Today's Work
I got truss to work and found estsearch was hanging when it went to set a lock. Otto determined that this was happening because the lock daemon was not running on happy. After he started the daemon, estsearch no longer hung. Instead, it now takes an error exit and says the index is invalid. This appears to be happening during a call to llseek.
I put that problem aside for the moment and started incorporating the new branding in the site. So far I have a new template file. Using the template on the mailing list is as simple as destroying the list and sending a new message to create it again. Before I do that, I want to clean up the template by fixing the indentation, etc. This also helps me understand the html used.
Posted on Thu, 26 Feb 2004 Tags: html
Today's Work
I still can't get this search to run without hanging. Right now I'm trying to run the script through truss, so far without success.
Posted on Wed, 25 Feb 2004 Tags:
Today's Work
I set up happy with my modifications and started testing I ran into a few problems with the glue script, which as it turned, out, were caused by pointing to the wrong directory. I fixed them and then modified the search form on the mhonarc test page. I then started testing the search and ran into a problem. The search hangs and doesn't return. I copied the directory tree over to my machine and it works fine there. So I'm still trying to figure out why things would hang.
Posted on Tue, 24 Feb 2004 Tags: mhonarc
Today's Work
The first thing I did today was review my open calls and close the ones that could be closed. Then I looked at the source code changes I made Friday on the "glue" script. I made some changes and then created the template files necessary for estaier to run. I tested the glue script, and fixed the bugs I found.
Posted on Mon, 23 Feb 2004 Tags:
Today's Work
I dotted my i's and crossed my t's on the estraier installation. I moved the executable over to happy and ran it by hand to make sure it worked. I modified the glue script to run estraier, though I have not yet tested the changes. Before I do that I have to set up the template and configuration files.
Posted on Fri, 20 Feb 2004 Tags:
Today's Work
I worked on the phonebook problem some more but was unable to solve it. So I moved to installing extraier on our production systems. The standard way to do this seems to be to put the software on workmate and aldur. I needed to install zlib on workmate first, which I did. I was also unable to get estserver to compile, but since I don't need it, that's OK. I compiled estrier on aldur, but I was unable to install it as the root password seems to be different there.
Posted on Thu, 19 Feb 2004 Tags: phonebook
Today's Work
I tested estraier on my web site and it seems to work fine. I started looking at a problem with the phonebook lookup on our development system. It seems to be a problem with database corruption.
Posted on Wed, 18 Feb 2004 Tags: phonebook web
Today's Work
I tried to modify our LDAP schema this morning to add the field Otto had requested. The test LDAP server refused to take the change. (It errored out.) This caused some consternation for me and Harry, but the problem has not yet been fixed. In the afternoon I tried to install Plucene but ran ino the problem that our production servers are still running Perl 5.005. This is something that should be fixed as 5.005 is badly out of date. So I went hunting for another free search engine and found estraier. I downloaded it and built it on my machine so that I can test it.
Posted on Tue, 17 Feb 2004 Tags: ldap perl
Today's Work
I debugged missing.pl and fixed the problem causing it to report fields as missing when they weren't. I checked the other ldap tasks for the same problem and fixed leave.pl. I downloaded and installed Jxplorer in an attempt to modify ldap for otto, but wasn't able to figure out how to add a new field with the tool.
Posted on Mon, 16 Feb 2004 Tags: ldap
Today's Work
I got retheme to work, which was a matter of of using the right templates on the right files. I moved onto doing the search and I decided to use Plucene. Plucene is a Perl port of Lucene, which is a well regarded Java search engine. Its advantage is that it's pure Perl and thus easy to install and interface with Mhonarc. I reviewed the documentation in preparation for installing and using it.
Posted on Fri, 13 Feb 2004 Tags: mhonarc java perl