Hi, Kundig Re: your e-mail from Tuesday, March 13, 2007 5:44 AM KA> Hello
KA> KA> I want to use ivy in several existing WSAD projects. KA> There are war projects, and ear projects which refer to the wars. KA> KA> A war project needs to know which jars he puts in the WEB-INF/lib directory for distribution, KA> which jars are distributed by the ear (the war must refer to them in the META-INF/MANIFEST.MF file) KA> and which jars he needs only for compilation (the jars the server provides to all ears). KA> KA> An ear project needs to know which jars he distributes. KA> KA> I am trying to keep all the information about jar dependencies in the ant/ivy system. KA> Is this possible ? KA> KA> thank you KA> KA> Andreas Kundig This is precisely type of a problem where Ivy shines. On our large JavaEE project we have many JavaEE modules / applications. We have standardized across the board to have the following Ivy configurations in our modules: - "compile". Determines compile-time classpath. - "war". Determines what JARs go into WEB-INF/lib. Also, the special case is a dependency on a WAR artifact, in which case, it is expanded and consumed (i.e. packaged) into an output WAR - "ear". Determines what JARs go into an EAR. Additional configurations that might be useful, could be "test" (in our case we simply use "compile" for test targets), and "ide" to pull sources and JavaDocs into IvyDE. And of course, you can tweak what configurations you decide to have to your heart's content. Dmitriy <1-127-441 @ICQ, DKroot @Skype, DKroot1 @AIM, dkroot1_at_gmail_dot_com @Google Talk>
