/ogl_bitmap.c     (local)
@@ -290,7 +290,7 @@ static ALLEGRO_LOCKED_REGION *ogl_lock_r
          }
          glPixelStorei(GL_PACK_ROW_LENGTH, pack_row_length);
          //FIXME: ugh. isn't there a better way?
-         upside_down(bitmap, x, y, w, h);
+         if(0) upside_down(bitmap, x, y, w, h);
       }
       else {
          //FIXME: use glPixelStore or similar to only synchronize the required
@@ -305,9 +305,9 @@ static ALLEGRO_LOCKED_REGION *ogl_lock_r
       }
    }
 
-   locked_region->data = bitmap->memory + pitch * y + pixel_size * x;
+   locked_region->data = bitmap->memory + pitch * (y + h - 1) + pixel_size * x;
    locked_region->format = bitmap->format;
-   locked_region->pitch = pitch;
+   locked_region->pitch = -pitch;
 
    return locked_region;
 }
@@ -328,7 +328,7 @@ static void ogl_unlock_region(ALLEGRO_BI
    if (ogl_bitmap->is_backbuffer) {
       GLint unpack_row_length;
       //FIXME: ugh. isn't there a better way?
-      upside_down(bitmap, bitmap->lock_x, bitmap->lock_y, bitmap->lock_w, bitmap->lock_h);
+      if(0) upside_down(bitmap, bitmap->lock_x, bitmap->lock_y, bitmap->lock_w, bitmap->lock_h);
 
       /* glWindowPos2i may not be available. */
       if (al_opengl_version() >= 1.4) {