Searching the archives is faster than typing such a long email :) Already fixed in CVS HEAD for Maven 1.1.
- Brett On Mon, 29 Nov 2004 20:32:59 +0100, Jesper Linvald <[EMAIL PROTECTED]> wrote: > Hi all, > > I think there might be a small bug in the way dependencies are currently > being resolved? > > SCENARIO: > > Two (or more) dependencies with same name but different extensions and types > are specified in the POM: > > <dependency> > <groupId>mydep</groupId> > <artifactId>mydep</artifactId> > <version>1.0</version> > <type>tld</type> > </dependency> > <dependency> > <groupId>mydep</groupId> > <artifactId>mydep</artifactId> > <version>1.0</version> > <type>jar</type> > </dependency> > > In the end only one of them (the tld dependency) gets properly resolved, > loaded and downloaded â no error message is conveyed in respect to failed or > erroneous dependencies! > The mydep-1.0.jar is not added as a dependency at all â to my best knowledge > it is completely ignored (maybe because there is a map impl. behind the > scenes?) > If the order of dependencies are reversed then the jar file is loaded and > copied! > > The documentation specifies that only jar, ejb and plugin are "known and > recognized" types. I have however seen extensive usage of other types such > as ears, wars etc. > > In my project we have further overloaded the type definitions to include > xmls and tlds because it is extremely useful â maybe the type of dependency > is not supposed to be arbitrary? I think it should! First class dependencies > should not be limited to only a specific set of types I think! The > repository is where I like to keep ALL dependencies and artefacts and a > dependency is a good and general way to specify a global path to get to it! > > In my way of Maven thinking these filetypes (tlds and xmls) are not simply > resources (for test or jar files as the documentation states) but full blown > dependencies without which I like my build to fail! > The issue might be more relevant in connection with deployment (where I > might like tlds to be deployed with the deploy:copy-deps goal) but I think > it is a general mistake if duplicates of artefact simple names are not > allowed if they otherwise differ on other parameters such as type/extension > or the likeâ. > > Does it make sense? > > ______________________________________________________________ > :)esper Linvald > > -- > No virus found in this outgoing message. > Checked by AVG Anti-Virus. > Version: 7.0.289 / Virus Database: 265.4.3 - Release Date: 26-11-2004 > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
