On, Tue Jan 08, 2008, Lenard Lindstrom wrote: [...] > The problem is that the major and minor axis of the array are reversed from > those of the surface. Try replacing lines 272-276 with: > > return numpy.ndarray(buffer=surface.get_buffer (), > dtype=numpy.uint8, > shape=surface.get_size (), > offset=start, > strides=(4, surface.get_pitch ())) > > The offset argument sets the relative position of the alpha byte. The > strides argument makes the second argument the major axis and the first > argument skip by the pixel size of 4 bytes. It works with the test case I > submitted and an application I wrote that relies heavily on arrays. Hope > this helps. You refer to rev. 1072 I guess, not to the fixed version? I cannot see any misbehaviour anymore. Regards Marcus
Attachment:
pgpBLH06Pnpco.pgp
Description: PGP signature