> I don't know if it's worth spending time adding more indirection to the > error path so python can print an error. If the error is fatal python isn't > going to be able to correct it and continue.
I don't plan to make it part of the standard run, but I plan to make it either something that happens in #define DEBUG, or add a debug flag (opinions?). The thing is, fatal errors happen sometimes in the C++ code, but you don't know what python thing triggered them. I'm specifically thinking about the drain events and such things that are initiated from python. It's very annoying to get a panic/fatal in C++, but not understand how you got there. In general, I'd like to move m5.opt/m5.fast to print just an error message and never a stack trace, and move m5.debug to either always print a full stack trace (c++ and python) or add a debug option to the simulator for that purpose. Nate _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
