The testcase was supplied by Chris Rompot in http://jira.codehaus.org/browse/MASSEMBLY-684
The test failed with sufficiently high frequency that I could "easily" track down the problem in a day or so. The bug is twofold; it regards overlapping calls to Method#setVisible between different threads, the singleton being the java.lang.Method object. The additional problem is that plexus-interpolation basically catches all these exceptions and puts these errors somewhere where most clients dont even bother to check. In .21 I only fixed the synchronization around setVisible, but I suppose we should also bring the exception into daylight. (We might simply just throw the exception instead....) Kristian 2014-10-18 20:26 GMT+02:00 Igor Fedorenko <i...@ifedorenko.com>: > Kristian, > > Where can I find the test case and more information about the problem > and the fix? Thank you in advance. > > -- > Regards, > Igor > > > On 2014-10-18, 14:22, Kristian Rosenvold wrote: >> >> Thanks to an excellent testcase provided by a JIRA user, a significant >> source of thread safety bugs has been located and fixed. The problem >> has an extensive reach, and can be the source of practically *any* >> issue in parallel builds, and is probably the root cause of most JIRA >> issues filed wrt parallel build problems. >> >> Fortunately, there is a very simple fix by putting the following >> inside your plugin configuration: >> >> <plugin> >> ... >> <dependencies> >> <dependency> >> <groupId>org.codehaus.plexus</groupId> >> <artifactId>plexus-interpolation</artifactId> >> <version>1.21</version> >> </dependency> >> </dependencies> >> </plugin> >> >> >> You should also put this jar file in the lib/ext folder of your maven >> installation: >> >> >> http://repo1.maven.org/maven2/org/codehaus/plexus/plexus-interpolation/1.21/plexus-interpolation-1.21.jar >> >> >> Which plugins are affected ? It's almost easier to tell which is /not/ >> affected. See the "relates" link on >> https://jira.codehaus.org/browse/PLXCOMP-258 to see which >> plugins/tools are affected. >> >> Kristian >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org >> For additional commands, e-mail: dev-h...@maven.apache.org >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
