Peter Kümmel wrote: >> my understanding was that the key problem is shared graph >> member of converters class and nowhere else. >> do we agree here? > > I don't know if Graph is used nowhere else, it happens so much > even when one only presses a key.
- all these graph backtraces i have seen up to now were going though converters. - when quickly went through export routines, all graph variables were local to the functions. - when studying possible conflict we both ended at the global converters object where is the conflict visible. - do we agree that locks are not enough in the current version. >We are on the secure side when we lock Graph. well i feel those ten locks ugly and deadlock prone (even if in future). they are just put there without analysis how exactly are the routines inside graph entangled (correct me if wrong). > But it is a principle question should we lock button-up or top-down? > Is there a general answer? i doubt and believe its different case by case, picking between effectiveness and error-prone scenarios. pavel
