[issue10181] Problems with Py_buffer management in memoryobject.c (and elsewhere?)

[Stefan Krah]

Stefan Krah <stefan-usenet at bytereef.org> added the comment:

Thanks for the comments. Most of them should be easy to fix.

Nick Coghlan <report at bugs.python.org> wrote:
> [...] expose the Py_buffer len field as "memoryview.size" instead of
> "memoryview.len" (to reduce confusion with "len(memoryview)" and
> to encourage a conceptual link with sys.getsizeof()).

I agree with this. The best thing is probably to have both versions
work in 3.3 and remove memoryview.len in 3.4 (or later).

> As noted in the review, I also think fixing #12384 should be fairly
> straightforward and it would be nice to have that working when the
> change is applied.

tobytes() currently calls PyBuffer_ToContiguous(..., 'C') from
Objects/abstract.c. Since the function repeatedly calls
PyBuffer_GetPointer(), it has this comment:

    /* XXX : This is not going to be the fastest code in the world
             several optimizations are possible.
     */

So if the fix for tobytes() should go in, I'd like to use the
copy_buffer() function from _testbuffer.c. PyBuffer_ToContiguous()
would need to be fixed separately.

test_buffer.py already asserts that for each result, result.tolist()
and result.tobytes() match the results obtained from PyBuffer_GetPointer(indices)
(See: test_buffer.py:779), so I'm not worried about using the same
implementation in the tests as in the production code.

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue10181>
_______________________________________