[Python-Dev] ssl module integration with asyncore

[Giampaolo Rodola']

On 29 Nov, 03:27, Bill Janssen <jans... at parc.com> wrote:
> > It does raise the same exception.
>
> Hmmm, not in my version.
>
> > Are there plans for fixing this?
>
> Yes, it's fixed in my CVS, and I'll upload a new version to PyPI when
> I get a chance.
>
> > Using that kind of workaround is not acceptable in any case (select
> > module shouldn't even get imported when using asyncore).
>
> Well, that's what the fix in the SSL module will do under the covers,
> so if there's a reason not to do it, or a better way of doing it,
> speak now.  

IMO, it's not reasonable since the application could use something
different than select.select(), like select.poll() or something else
again.
I guess that the best thing to have in such case would be having an
"ssl-ized" wrap of the socket which follows the same behaviour of a
classical non-blocking socket, then it would be up to the application
deciding what to do with it.
That way the asynchronous application which wants to switch to a
secure connection would just need to wrap the socket by using
ssl.wrap_socket() without altering the existing code.

Note that I'm just talking about end-user API design.
Actually I don't know what happens under the hoods so I'm not even
sure which are the real implications.