[MKSearch-dev] Week 8 round up
Phil Shaw
phil at mkdoc.com
Fri Dec 3 20:30:19 GMT 2004
Completed the first significant integration this week by completing
the JTidy plugin for JSpider. Now you can point JSpider at a Web site
and make a tidied mirror copy on the local file system.
That's a pretty useful thing itself, but the next step is to extract
the metadata...
Best regards,
Phil
Thursday
--------
Bruno pointed me to Swoogle and this led me to some RDF crawler
applications, which I added to the research area of the Web site.
None suitable for our immediate needs. I also found some notes on RDF
and database storage which I have bookmarked for later.
http://swoogle.umbc.edu/
Picked up a tip on the GCJ mailing list and simplified the source
file input for JSpider to pipe from find.
Brought the Ant build file up to equivalence with the bash scripts
for compiling and archiving the library classes. Found another GCJ
bug in the process and reported number 18789.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18789
Got the first draft JTidy plugin for JSpider working with some hacks
configuration. There are flaws in the Diskwriter plugin it was based
on.
Friday
------
Built up a full set of JUnit test cases around the JTidy plugin to
isolate and correct the configuration problems. As I got a clearer
understanding of what the original code was doing, I also refactored
some of the original methods to correct and enable complete testing.
Created a set of JSpider mock objects to test with and a Hansel
coverage test to ensure all branches are covered. All tested and
complete!
More information about the MKSearch-dev
mailing list