On Sat, 19 Apr 2003, Antoine Levy-Lambert <[EMAIL PROTECTED]> wrote: > * In the code of the War.java task, the lib attribute is defined as > a ZipFileSet. In the documentation, the same lib attribute is > defined as a FileSet. Should we fix the code or the documentation > of the War task ?
Which documentation? We don't tell people that <lib> is a zipfileset because they may get confused when they try to use the prefix attribute then. It must be a zipfileset, so that it can get mapped to prefix="/WEB-INF/lib". > > * the failure of the test was due to : > > - this code in ZipFileSet#getRef(Project p) which is good in situations where you really have to be sure that you are dealing with a zipfileset - and bad when you don't care. There is no good implementation that works in both cases, I'm afraid. > a) remove the ZipFileSet#getRef method , which means a fall back > on > AbstractFileSet#getRef (status-quo ante), which should work too. I think this is what Conor had to do for <classfileset> to work. Not sure whether it is the correct solution, though. But then again, you can always cast the returned object, when you really must be sure that you are dealing with - say - a zipfileset. > c) or write in WHATSNEW something like : - only references to > zipfileset(s) are accepted by tasks using zipfileset(s) as > attributes or nested elements. Is this really new? Stefan