calldll and threads...
Steve
sjh at shic.co.uk
Wed Mar 27 05:44:26 EST 2002
More information about the Python-list mailing list
Wed Mar 27 05:44:26 EST 2002
- Previous message (by thread): calldll and threads...
- Next message (by thread): communication with a java applet
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Mark Hammond <mhammond at skippinet.com.au> wrote in message > This is a common problem, not just with calldll. Although you released > the lock OK, you will probably find that you did not re-acquire the lock > as calldll made callbacks into your code. What was the callstack when > you crashed? I won't bother you with the call stack (which was most uninteresting - depth 2 in the pre-compiled python.exe when it had an address exception)... I hadn't realised that the C-API calls to PyArg_ParseTuple() or Py_BuildValue() constituted call-backs into Python in the context of the global lock. (I think it would be an improvement if the manual made this point explicitly.) Anyhow... my trivially modified calldll module is now happily allowing me to make blocking calls in spawned threads without blocking all my python threads. All's cheerful this bright spring morning - thanks... Steve
- Previous message (by thread): calldll and threads...
- Next message (by thread): communication with a java applet
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Python-list mailing list