@scottb: Yeah the wording there is a bit awkward. I'll blame that on the description growing organically as the patch progressed rather than being written as a summary at the end.
http://gwt-code-reviews.appspot.com/1097801/diff/1/2 File dev/core/src/com/google/gwt/dev/shell/jetty/JDBCUnloader.java (right): http://gwt-code-reviews.appspot.com/1097801/diff/1/2#newcode33 dev/core/src/com/google/gwt/dev/shell/jetty/JDBCUnloader.java:33: public static void unload() { On 2010/11/10 21:17:27, jat wrote:
Could you avoid making the reflective call by just making this a
clinit instead? No, it needs to be loaded by the WebAppClassLoader. Without the reflective call, this would happen in the main app classloader which is not what we want. http://gwt-code-reviews.appspot.com/1097801/diff/1/3 File dev/core/src/com/google/gwt/dev/shell/jetty/JettyLauncher.java (right): http://gwt-code-reviews.appspot.com/1097801/diff/1/3#newcode616 dev/core/src/com/google/gwt/dev/shell/jetty/JettyLauncher.java:616: * in the Apache Tomcat project at On 2010/11/10 21:17:27, jat wrote:
You should describe the modifications, which in this case I presume
are all
about GWT's logging, right?
I stripped the logging, the configuration, and it's not a LifeCycleListener since that's a Tomcat thing--not a Jetty thing. Given that these are all pretty inconsequential I don't see the need to describe them in great detail. If someone cares they can easily pull up a diff. http://gwt-code-reviews.appspot.com/1097801/diff/1/3#newcode628 dev/core/src/com/google/gwt/dev/shell/jetty/JettyLauncher.java:628: private void jreLeakPrevention(TreeLogger logger) { On 2010/11/10 21:17:27, jat wrote:
Is there a reason this isn't static?
Nope. It probably should be. I'll fix that. http://gwt-code-reviews.appspot.com/1097801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
