On Tue, 2 Mar 2004, Edward S. Peschko wrote:

> The fact that curl doesn't handle automatic proxy configuration - which is
> standard in all browsers - *does*.

Javascript is indeed a browser-thing. curl is not a browser.

> That's why it should support it - its directly in the problem domain that
> curl sets up for itself.

I disagree with that description. curl deals with proxies, it does not deal
with figuring out which proxy to use. Most sites that use proxies have the
proxy info stored somewhere. Some of them use this horrible "pac" stuff, other
use static host names or whatever. In all these cases, the curl user needs to
get the info and provide it to curl.

At least that is how it works today.

> 2) I can't use curl because of this authentication method that curl
> doesn't support.

This is not authentication.

> But I hope that what I'm saying doesn't get ignored, or swept under. Like I
> said, I'm not the person to code this in the first instance, but I can give
> testing and porting support.

curl is free and open source. No one gets paid anything for making it. I don't
see a huge personal gain in spending a whole lot of time and energy on a
feature you want to use to save yourself (or your company) time, effort and
energy.

I usually add features I miss myself or that I think is fun to write. On my
spare time.

It's generally the same reason you don't see this feature as you don't see a
bunch of other features too: no one has yet missed them so much they
implemented it.

If anyone wanted it badly enough but couldn't do it because of a lack of skill
(or similar), it would be very easy to pay someone to provide the necessary
code.

Maybe one day someone wants this badly enough to make it a feature of curl. Or
maybe not.

I would probably argue that what you want is basicly a tool that can get a
HTML page and interpret a limited set of javascript that in the end produces a
host name (and portnumber etc). I don't think that tool is curl. It could just
as well be a libcurl + javascript interpreter tool that can fetch a proxy name
to be used with whatever tool that needs to know about the proxy. It would
make more sense than having it built inside curl. IMHO. All this very
hypothetical of course.

-- 
    Daniel Stenberg -- http://curl.haxx.se/ -- http://daniel.haxx.se/
   [[ Do not send mails to this email address. They won't reach me. ]]