Hi,
I'm using Debian so the installation was performed using the Debian package
manager. I used 'dpkg -s htdig' to find out the version of the htdig
installation (which dpkg reported as 3.2.0b4-1). The htdig search results
page reports the version as 3.2.0b4-20020915. Sorry for any confusion this
caused.
The symptoms of the problem are a worse than I originally described. All of
the following phrase searches:
"and testing"
"or testing"
"testing and"
"testing and something"
"testing not something"
Resulted in an internal server error being reported (after the browser sat
there doing nothing for quite a while). This only happened when using the
phrases with a method of "boolean" (the search methods "or" and "and" work
fine).
Also, the phrase "or testing" failed as well, "or" is not set in my
bad_words file. The phrase "from testing" and "find from testing" worked
alright ("from" is in my bad_words file).
Originally I thought the problem was with a cgi wrapper that I was using to
validate htsearch input. But I still get the problem after using the
search.html page to directly invoke htsearch.
I will try and update to a more current version and check if I still have
this problem.
thanks for your help,
Andrew Patterson
-----Original Message-----
From: Jim Cole [mailto:greyleaf@;yggdrasill.net]
Sent: Saturday, 19 October 2002 15:27
To: Geoff Hutchison
Cc: Patterson, Andrew; '[EMAIL PROTECTED]'
Subject: Re: [htdig] SEC: UNCLASSIFIED: Problem with 3.2.0b4-1
Geoff Hutchison's bits of Fri, 18 Oct 2002 translated to:
>> for additional stop words. If I recall correctly this problem
>> arises anytime that a phrase search contains a stop word that is
>> bracketed by non-stop words.
>
>This should be fixed in current snapshots. What snapshot of 3.2.0b4 are
This still appears to be a problem as of 3.2.0b4-20021013. Just
to make sure we are talking about the same thing, if a site
contains something like
This time period created much confusion and dissension among ...
a query for "confusion and dissension" will result in no matches.
A query for "and dissension among" matches 'dissension among'.
A query for "much confusion and" matches 'much confusion'.
If it is meaningful, removing 'and' from the bad_word file after
indexing causes all of the above queries to fail. Removing 'and'
and then reindexing causes all of the above queries to succeed.
By succeed, I mean both that matches are found and the full
query, including 'and' is highlighted in the excerpts.
Jim
-------------------------------------------------------
This sf.net email is sponsored by:
Access Your PC Securely with GoToMyPC. Try Free Now
https://www.gotomypc.com/s/OSND/DD
_______________________________________________
htdig-general mailing list <[EMAIL PROTECTED]>
To unsubscribe, send a message to <[EMAIL PROTECTED]> with a
subject of unsubscribe
FAQ: http://htdig.sourceforge.net/FAQ.html
