Le 24 août 2010 à 00:47, Carl Myers a écrit : > The "common and reusable" pattern is "I have a mess of files". Why does it > need to be more specific? > > Some examples: > > 1. Our custom build system consists of several xml and properties files. We > would like to version our build system itself so that just a small ant script > is used to pull the build system, then it is used to perform the rest of the > build. We would like to have the build system cached, then extracted to a > certain location, after an ivy resolve, but this must work within the eclipse > workflow and the CLI workflow (but, I suppose, either way it will be the dev. > running an ant script probably, so maybe IvyDE doesn't need to support it > explicitly)
quite off topic, but this is amazingly exactly what I want to put in place for my projects. I think Easyant can help with that, but it is not yet there, so my recent long mail on easyant dev mailing list. > > 2. Sql files, or other generated artifacts that are not Jars. I have a > directory full of Sql files that several packages need to depend upon. What I > really want to do is "build" these sql files (run tests, validate them), then > publish them using ivy. But when other things depend on them, they need the > files in a certain location, not on their classpath. I want to be able to > use a construct where I say "resolve this package and place its contents > here". > > 3. Configuration only packages. This makes sense once your codebase gets big > enough - and similar to #2 above, you want to have a bunch of xml or > properties files or whatever and you want them to end up in a location on > disk, not in the ether. > > I think this is a very general (and useful) case for Ivy to solve. So you need IvyDE to launch a resolve of some ivy.xml file, and then retrieve the appropriate files in some defined place. IvyDE can already do that: http://ant.apache.org/ivy/ivyde/history/latest-milestone/cpc/retrieve.html As I wrote, it is tied to the Java nature of a project, and an IvyDE classpath container has to exist even if you don't use it. It should be useable though. But for IvyDE 2.2, I have just implemented a retrieve mechanism for non java project. So you would be able to right click on a project, select the Ivy menu and then select "Retrieve 'sql files'". And with Clint Burghduff patch I integrated too, the files get properly refreshed in Eclipse. I wouldn't consider the 2.2 as production ready, but an early feedback on that would be appreciated. Nicolas > > -Carl > > On 08/19/2010 01:00 PM, Nicolas Lalevée wrote: >> pe, Ivy doesn't care. I have been able to make Ivy manage dependencies >> between flex projects. IvyDE on his side is mainly intended to be used in a >> Java projects (probably too tied to Java, > > -- > Carl Myers > Palantir Technologies | Internal Tools Software Engineer > [email protected] > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
