On 01/15/2013 07:17 PM, Dennis Lundberg wrote: > On 2013-01-12 15:03, Thomas Neidhart wrote: >> Hi, > > Hi Thomas > > A while back I made changes to the Maven build so that it produces the > same output as the Ant build. The should mean that we can get rid of the > old Ant build if we want to.
I think the ant build script it is used by gump atm. > One thing that I'd like to do is to restructure the source code into > several separate Maven modules, so that there is a specific module for > commons-logging-api. The current setup is error prone, as it is > extracting certain files from certain archives and then repackaging them > again, with the risk of loosing meta data like MANIFEST.MF. Yes I like that idea, the current packaging and testing is quite odd, e.g. one has to run mvn integration-test to do the actual unit tests. btw. does anybody know how to automatically do a verify phase when doing a mvn site, otherwise the results of the integration-tests are not published. >> I would like to do a similar cleanup as for email also for logging and >> aim for a 1.2 release in the coming weeks. The things I have in mind: >> >> * update to Java 5 > > I'm -1 on this change. I don't see any reason to do it. We don't need > features from a more recent Java version in commons-logging. As others > have said: most users of commons-logging are old and older apps. In general I am fine with keeping the current JDK compatibility, but there are also a few points in favor of such a change: * how can we ensure compatibility with such outdated JDKs today? minimum jdk I have installed is JDK 1.5 * the codebase is at parts quite complex as it has to deal with such a wide range of supported JDKs. Reducing the number for (future) releases would simplify the maintenance (and we could remove some old code for pre-1.4 JDKs). There are still lots of other (non-legacy) projects that use commons-logging, e.g. spring. But maybe we could also keep the 1.x branch on java 1.1, and create a 2.x branch which targets java 5. >> * comply to default maven structure > > +1 already did so. >> * update to Junit 4 > > +1 still open. >> * fix the open issues wrt thread safety Sebb created some issues about this topic, and there is one (reproducible) deadlock scenario with the WeakHashtable. Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
