>
> I didn't know whose curly bracket style that was, yours
> or openssl's. As long as we don't adopt the thousand
> line function style of openssl, I'll be happy. :-)
>
> #149 looks good. If Py_BuildValue returned NULL, it
> was like the alias didn't exist. I contemplated an
> error message, but chose not to. With the new code
> maybe we should.
>
I wasn't sure how that Py_BuildValue might be induced to fail. I thought it would have to be
an internal CPython malloc failure or something like that. If the alias doesn't exist, doesn't
that mean the whole block with Py_BuildValue will be skipped? Or am I misunderstanding something?
>
> I didn't know whose curly bracket style that was, yours
> or openssl's. As long as we don't adopt the thousand
> line function style of openssl, I'll be happy. :-)
>
> #149 looks good. If Py_BuildValue returned NULL, it
> was like the alias didn't exist. I contemplated an
> error message, but chose not to. With the new code
> maybe we should.
>
I wasn't sure how that Py_BuildValue might be induced to fail. I thought it would have to be
an internal CPython malloc failure or something like that. If the alias doesn't exist, doesn't
that mean the whole block with Py_BuildValue will be skipped? Or am I misunderstanding something?
Jean-Paul