Please don't let them be as boring as Brian's friends

I made an emergency release of the chronicle blog compiler yesterday, after noticing that it was truncating titles containing periods.

That was a bit of a mea-culpea moment, but I guess mistakes happen.

The new release is in perfect shape for Lenny, and now includes two new scripts installed into the examples/ directory:

• The RSS-feed importer I mentioned prviously.
• A quick hack to detect duplicate post/entry titles.

The latter was applied to my own blog, and I discovered several duplicates. I guess my film quotes having only a limited source collection to work from could also include duplicates - so I've updated my Makefile to only build and rysnc my blog if there are none detected.

(In many ways that films site is the precursor to this blog; it uses a collection of text files, one per film, and generates a cross-linked HTML output of film entries. Sadly it is out of date, because entering titles is a real pain..) Chronicle Comments

I'm pleased with the comment process now though, the CGI comment submission script simply archives each submitted comment into a "comments/" directory on the webserver.

There a cron-job passes each one through a bayasian filter and moves the file(s) to either "comments/good/", "comments/bad/" or "comments/unsure/".

When I come to rebuild the blog I rsync the "comments/good" directory to my local machine, rebuild and then rsync the output back to my remote webserver.

(On a single machine this would be much simpler process!)

I've imported my blog source into a mercurial repository, so the client-side is consistent. I have a bad habit of making new postings from wherever I happen to be and having a central repository will make that less prone to diaster.

Just running "make steve" against the Makefile is sufficient to rebuild everything and sync it to my live system.

ObQuote: Kalifornia

Syndicated 2008-05-03 14:41:36 from Steve Kemp's Blog