On, Sun Jun 08, 2008, Brian Fisher wrote: [...] > The problem as I see it is that seg faults and corrupted memory are both > really really bad bugs that I wish I could never get when using python as > they are the sort of bugs whose behavior is inconsistent so you could miss > them when developing but leave in production, and because they are not > strongly tied to their actual cause Sure, this whole issue with the locks could be avoided by a completely reworked locking approach, which however will make it necessary to change large parts of the current pygame code, but will become a necessity by time we use direct pixel access extensively in various parts of the code. > Also, if people aren't sticking to the docs, then the docs aren't effective If people do no read the docs, they should not wonder about any bugs and failures caused by using the stuff in a wrong way. Or would you blame C standard docs for people not reading them and passing wrong stuff to the functions? I agree with you that the whole locking problems we have within the aurfarray and PixelArray code (and other parts maybe), have to be solved. Regards Marcus
Attachment:
pgpXUOTwrB5AU.pgp
Description: PGP signature