--- In [email protected], "swzoh" <sean...@...> wrote: > > --- In [email protected], "entropyreduction" > <alancampbelllists+yahoo@> wrote: > > > > The bad news is I'm avoiding the problem by simply not calling > > CoUninitialize() before my worker thread dies on a com.unload. > > So if you call com services and unload it alot, Something Bad might happen. > > Dunno what the Something Bad might be. > It could cause problems with local servers like Excel, Word, etc. If users > doesn't do clean-up completely via Release, then a lot of local servers will > be hanging around, not being closed.
unload causes all retained pointers to com objects to be Released. So in theory com plugin shouldn't tie up any servers. However, my current theory re CoUninitialize hanging is that somehow I'm indeed leaving something unreleased, and CoUninitialize won;t let go until I do. Can;t for life of me see what I've left undone. BTW if I modify any scripts to only do a get_object, then some methods, then unload, all's well. A script with only a create_object and an unload hangs. Therefore it's got to be my create_object code that's a problem...
