I assume that the 'generated URLs' are coming from swish.cgi. Can you see what 
the cmd line returns?

% swish-e -w shell

?

You probably need to configure swish.cgi to set the correct base URL, and/or 
play with the ReplaceRules config option.

John Young scribbled on 3/15/05 1:02 PM:
> Hello,
> 
>     I am trying to set up Swish-e 2.4.3 on a Solaris box to 
> use with local copy of the UNIXhelp web pages from the University 
> of Edinburgh (http://unixhelp.ed.ac.uk/index.html) on a non-public 
> web server.  The server is running Apache httpd 2.0.53.
> 
> My swish.conf file looks like this:
> #
> # SWISH configuration file
> #
> 
> IndexDir         /usr/opt/MSMBweb/htdocs/Edinburgh
> IndexFile        /usr/opt/MSMBweb/htdocs/Edinburgh/index.swish
> IndexName        "Index of UNIXhelp 1.3"
> IndexDescription "This is a full index of UNIXhelp release 1.3."
> IndexPointer     "http://msmb.larc.nasa.gov/cgi-bin/unixhelp_search"
> IndexAdmin       "webmaster"
> 
> IndexOnly .html
> IndexReport 3
> NoContents .gif .xbm .au .mov .mpg
> IgnoreLimit 70 200
> #
> 
> I tried indexing the pages by doing:
> 
> %  ../cgi-bin/swish-e -i ./Edinburgh -c ./Edinburgh/swish.conf
> 
> Which resulted in a great deal of output ending with:
> Removing very common words...
>   Getting IgnoreLimit stopwords: Complete                            
> 13 words removed by IgnoreLimit:
> a, gov, by, of, to, larc, the, site, maintained, this, msmb, nasa, help, 
> Writing main index...
> Sorting words ...
> Sorting 3,952 words alphabetically
> Writing header ...
> Writing index entries ...
>   Writing word text: Complete
>   Writing word hash: Complete
>   Writing word data: Complete
> 3,952 unique words indexed.
> 4 properties sorted.                                              
> 884 files indexed.  1,307,022 total bytes.  87,710 total words.
> Elapsed time: 00:00:07 CPU time: 00:00:07
> Indexing done!
> 
> It seems to have worked, and searching for, say, "shell"
> produces a reasonable list of pages.  But the generated URLs 
> are incorrect, e.g. the first result is for
> 
> "1. Using the Bourne shell to interpret a shell script"
> 
> and the associated URL is 
> 
> http://msmb.larc.nasa.gov/cgi-bin/Edinburgh/scrpt/scrpt1.2.4.html
> 
> but it *should* be
> 
> http://msmb.larc.nasa.gov/Edinburgh/scrpt/scrpt1.2.4.html.
> 
> I cannot figure out why "cgi-bin" is being incorrectly inserted into
> the results.  I have searched the documentation on http://swish-e.org
> but I still have not found the answer.  Any suggestions?  Is anyone 
> else on this list using the Edinburgh UNIXhelp pages with Swish-e?
> 
> 						JY
> ------------------------------------------------------------
> John E. Young				    B1148/R202
> Analytical Services and Materials, Inc.     (757) 864-8659
> 
> 
> 

-- 
Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com