More Java Woes
At the end of last week I was trying to run a release of a Maven2-based project. This involves the "mvn release:prepare" and "mvn release:perform" commands. While running the perform step, the project's javadoc creation was triggered and this hung the build entirely. The javadoc builder went out to lunch while using 100% of one of my CPUs. Lovely.
A colleague suggested switching from the Apple JVM to the soylatte JVM. I tried this and not only did the build not hang, but the build went noticeably faster. So I've switched to soylatte as my default JVM on my MacBook. I did a slightly more scientific speed comparison by recording how long it took to do a "mvn clean test" for a few of the projects I worked on and didn't see any significant difference between the two JVMs. So whatever difference I saw was either specific to javadoc or a result of disk cache, etc.