Hello, I've long heard "that Gump can't *yet* run Maven", and the emphasis is always on the yet. I'm assisting with a re-work of Gump (into Python) and I'm curious whether making Gump run Maven is practical. If it is, and folks here are game to try, maybe I can look into coding the Gump side.
I've seen other folks try having the build tool [and all it's components] be built by Gump, and then the build tool used to build other Gump tasks. Clearly it works for ant, however ant is pretty stable these days and (at core) has few dependencies. That said, even ant has a bootstrap issue. Could Maven be built from ant builds of it's sub-components? Would it be a stable stack? The problem comes when the build tool is sufficiently dependent on a stack of sub-components, that any instabilities in that stack cause the build tool not to build, and hence cause the dependee projects not to get Gumped. This means that dependee projects loose valuable chances at spotting a glitch in their runtime dependencies. This could be a big negative, and maybe even enough reason to keep things as they are. [Gump builds/test don't have to be full distribution builds, just enough to compile/test interfaces, so maybe don't need fancier features Maven has, not sure...] Also, I assume that Maven uses the POM in order to manage the classpath for builds. Ant allows the environment to take control, overriding the ant script. [See: http://ant.apache.org/manual/sysclasspath.html] Does Maven have any such concept? Anybody here able to shed further light on this matter? What would you guys like to see done, and how? regards, Adam -- Experience Sybase Technology... http://www.try.sybase.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
