Couple of parsing failures.

[Derek Pomery <pomerydc@thc-tsg.com>]

The following were with the latest release.
---------------------
<a 
HREF="http://oursite.reallyreallylongdomainname.net:80/project/contents/
process/design/UserInterface/prototype/somescreen.jsp:%20Jan%2024%202002
:%20221:%20%20%20%20%20<TD%20class='row-hilite-0'><input%20type="radio%2
0name="test"%20CHECKED></td>">http://oursite.reallyreallylongdomainname.
net:80/project/contents/process/design/UserInterface/prototype/somescree
n.jsp:%20Jan%2024%202002:%20221:%20%20%20%20%20<td%20class='row-hilite-0
'><input%20type="radio"%20name="test"%20CHECKED></td></a></b>, <br>

The following malformed url showed up in one of my test searches.  Had 
jump-to-line turned on, but it looks like a failure in parsing the 
glimpse results.
Still trying to figure out what went wrong, any advice welcome.
---------------------
More carefully examined is a possible difference in date string on my 
server.
The RankHits.pm module was incorrectly attempting to parse $date
which was the string "Fri, 25 Jan 2002 10:34:33" in this case.
I modified the string to match what I have on my machine, but obviously a 
smarter regex will be needed if date strings vary (or can a timestamp in 
seconds be grabbed from somewhere?)
---------------------
Finally, I had to take the substr() out again, it was still set to 10000 
characters, but I was once more losing the ends of filenames (an effect 
that disappeared when I removed the substr). Will try to identify what is 
going on in more detail.