Re: [dav-dev] WebDAV, IE5 and Squid?

[Discussion moved to squid-dev@ircache.net. I think posting to this list
is open for non-subscribers, but perhaps not.. if it isn't then send any
replies to squid-bugs@ircache.net]

As you note Squid is not an HTTP/1.1 proxy, and any request passed
throught Squid is automatically downgraded to HTTP/1.0 as required by
HTTP specifications.

I agree that this matter of "unknown" or "extension" methods should be
dealt with. As I understands it the correct way (even for HTTP/1.0 to
some extents) is to forward the request and invalidate any caced
object(s) on the same URI.

---
Henrik Nordstrom
Spare time Squid hacker
Greg Stein wrote:
>
> Mike Barnes wrote:
> >
> > I'm coming in late here, but I did a quick skim over the archives and
> > couldn't find any mention of this ... apologies if I missed something and
> > I'm bringing up the obvious.
>
> Nope, first time, so no worries :-)
>
> > After hours of bashing my head against my DAV config, I had the bright idea
> > of turning off my proxy in IE5. Everything worked. Hooray. The Squid logs
> > revealed a whole bunch of "error:unsupported-request-method" messages.
> >
> > So, where does the problem lie? Does Squid need to be 'DAV-Aware'? Is it
> > Microsoft's fault? What should I start prodding and hacking? It's not a huge
> > issue to exclude my DAV-happy servers from the proxy list, but it'd be nice
> > to have things working 'perfectly'. :)
>
> Given that IE5 and mod_dav interoperate fine (so far that I've seen),
> and that it started to work for you once you took Squid out of the
> picture... I'd have to say that Squid probably does not deal with the
> DAV methods properly.
>
> I'm copying this email to the squid-users mailing list. I looked thru
> the Squid mail archive for the month of March and couldn't find anything
> that looks DAV-related, so this is probably news to them, too. I also
> used their search tool. Still nothing so far.
>
> Squid probably needs to simply pass-thru any HTTP methods that it does
> not understand.
>
> Ah. Some related information... found in the HTTP/1.1 spec (draft 6),
> section 13.10:
>
> "A cache that passes through requests for methods it does not understand
> SHOULD invalidate any entities referred to by the Request-URI."
>
> There are two caveats here: Squid advertises itself as an HTTP/1.0 proxy
> (and DAV requires HTTP/1.1), and the above text from 13.10 doesn't state
> that a proxy MUST pass through requests for unknown methods.
> [ oh. just read further. section 13.11 says the method must be passed
> through. ]
>
> Whatever... again, the "right" thing to do is pass them through. We'll
> see what the squid developers say :-)
>
> > Anyone dealing with this sort of thing?
>
> Not so far. Thanx for reporting it! I hope this can be tracked down and
> fixed.
>
> Squid guys: please let me know how I can help you.
>
> Cheers,
> -g
>
> --
> Greg Stein, http://www.lyra.org/
>
> _______________________________________________
> dav-dev maillist  -  dav-dev@lyra.org
> http://dav.lyra.org/mailman/listinfo/dav-dev