Gerben Welter wrote:
>
> At 11:40 16-6-00 +0300, you wrote:
>
> >Hello,
> >I have a question:
> >
> >How can squid be told to leave a URL request such as amazon
> >in, eg, netscape, to go on its way and become http://www.amazon.com ?
>
> Browsers like Exploder and Netscape do this by themselves. But I think that
> only works if the browsers are configured for direct of transparant proxy
> access. No need for squid to start guessing the right url.
>
> Grtz Gerben.
Our proxy isn't trasparent. So people who manually put the
proxy in their browsers, loose the following functionality:
If one puts just a word in the URL field of netscape, it
makes a query to keyword.netscape.com and then
- if there's a site www.'word'.{com,org,...)
it redirects you there.
- If it's ambiguous (eg. squid-cache) it brings you to a list of choices.
You can also put a phrase in there, and have it searched with some engine.
Phrases aren't forwarded to the proxy, but single words are.
For instance, I put "sun" in the url field, and
- if I have the proxy configured, squid complains it can't resolve "sun"
- if I have "direct connection", I come to www.sun.com
What I imagine as a solution, is to have squid ignore anything that doesn't
have a dot in. But I can't say how I could instruct it to do so.
"Jens-S. Voeckler" wrote:
> Anyway, if you still insist, you could let Squid append a domain suffix,
> see squid.conf.default. But, choosing a .com suffix, and there is no
> server http://amazon.com/, you are out of luck.
>
> Alternatively, a redirector (an external program) might yield the desired
> output.
A redirector, or append_domain would try to do netscape's work, while what
would be preferable is just have such queries passed through squid, on to
keyword.netscape.com
Thanks for your replies,
regards,
Thanos Siaperas
NOC / AUTH / Greece
thanus@auth.gr
Received on Fri Jun 16 2000 - 05:27:31 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:03 MST