On Thu, 08 Mar 2001 13:15:25 +1100, Peter Donald wrote: >At 05:45 7/3/01 -0800, David Rees wrote: ... > >>I think we need to look at better understanding the fact >>that the attribute/element is very blurred in XML and that Ant blurs >>it much more. We need to support a common way for indicating either >>will work in a Task. Includes being good example. Do we even really >>need a element for that, or could we support >>setIncludes(String aString) > >Well I agree though I would do it like > >setInclude( IncludeEntry[] entrys ); > >because include will most likely have multiple attributes. We could still >support > > > >my implementing converters that convert from string to IncludeEntry type. >(A feature that is much requested in certain domains like web-site >development). > >so > > > > > > > >would be identical (assuming myinclude refers to an include datatype). They >would all end up doing roughly > >setInclude( new IncludeEntry[] { new IncludeEntry( "**/*.java" ) } ); > I definitely agree with this, but I am thinking that we need be more generic for set vs. single. Something like: setInclude(AnObject obj); setIncludeSet(AntSet aSetOfAnObjs) We shouldn't need to create AntSets for everything we support more than one of. And I still like the idea of a simple task supporting setInclude(AnObject obj) and Ant (or some batch task :)?) iterating through the values. Then is someone writes a reverse task to reverse the contents of a file he can just write code that supports setFile(File aFile) and the code (or the developer) doesn't have to know anything about AntSets. And yet the following would all work: Of course, the last two should be supported anyway by inheritance (IMHO). d --------------------------------------------------------------------- To unsubscribe, e-mail: ant-dev-unsubscribe@jakarta.apache.org For additional commands, e-mail: ant-dev-help@jakarta.apache.org