For instance, URL http://foo.com/top.html has an <ISINDEX> tag in it.
If the user attempts to search this document for the keywords ``ox''
and ``stoat'', the browser will request the URL
http://foo.com/top.html?ox+stoat. Note that the browser is not
performing the search; it simply requests a URL from a server.
The server must do the search. The standard CERN server doesn't
search at the moment. Plexus, the super sexy perl-based server does.
I have made crude and rude patches to allow searching, but they suck
and have been improved on by other people.
Your options, to perform searches, are:
--- don't
--- run a server that allows searches
--- make a shell script that has the cover-page for the search (the
page with the <ISINDEX> in it) and code to do the searching.
Then run this from inetd on a different port to your standard
server. Then map from the old cover page to the new one with
your server.
--- wait until the new CERN server comes out which will allow
searches.
Cheers;
Nat