Tim Peters wrote: > PythonLabs never found the bandwidth for this, as we all spent the > next year trying to teach Barry how to use Unicode. I don't even > remember what the thread above was about, but it would be great if the > PEP contained enough relevant words that I could finally clean that > message out of my inbox ;-)
Sure: I think many of these cleanup points overlap, in the sense that cleanup should get invoked if the module object is discarded. I can put that in the discussion of the PEP (and it's in the python-3000 archives meanwhile to remind me): > - when the module object is deleted from sys.modules This may not have immediate effect, since somebody might still holding a reference to the module. A new import will create a new module object, so I guess it would be best if indeed the state lives in the module object. > - when reload() is used on an extension This is trickier: I guess reload should invoke the finalize function, discard the memory for the state, and then invoke the initialize function again, which will allocate a new state block. > - when an interpreter created with PyInterpreter_New() is finalized This shouldn't cause any module finalization directly. Instead, sys.modules of the interpreter gets cleared, which should cause the module objects to be released. > - when Py_Finalize() is called Likewise. Of course, a forced clearing of all modules might be desirable, since the modules might keep a cycle alive. This raises the question whether modules should be able to participate in cyclic GC... > - when Python exits That should not cause action: Py_Finalize should clear it all. > - when the Python DLL is unloaded (Windows only) IMO, the DLL should only get unloaded when all modules referring to it have been unloaded. So, when the DLL is unloaded, no action should be taken. I also think it's not Windows only: on dl, we should invoke dlclose when all modules have been finalized. Regards, Martin _______________________________________________ Python-3000 mailing list Python-3000@python.org http://mail.python.org/mailman/listinfo/python-3000 Unsubscribe: http://mail.python.org/mailman/options/python-3000/archive%40mail-archive.com
