[Adium-devl] Facebook/Safari

[Evan Schoenberg]

On Jan 5, 2009, at 11:52 AM, Zachary West wrote:

> If it passes the above test (as well as the one I wrote[1]) then it  
> isn't stripping outgoing cookies. When I tested - 
> [NSMutableURLRequest setHTTPShouldHandleCookies:NO], my cookie test  
> passed as valid as well, with the semi-bad patch[2] I used to test it.

I wrote AICookieJar to be more generic, with the goal of being able to  
contribute it elsewhere (e.g. Flock, where it's in demand afaik), but  
used fundamentally the same approach as your patch.

> I don't think switching to Pidgin is a great move, since we  
> implement nicer the "protocol", as well as the extra benefits of the  
> Info window, etc. It's an option, of course, and the interaction  
> between Safari and Facebook is really annoying.

Actually, I think that the pidgin-facebook implementation has more  
thorough protocol-level error checking at this point.

I haven't used it, but looking at the code, it appears to have the  
info window stuff just like I did for Adium, as well, and it also has  
a few features we don't implement (e.g. buddy management). I think  
it's feature-equivalent-plus. Eion, is that right?

> It doesn't strip outgoing cookies, we're just not behaving with  
> Facebook correctly for some reason (they submit multiple cookies to  
> the browser by the same name, I'm not sure if this is the issue).
> I've been trying to figure this one out too, and I've had no luck so  
> far. It's confusing, to say the least. Perhaps we're sending the  
> cookies outgoing wrong, for multiple cookies? That case, I didn't  
> test.


pidgin-facebook uses a hash table by name to track the cookies to  
send, so it shouldn't be the case that it's sending multiple cookies  
with the same name... and yet it works.

David wrote:
> Random possibility would be to check that we aren't somehow  
> disabling cookies for Adium via a preference we set on one of our  
> webviews.

We're not... everything is fine if you let that particular  
NSConnection do its default cookie handling.

> Other than that, I like the idea of less code to compile/maintain,  
> and fixing bugs "for free", but I reaaaaaally wish facebook would  
> get off their butts and get that xmpp server working so we could  
> drop the whole thing.

Indeed.

-Evan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://adium.im/pipermail/devel_adium.im/attachments/20090105/f18e351f/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 194 bytes
Desc: This is a digitally signed message part
URL: <http://adium.im/pipermail/devel_adium.im/attachments/20090105/f18e351f/attachment.sig>