Re: [patch] Add an evaluation function hook to Python breakpoints.

[Phil Muldoon <pmuldoon at redhat dot com>]

Tom Tromey <tromey@redhat.com> writes:

>>>>>> "Phil" == Phil Muldoon <pmuldoon@redhat.com> writes:
>
> Phil> +	  PyObject *gdbpy_bp_eval = PyString_FromString ("evaluate");
>
> This is leaked.  However...
>
> Phil> +	  if (PyObject_HasAttr (py_bp, gdbpy_bp_eval))
> Phil> +	    {
> Phil> +	      PyObject *result = PyObject_CallMethodObjArgs (py_bp,
> Phil> +							     gdbpy_bp_eval,
> Phil> +							     NULL);
>
> You can just use PyObject_HasAttrString and PyObject_CallMethod instead.

Ok, thanks.

> Phil> +gdb_py_test_silent_cmd  "python eval_bp1 = bp_eval(\"$bp_location2\")" "Set breakpoint" 0
> Phil> +gdb_py_test_silent_cmd  "python also_eval_bp1 = bp_also_eval(\"$bp_location2\")" "Set breakpoint" 0
>
> I think there should also be a test for the case where there is an
> ordinary user breakpoint at the same location as the Python breakpoints.
> In this case, the user breakpoint should stop, but all the Python
> methods should be invoked first.

All the breakpoints at a location will be evaluated, but does the order
matter? I can add the test to prove the point, but the Python breakpoints will
have the evaluation function called regardless if the user breakpoint
was first. 

FWIW in bpstop_stop_status a bpstat chain is built from the location
where the inferior stopped.  Then later, we iterate through that chain and
eventually call bpstat_check_breakpoint_conditions.  Then finally in
that condition call we evaluate the Python function.  This loop does not
exit until the all the bpstats in the chain have been iterated.

Cheers

Phil