>>> Any chance you can take this issue over to the curl-library list and try to
>>> debug the Curl_proxyCONNECT function? Adding a bunch of printfs or whatever
>>> could be a perfect start!
>>
>> I'll gladly mail this to wherever you think is most appropriate but I'd like
>> to avoid having to compile stuff myself here, that would lead me to far...
>
> Fair enough, but then I also don't know how to proceed from here.

I would have expected a debug version or mode to exist, having to write it on the spot seems a bit excessive. However if you're willing and able to walk me through it (getting it to build and telling me where to look in the code and/or getting the printfs in), I'll do it.

Fyi, I posted it to library. The length of the mail probably scared everyone.

>>> I'd say RFC3986 disagrees with you.
>> I'd say RFC 3986 is dated January 2005, whereas the thesis first detailing
>> REST was written in 2000.
> So you argue that you are right and RFC 3986 is wrong?

No, I'm arguing that I think both approaches have conflicting interests and that this should have been untangled in the first place by the creators of RFC 3986 not claiming RFC 3986 is valid for everything and anything. Specifically not expecting this behaviour client side but restricting it as a prescribed behaviour for 'typical' applications.

> It seems like a tough position to take. RFC 3986 may be modern but it is just one out of a series of
> specs out of which the first was published long before that. Also, IETF people
> don't just invent things out of the air.

Indeed and I haven't followed the whole discussion so I don't know how they arrived at their decisions but from my high and mighty position as end user I'm just not convinced this was the right decision. So I don't really want to push hard on this, I just wanted to voice my concerns hoping that someone would be able to take them away with a good argument.

>> Regarding making a patch for it : I'd love to but there's only so many
>> projects I can handle
>
> Right I understand, but we're all in that position! Additionally, just
> providing a convincing argument would take this a long way and after being
> convinced the actual writing of a patch for this would be a small thing.

If convincing arguments would have to take the form of an RFC reference, I'd have to dig for inconsistencies in the specifications or revisions of RFC 3986 which I doubt exist or you would be aware of them. That seems like a job for a lawyer.

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-users
FAQ:        http://curl.haxx.se/docs/faq.html
Etiquette:  http://curl.haxx.se/mail/etiquette.html
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-users
FAQ:        http://curl.haxx.se/docs/faq.html
Etiquette:  http://curl.haxx.se/mail/etiquette.html