Re: cvs commit: jakarta-gump/project jakarta-ant.xml jakarta-commons.xml jakarta-taglibs.xml uddi4j.xml xml-crimson.xml xml-xerces.xml
Christopher Lenz <[email protected]> Sun, 05 Jan 2003 23:23:47 +0100
| Newsgroups | gmane.comp.jakarta.alexandria.devel,gmane.comp.jakarta.gump |
|---|---|
| Message-ID | <[email protected]> |
Nicola Ken Barozzi wrote: > Sam Ruby wrote: >> For projects I care about (and that's a *lot* of projects), I want the >> descriptors to *either* be actively maintained or someplace that >> bodewig, conor, Stephen, myself, and several others (hi leosimons and >> mpoeschl and mvdb and cmlenz!) can update. > > This is not your requirement, it's a Gump need. It's quite evident to all. > > Your way of making Gump runs work is to have a deep knowledge aff all > Jakarta codebases and fix them yourself. > > My view is that the Gump descriptors are the base of the daily builds of > a project, so that users will be *forced* to maintain most of it, or > else the project won't compile. I agree. For Cactus, we're keeping a pretty damn good eye on the Gump results and tend to react to problems very quickly, because we use Gump for our nightly builds. I get cold shivers when I see a Cactus build soaked in a red background :-) The vast majority of projects doesn't really on Gump for anything, though, which gets very evident in the universe of Maven built projects, for example. I think keeping the Gump descriptor close to the project makes a lot of sense. But other situations like the recent mass-optionalizing of jaxp are also a concern. Perhaps, just following the guideline that projects should be able to maintain their gump descriptor themselves, as long as they really *care* about the results, is sufficient here. For now, I'll leave the Cactus descriptor in jakarta-cactus. -- Christopher Lenz /=/ cmlenz at gmx.de