Ron, > Sorry, but I don't understand, *how* could any OLE control still own > resources created in PRG, *after* such control is explicitly released?
I'm affraid it may happen, at least with buggy OCXs. Long time ago, I had to deal with this same problem. I Xailer user sent me a sample of using certain OCX that GPFs at exit. Then, I discovered the reason, and solves it by calling to OleUninitialize() explicitly before program ends. However, I don't find that sample now, and I cannot test it again, sorry :-( > IMO, Miguel's problem must be one of two possibilities: > > 1. Failure to explicitly release and/or terminate OLE resources [in > correct order]. > > 2. Bug in OLE control, which attempts to access resources *after* > it's been requested to release them. 3. While executing OleUninitialize(), the OCX is sending an event to the program (something like 'OnDestroy'), and the program tries to call the corresponding object method and GPFs. Perhaps this case could be avoided by the OLE container by testing ( hb_vmRequestQuery() == HB_QUIT_REQUESTED ) to not call any object method nor property. Is this correct? > Sorry, but OleUninitialize() should not have any impact if performed > *after* all OLE operations have been completed correctly. Yes, I agree. But, as I've said above, it may happen. Regards, Jose F. Gimenez ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ xHarbour-developers mailing list xHarbour-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xharbour-developers
