Hi Jacek,
thanks, this change in behaviour is not intentional. I will recent with your suggested change.
Regards, Paul.
On 3/1/22 19:45, Jacek Caban wrote:
Hi Paul,
On 2/24/22 14:15, Paul Gofman wrote:
+/***********************************************************************
- * __wine_get_wgl_driver (win32u.@)
- */
+struct opengl_funcs * CDECL __wine_get_wgl_driver( HDC hdc, UINT version ) +{ + BOOL dibdrv; + DC * dc;
+ if (!(dc = get_dc_obj( hdc ))) return NULL; + if (dc->attr->disabled) + { + GDI_ReleaseObj( hdc ); + return NULL; + } + dibdrv = !dc->is_display; + GDI_ReleaseObj( hdc );
+ if (dibdrv) + { + if (version != WINE_WGL_DRIVER_VERSION) + { + ERR( "version mismatch, opengl32 wants %u but dibdrv has %u\n", version, WINE_WGL_DRIVER_VERSION ); + return NULL; + } + return dibdrv_get_wgl_driver(); + } + return user_driver->pwine_get_wgl_driver( version ); +}
I didn't mean to suggest dropping the check for memdc, this changes behaviour over current code. Although such change may make some sense (I'm not sure), it would least deserve a separated commit and consideration about stock bitmap handling in osmesa code. How about doing it like in the attached, modified version of your patch?
Thanks,
Jacek