On Wed, 19 Aug 2009, Michael Wood wrote:

>> I argue that specifying the order is unnecessary. Am I wrong?
>
> Yes, I should think it would be silly for a server to rely on something like
> this, but I can't comment on whether or not any server actually does rely on
> it.

Unfortunately calling things by their true name (silly!!) doesn't always help
in the world of standards and web people taking shortcuts.

Apparently IE once started that sorting and apparently all the major browsers
followed that and apparently there are a small fraction of servers out there
today that break if you pass them cookies in another order. There's a rather
heavy push towards putting that sorting in the draft-in-progress for how
cookies work and should work. Even though it would _also_ say that servers
MUST accept them in any order.

libcurl has supported cookies for almost 11 years now and I've not yet once
been confronted with such a site...

-- 
  / daniel.haxx.se