I'm trying to get htdig working on a Solaris 2.7 system with the Sun C
compiler. Here's the notes I kept as to what I did to get where I am.
I can now go to my local search.html Web page, and enter a word, but
no matter what word I try, I always get no hits. Any suggestions are
welcome.
1.35.1 Get htdig
To get it, web to http://www.interweft.com.au/other/htdig/where.html
and click on the htdig-3.0.6.tar.gz link.
gunzip htdig-3.0.6.tar.gz
tar xf htdig-3.0.6.tar
rm htdig-3.0.6.tar
cd /usr/web
mkdir htdoc
cd htdoc
cp /usr/src/htdig-3.0.6/htdoc/* .
cd ..
chown -R siemsen htdoc
chgrp -R datacomm htdoc
1.35.2 Build htdig
cd /usr/src/htdig-3.0.6
Use a web browser to look at /usr/web/htdoc. Under Installation,
you'll find a page that describes editing of the CONFIG file. I set
values in this file as follows:
DEST= /opt/htdig
CGIBIN_DIR= /usr/web/nets/cgi
IMAGE_DIR= /usr/web/nets/images
IMAGE_URL_PREFIX= /nets/images
SEARCH_DIR= /usr/web/nets/Forms
Now configure it:
./configure
The next step is to run "make". If you do, you'll get compiler
warnings, and then a compile failure. To get past the failure, I
hacked the source files. Scary, but I didn't know what to do. I
tried copying the distribution to another system that has GCC and
doing a "configure" and "make" there. On that system, I got even more
warnings, and then it failed on the same source file. So...
Edit htfuzzy/Exact.c and add
return 0;
as the body of the openIndex function.
Edit htdig/Document.cc and change
typedef void (*SIGNAL_HANDLER) (...);
to
typedef void (*SIGNAL_HANDLER) (int);
Edit htdig/SGMLEntities.cc and change
return (unsigned char) (*junk.trans)[entity];
to
return (int) (*junk.trans)[entity];
Edit htdig/Makefile and change
$(TARGET): $(OBJS)
$(CXX) -o $(TARGET) $(LDFLAGS) $(OBJS) $(LIBS)
to
$(TARGET): $(OBJS)
$(CXX) -o $(TARGET) $(LDFLAGS) $(OBJS) $(LIBS) -lsocket -lnsl -lxnet
1.35.3 Build and install htdig
cd /usr/src/htdig-3.0.6
make
make install
chown -R siemsen db
chgrp -R datacomm db
1.35.5 Configure htdig
A file named /opt/htdig/conf/htdig.conf controls the operation of the
indexer. Edit it to reflect our web:
start_url: http://netserver.ucar.edu/
exclude_urls: /nets/cgi/ .cgi
1.35.6 Make a Web page to test searching
The make install copied an HTML file named search.html into
/usr/web/nets/Forms. The file is screwed up ina few ways. For one, it
contains these lines:
<input type=hidden name=config value=$(CONFIG)>
<input type=hidden name=restrict value="$(RESTRICT)">
<input type=hidden name=exclude value="$(EXCLUDE)">
...which seem wrong. When the Web server starts htsearch, htsearch
produces the error
Unable to read configuration file "/opt/htdig/conf/$(CONFIG).conf"
After tearing my hair for a while, I discovered that there was a
version of search.html in .../installdir/search.html.orig that seemed
more correct. Instead of the above three lines, is had just
<input type=hidden name=config value=htdig>
So:
cp /usr/src/htdig-3.0.6/installdir/search.html.orig /usr/web/nets/Forms
Edit /usr/web/nets/Forms/search.html.
Change the header sentence to
This page allows you to search the contents of the NETS Web pages.
Change
<form method="post" action="/cgi-bin/htsearch">
to
<form method="post" action="/nets/cgi/htsearch">
1.35.7 Build the htdig index
/opt/htdig/bin/rundig -v
1.35.8 Test a search
Start a web browser, and go to this URL:
http://netserver.ucar.edu/nets/Forms/search.html
This "works", except that it always says there are no hits.
Please help. HTDIG looks like a neat tool, and I think I'm very close
to making it work.
-- Pete
----------------------------------------------------------------------
To unsubscribe from the htdig mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the body of the message.
