Alexander Belopolsky wrote:
> The patch looks good. Just a question: I thought the strings returned
> by PyUnicode_AsStringAndSize are 0-terminated, while your patch at
> several places attempts to explicitly 0-terminate a copy of such string.
> Are you sure this is necessary?
I wasn't sure if the strings returned by PyUnicode_AsStringAndSize were
0-terminated, so I didn't take any chance and explicitly terminated
them. But I just verified myself and they are indeed 0-terminated. So, I
modified the patch in consequence.