[Python-Dev] FileCookieJars

[Demian Brecht]

On 2013-03-11 5:44 AM, R. David Murray wrote:

> though some patience
> and persistence may be required.

I have a wife and kids. This, I've become quite good at ;)

> Take a look at http://bugs.python.org/issue2193 (for example), and see
> if you still want to tackle this topic :) (I hope you do).

Egad. I knew that cookies were quite the can of worms prior to digging 
into this as much as I have, but I didn't realize that the RFC had been 
written /after/ cookie implementations had already surfaced in the wild 
(I guess I shouldn't have actually been surprised either). Just makes 
this more challenging and therefore interesting to work on imo :)

> The other reality is that our cookie support won't be very useful if
> it adheres strictly to the RFCs, since the servers and browsers don't.
> What we need is something practical...which may differ to a greater or
> lesser degree from what we currently have.

Yes, I wasn't sure of the general standpoint of Python stdlibs in terms 
of practicality versus strict adherence. While adhering to Postel's law 
in cases such as cookies can definitely make an implementation much more 
tricky, it increases its practical usage (I didn't realize just how 
deviant servers and browsers were for this particular topic until after 
reading through issue 2193).

> But there is, and in fact it *is* useful and used by many people, so
> IMO it is worth maintaining.

I see your point here and agree. It's much different when changes can be 
dictated in closed source packages (what I'm most accustomed to) than 
dealing with an open source project at the scale of Python and the stdlib.

> Excellent.  If you aren't already on the core-mentorship mailing list, you
> might want to sign up.  Your approach (adopting modules without current
> maintainers) is a good one.

Thanks, I wasn't aware of the core-mentorship list. I'll be signing up 
shortly. Good to know my approach is sane :)