Am Donnerstag, 18. August 2005 16:41 schrieb Sean Schofield: > > Hi, > > > > to summarise: Up to now all problems were related to the migration from > > myfaces 1.0.9 to the nightly build from 20050812. It would have saved me > > a lot of time, if the necessary changes would have been documented > > somewhere > > What changes were necessary? Removing myfaces.jar? If that's what > you mean, then yes that should be documented. Your tree2 code should > not have to change though.
[...] Well, it took me a day to get things running. Maybe it was all my fault -- yet some documentation might have saved me a lot of time. (1) The first point was to clean up my project from myfaces.jar which I overlooked way too long. (I removed the library from the project right away but that was not enough. I am using WTP with eclipse, and I also had to remove the lib from the .deployables subdirectory of the project by hand. Well, this is not really an issue of myfaces, but a hint could save other people a lot of time too. Maybe it is a bug in WTP...) (2) The TreeBacker has changed. It has four new methods. Without documentation I don't know which ones are really necessary. As far as I understand the (one) comment in the source of the example, one has to define only one of the two methods public TreeNode getTreeData() and public TreeModel getExpandedTreeData() But public void setTree(HtmlTree tree) public HtmlTree getTree() public String expandAll() all have to be present? I didn't check it systematically afterwards, but I think the old treeBacker -- which is missing the last four methods -- does no longer work. Without documentation and the sources of the examples it takes some time before one browses the CVS online to find out... Which of these methods are necessary? (3) Maybe my example was wrong before and it didn't matter, but I think that the meaning of the last parameter in the constructor of the TreeNodeBase has changed: leafs are 'true' and nodes are 'false' now. My old example worked the other way arround before -- for whatever reason. Regards Andreas Schenk
