Operation of the program

The script runs as a cron job, at some reasonable time after the log files are rotated (say, an hour).

The script should be run on infohost for performance. The line to start it should look like this:

    nice +5 zcat access_log.1.gz| \
      /u/www/docs/tcc/projects/weblog/run

The program reads the new log file from its standard input, merges it with its own summary database at a known location, rewrites the summary database, and builds the output files (q.v.).