On, Thu Dec 11, 2008, Lenard Lindstrom wrote: > Hi James, > > I don't have this problem with my antiquated PC. It is an endian issue, > sort of. The rgba mult code is faulty. For a 4 byte source and > destination surface the blit assumes both have identical masks. Not > true. For my Intel machine I get this: [...] That applies to all blend modes for 4 bpp surfaces. We have to respect the RGB layout there. Regards Marcus
Attachment:
pgp5VTfq9hvOD.pgp
Description: PGP signature