freewais and WWWWais

junga@informatik.tu-muenchen.de (Achim Jung)
Errors-To: listmaster@www0.cern.ch
Date: Mon, 13 Jun 1994 22:07:23 +0200
Errors-To: listmaster@www0.cern.ch
Message-id: <2tie48$k4c@hpsystem1.informatik.tu-muenchen.de>
Errors-To: listmaster@www0.cern.ch
Reply-To: junga@informatik.tu-muenchen.de
Originator: junga@hphalle10.informatik.tu-muenchen.de
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: junga@informatik.tu-muenchen.de (Achim Jung)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: freewais and WWWWais
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Organization: Technische Universitaet Muenchen, Germany
Organization: Technische Universitaet Muenchen, Germany

Last week I tried to install WWWWais on www.informatik.tu-muenchen.de
Compilation and installation worked fine ... BUT:

	freewais creates corrupt index-files :-(((

(Happens on version 0.202 und 0.3 ! (with HP-UX 9.01))

The indexing was done by

 	find $top_dir/* -name "*.html" -print | xargs \
 	waisindex -d $index -a -t URL $top_dir $top_url 

whereas the vars are set to:

 	top_dir 	document-root
 	top_url 	URL zum document-root
 	index		Name des Index

The first waisindex was called by "-export" rathen than "-a". 
(This will overwrite an existing index.)

>From the beginning of indexing this warning appears:
 	Warning: couldn't open /usr/proj/infosys/wais-db/html.syn - translation disabled
(Here: $index = /usr/proj/infosys/wais-db/html )
The created index only works for very few words. In most cases freewais doesn`t find
any hits or dumps core :-(

Now I`ve done this before indexing:
 	touch /usr/proj/infosys/wais-db/html.syn
After some time of indexing (does not depent on the order of the indexed files) this
error-message appears:
 	 100: Total word count for dictionary is: 15929
 	  -1: total_word_count not found in dictionary. This is either an error, or the database is old. 0 out of 1

Needless to say that freewais sucks like above :-(

Can anybodey tell me what I`ve done wrong???
I`m using freewais in some other cases where it works fine ...

Ciao, Achim
-------------------------------------------------------------------
Achim Jung        IRC: Flops        junga@informatik.tu-muenchen.de
http://www.informatik.tu-muenchen.de/~junga/          junga@leo.org 
http://www.leo.org/~junga/                            flops@leo.org