Well,
with NTLM looking like it's all-but-finished, and cygwin running along
quite nicely I'm putting my hand up for another small-mid sized squid
project. Anyone have a particular preference as to the next thing tackled so
I don't duplicate effort?
====
Also does anyone have a list of what squid needs to do to be http/1.1
conformant? I don't mean read rfc 2616 !
What I am think of is something like
RFC 2616
13.5.4 a client MUST xyz , squid currently does/does not/sometimes
conforms , estimated difficulty.
If everyone's just scratching their head and saying "well I know x needs
doing" that's cool - just drop me all the infor you have and I'll colate it
and post it to the list. If someone does have the full list I'd be very
interested in looking through it.
Development approach (way down the track here)
I think that squids internal architecture should survive 1.1 being bolted
on - but it might not be pretty :-]
Do we want to end up with a http/1.1 and not a 1.0 proxy? or a 1.1 and 1.0
proxy?
My vote is to support both.
checking the client header should just switch us into the appropriate mode
for each request, and it would allow careful adoption rather than an
all-or-nothing approach.
That raises an interesting question: what should happen when a client sends
a http/1.1 request and then a http/1.0 request on the same connection?
Rob
Received on Mon Oct 16 2000 - 17:02:08 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:43 MST