So it's basically arguing that server-side code benefits from logging to stdout rather than a named file? Interesting idea if you aren't using stdout for something else, but it doesn't seem particularly radical since you can just use cat.
A few months ago I published a suite of scripts that allow me to run complex queries using shell pipes. They work by communicating in YAML. The first stage, cat_logs, simply parses logs from a file and emits yaml for each line. cat_logs understands that log files are split across multiple files, and that they may be gzipped.
A few months ago I published a suite of scripts that allow me to run complex queries using shell pipes. They work by communicating in YAML. The first stage, cat_logs, simply parses logs from a file and emits yaml for each line. cat_logs understands that log files are split across multiple files, and that they may be gzipped.
http://akkartik.name/blog/2010-09-25-05-29-01-soc