The hardware breakpoint handler is executed only once and then the debugger stops the application.

Seems like the single step exception handler does not reset _restore_breakpoint properly. The next changes to pydbg.py worked for me.

========================================= --- pydbg.py.orig 2009-03-12 22:41:02.000000000 -0400 +++ pydbg.py 2009-03-12 22:51:45.000000000 -0400 @@ -1698,7 +1698,10 @@

     else:

continue_status = DBG_EXCEPTION_NOT_HANDLED

• • reset restore breakpoint variable.

• self._restore_breakpoint = None

•  # if we are handling a hardware breakpoint hit and it still
  

  exists, ie: the user didn't erase it during the # callback, and the breakpoint is flagged for restore, then tell the single step handler about it. furthermore, # check if the debugger is still active, that way we don't try and single step if the user requested a detach. @@ -1714,11 +1717,11 @@ self._restore_breakpoint = prev self.single_step(True)

+ self.bp_del_hw(slot=prev.slot)

• reset the hardware breakpoint hit flag and restore breakpoint

  variable.
• self.hardware_breakpoint_hit = None
• self._restore_breakpoint = None

• reset the hardware breakpoint hit flag.

      f.hardware_breakpoint_hit = None
  

return continue_status