On 10/12/2011, at 9:44 AM, Benson Margulies wrote: > On the substance of the original problem: > > user runs 'mvn site:site' > > site:site finds javadoc and checkstyle, trigger a forked lifecycle to > generate sources. > > In this particular case, 'generate-sources' in module 'b' depends on > running out through 'process-classes' in module 'a'. b has a > compile-scope dependency on agroup:aartifact:aversion:wsdl.
The Javadoc plugin specifies this: @execute phase="generate-sources" So there isn't a way to convince it to do something else in the forked lifecycle. > > Is the lifecycle calculator supposed to figure this out? It has no way > to tell which execution in 'a' produces > agroup:aartifact:aversion:wsdl, so it has no way to tell that > 'process-classes' is going to be required. This is the more fundamental question that would need to be addressed - there's no way to express that in Maven right now. > > However, adding 'process-classes' to the top-level command line > doesn't help, apparently because it doesn't travel from the initial > execution plan to the forked execution plan. > > Once you have the fix I made today, you can get around this by running > 'mvn install' to stuff the wsdl artifact into the local repo before > running site:site. > > But if the intent is to avoid this requirement, how is it supposed to work? Perhaps an alternative would be to additionally bind that generation to the pre-site phase in 'a' as well. Not sure if that'll make it available, but worth a try. That would require running "mvn site" instead of "mvn site:site" (which is standard anyway). Otherwise, I think you're faced with documenting "mvn install" without doing major lifecycle surgery. - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
