[RFC] Python Finish Breakpoints

[Kevin Pouget]

On Fri, Oct 21, 2011 at 9:45 AM, Kevin Pouget <kevin.pouget@gmail.com> wrote:
> On Thu, Oct 20, 2011 at 10:41 PM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>>> "Kevin" == Kevin Pouget <kevin.pouget@gmail.com> writes:
>>
>> Tom> You need to store a gdb.Type wrapper.
>> Tom> A 'struct type' can also be invalidated when an objfile is destroyed.
>>
>> Kevin> I wouldn't mind, but I can't see how gdb.Type ensures validity, as far
>> Kevin> as I've seen, there is no "is_valid" method and I can't and no further
>> Kevin> verification during the Python -> C translation:
>>
>> When an objfile is destroyed, py-type.c:save_objfile_types is called.
>> This makes a copy of the struct type underlying any relevant Type object.
>
> I'll take a look at that, I don't remember what was the problem!

I understood that, and replaced 'struct type *' fields by the
corresponding 'PyObject/Type' python objects

>> [ out_of_scope_notif ]
>> Kevin> - avoid calling `out_of_scope' every normal_stop when the breakpoint
>> Kevin> is not anymore in the callstack
>>
>> It seems to me that manually managing this is not the best we could do.
>>
>> I was re-reading the code and I realized that this isn't actually
>> working the way I had remembered it working:
>>
>> +A @fdb{finish breakpoint} is a breakpoint set at the return address of
>> +a frame, based on the "finish command. @code{gdb.FinishBreakpoint} extends
>> +@code{gdb.Breakpoint}
>>
>> This approach is fine; but once a frame is gone, it is gone -- we should
>> just destroy the breakpoint at that point.  Maybe this would make the
>> internal logic simpler as well.
>
> yes, I think you're right with that, "once a frame is gone, it is
> gone", I actually don't use myself this 'permanent finish breakpoint'
> feature, so I'll change it.
> So basically, the BP must be deleted right after it has been it (maybe
> through the 'temporary' mechanism). 'out_of_scope' will be triggered
> in the same conditions as before, but instead of playing with
> 'out_of_scope_notif' I'll delete the GDB breakpoint (thus invalidating
> the Python BP)
>
> I'll post a patch going this way later today

So I've changed the way FinishBreakpoints are 'garbage collected'.
Now, they're more similar to Temporary Breakpoints, that is, they're
deleted *soon* after they've been hit.
Soon, because when Breakpoint.stop tells GDB *not* to stop at the
[finish] breakpoint, there is no way to delete this breakpoint (and I
guess that may be the reason why Temporary Breakpoints are not
available in Python yet).

So I disable it, and "b->disposition == disp_del" will ensure that the
BP will be removed when it's possible.

The `out_of_scope' callback will also delete the breakpoint.


I've updated the doc and testsuite accordingly, and as far as I've
seen, there is no regression on X86_64/Fedora
(I need still need to improve my test environment)


Thanks,

Kevin

2011-10-24  Kevin Pouget  <kevin.pouget@st.com>

	Introduce gdb.FinishBreakpoints in Python

	* Makefile.in (SUBDIR_PYTHON_OBS): Add py-finishbreakpoint.o.
	(SUBDIR_PYTHON_SRCS): Add python/py-finishbreakpoint.c.
	Add build rule for this file.
	* breakpoint.h (struct breakpoint): New field: is_py_finish_bp.
	* breakpoint.c (init_raw_breakpoint_without_location):
	Initialize is_py_finish_bp.
	* infcmd.c (get_return_value): New function.
	(print_return_value): Split to create get_return_value.
	* inferior.h (get_return_value): New prototype.
	* infrun.c: Include python/python.h.
	(stop_registers): Mention FinishBreakpoint in description.
	(normal_stop): Set stop_registers if stopped at FinishBreakpoint.
	* python/py-breakpoint.c (breakpoint_object_type): Make non-static.
	(bppy_pending_object): Likewise
	(gdbpy_should_stop): Disable temporary breakpoints.
	(gdbpy_breakpoint_created): Set is_py_finish_bp is necessary.
	(struct breakpoint_object): Move to python-internal.h
	(BPPY_REQUIRE_VALID): Likewise.
	(BPPY_SET_REQUIRE_VALID): Likewise.
	* python/python-internal.h (breakpoint_object_type): Add as extern.
	(bppy_pending_object): Likewise.
	(typedef struct breakpoint_object) Removed.
	(struct breakpoint_object): Moved from py-breakpoint.c.
	(BPPY_REQUIRE_VALID): Likewise.
	(BPPY_SET_REQUIRE_VALID): Likewise.
	(frame_object_to_frame_info): New prototype.
	(gdbpy_initialize_finishbreakpoints): New prototype.
	(bpfinishpy_is_finish_bp): Likewise.
	* python/py-finishbreakpoint.c: New file.
	* python/py-frame.c(frame_object_to_frame_info): Make non-static and
	accept PyObject instead of frame_object.
	(frapy_is_valid): Don't cast to frame_object.
	(frapy_name): Likewise.
	(frapy_type): Likewise.
	(frapy_unwind_stop_reason): Likewise.
	(frapy_pc): Likewise.
	(frapy_block): Likewise.
	(frapy_function): Likewise.
	(frapy_older): Likewise.
	(frapy_newer): Likewise.
	(frapy_find_sal): Likewise.
	(frapy_read_var): Likewise.
	(frapy_select): Likewise.
	* python/python.c (gdbpy_is_stopped_at_finish_bp): New noop function.
	(_initialize_python): Add gdbpy_initialize_finishbreakpoints.
	* python/python.h: Include breakpoint.h
	(gdbpy_is_stopped_at_finish_bp): New prototype.

doc/
	* gdb.texinfo (Breakpoints In Python): New subsection: Finish
	Breakpoints.

testsuite/
	* gdb.python/py-breakpoint.exp (mult_line): Define and use variable
	instead of line number.
	* gdb.python/py-finish-breakpoint.c: New file.
	* gdb.python/py-finish-breakpoint.exp: New file.
	* gdb.python/py-finish-breakpoint.py: New file.
	* gdb.python/py-finish-breakpoint2.cc: New file.
	* gdb.python/py-finish-breakpoint2.exp: New file.
	* gdb.python/py-finish-breakpoint2.py: New file.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Introduce-gdb.FinishBreakpoints.patch
Type: text/x-patch
Size: 53570 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/gdb-patches/attachments/20111024/ac829909/attachment.bin>