Skip to main content.
home | support | download

Back to List Archive

Re: why can't i find the page that's indexed

From: Bill Moseley <moseley(at)>
Date: Mon May 05 2003 - 17:16:44 GMT
Please CC the list, too.

On Mon, May 05, 2003 at 12:47:03PM -0400, Jean Mao wrote:
> I tried search directly with swish: ./swish-e -w washington and did get hits
> (see below). does that mean i should try your dev version? Thank you very
> much!
> # SWISH format: 2.2.3
> # Search words: washington
> # Number of hits: 6
> # Search time: 0.029 seconds
> # Run time: 0.116 seconds
> 1000 "repeatmasker_help.txt" 97988
> 1000 "HMMER on Biowulf" 13626
> 974 "Documentation of TREE-PUZZLE 5.0"
> 72756
> 974 "hmmer_homolog.out" 168124
> 878 "hmmer_pfam.out" 23495
> 878 "RepeatMasker on Biowulf" 2349

Where's the "." at the end of the list?  Did you just not copy that in the email?  If that 
was missing you would probably get that error.  The script looks for the lone "." to know 
that it has seen all of the results.

Did you look at the web server's error log?  It would be nice to see why it's failing.

There's a debugging section in the swish.cgi manual.  It's titled "DEBUGGING" so should be 
easy to search for.

In short, you can run the program interactively outside of the web server.

So you could do (assuming you are running a borne-type of shell like bash):

           $ SWISH_DEBUG=basic ./swish.cgi >outfile
           Debug level set to: 1
           Enter a query [all]: washington
           Enter max results to display [1]: 20

and then see what happens.  Look at "outfile" to see what output would have gone to the web 
client, and any errors will be displayed on your console unless you also redirect stderr.

If that works then you have a problem with your environment.  Maybe you have an older 
version of swish-e installed and when running under the web server you are using that 
instead of the newer 2.2.3 swish-e binary.

Bill Moseley
Received on Mon May 5 17:17:12 2003