> One thing that I have is constraints. It's a way to describe under what > circumstances a tag makes sense. These constraints are currently based > on whether the class is instance of a certain class/interface, and > whether or not an other tag is present. This is necessary if we want to > provide proper validation.
IMHO it's also good to provide a validator="com.foo.MyValidator" attribute or even a <validator/> element. There can be a default validator which uses hierarchy/etc info by default as you mentioned above for we can also write some smart validators too this way for other more complicated cases (if stateless session and having method x and bla bla so...). > I also saw from you source code that you're using println() to generate > sources. Why not use XDoclet's template engine? You don't need java > classes to use the template engine you know. Take a look at my recently > committed xdoclet.ant package and the samples' appxml target. I think > you could use the same approach, but extract the info from TagMetaData > instead of a Hashtable like I did. I think he's more comfortable with Velocity template engine. We may later use Velocity but for now please use xdoclet's template engine. The template engine is not important much for our current plans (gui/xtags/xjavadoc/etc). Having a template engine of our own gives some flexibility, for example take a look at what Vincent has done to handle merge file timestamps, I have no idea how we can implement it with Velocity! Ara. _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com _______________________________________________ Xdoclet-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/xdoclet-devel
