----- Original Message -----
Sent: Thursday, July 05, 2001 4:28
PM
Subject: RE: [htdig] laola2html.pl
Believe me, I'm getting ready to spend a bit, but for
some political reasons on this project, it would probably be my own pocket
money!
As
for catdoc, I simply haven't been able to use it. The current and
previous version wouldn't compile for me (on Cygwin on NT) and the binary I
tried just emits very unhelpful errors.
What's frustrating is that LAOLA works like a charm
using laola2html or doc2html on the command line. It is only during a
dig, inconsistently, that I get an excerpt like I showed before (always the
same exact odd characters, BTW). The first, small test dig I did
appeared to be fine (which is why I posted the script and reported seeming
success, silly me) but when I went to dig our whole intranet site this thing
started cropping up.
If I
take one of the exact Word docs that had the weird excerpt and process it on
the command line with laola2html.pl or doc2html.pl (calling the other), it
spits out a perfectly fine web page (that should be indexed perfectly
well if returned to htdig during the dig). Weird.
I'll
probably end up buying wp2html and using it with your standard scripts; all I
can say is this must be one messed up file format if it is that difficult to
dump text out of it!
I have no experience with laola, or other Perl software
for MS files. I use a combination of wp2html and catdoc, and though I
can't recall ever seeing anything like ��ࡱ�, I could well believe that
catdoc could produce such output. However, I've no explanation for
what you are seeing. Htdig launches an external parser and waits for
it to complete, as you know from your expereince with word2x
:)
If your Word documents are Word97 and later you might find
the best solution is to spend a few pounds on wp2html.
I don't recall you every giving a reason for rejecting
catdoc?
--
David Adams
Computing Services
Southampton
University
