[Python-Dev] Commit-ready patches needing review

[Antoine Pitrou]

On Sun, 27 Apr 2014 12:10:46 -0700
Nikolaus Rath <Nikolaus at rath.org> wrote:
> 
> * http://bugs.python.org/issue20951 (SSLSocket.send() returns 0 for
>   non-blocking socket)
> 
>   In this case someone just needs to decide if we want to (a) document
>   the current behavior, (b) deprecate the current behavior or (c) change
>   the current behavior. I have attached patches for (a) and (b), and if
>   (c) is the desired route I'll be happy to create a patch on short
>   notice.

In this case I'd be inclined to follow Ben Darnell's advice and change
the current behaviour (i.e., let the exception bubble up rather than
catch it). This is what your initial patch does. However, it would need
a documentation addition to explain the change (and perhaps a test,
though that doesn't seem terribly necessary here).

Regards

Antoine.