Hi, Petr found a bug when generating a backtrace from inside the Code observer updateHit() method. This didn't show up in other places since we would set the pc up correctly immediately after we detected a breakpoint, but after calling the Code observers. It would also only show if you were unlucky enough to set a breakpoint on an address right before the dwarf cfi info would mark a cfa change. Of course Petr's test hit both cases. Fixed as follows: frysk-core/frysk/proc/live/ChangeLog 2008-04-11 Mark Wielaard * LinuxPtraceTaskState.java (Running.setupSteppingBreakpoint): Removed. (Running.handleTrappedEvent): Don't call setupSteppingBreakpoint(). (Stepping.handleTrappedEvent): Don't do stepping breakpoint sanity check. Don't call setupSteppingBreakpoint(). * LinuxPtraceTask.java (notifyCodeBreakpoint): Add stepping breakpoint sanity check. Set task pc when breakpoint found. Set steppingBreakpoint. frysk-core/frysk/stack/ChangeLog 2008-04-11 Mark Wielaard * TestFrame.java (testBogusAddressPrevFrame): Resolved. No regressions on x86 and x86_64 and the test now passes on both architectures. Committed and pushed, Mark