That's an interesting point. So you would be in favor of leaving it alone? If so, I still think we could change TreeNodeBase so that it could check its children count to get the answer. You could always override this method in your own TreeNode implementation.
What do you think about that? sean On 4/27/05, Sean Radford <[EMAIL PROTECTED]> wrote: > Jon Travis wrote: > > > Seems like in the very dynamic case, where you are only > > generating part of the tree at a time, that isLeaf might > > be useful. You might not know how many children you > > have without doing an expensive operation, but you might > > know that you are not a leaf. > > > And I will be doing it as such in a new app in a few weeks time... > > > isLeaf() isn't even used. Making it the exclusive way > > to determine if there is an expansion button beside it > > seems like a better API. > > > > -- Jon > > > > > > On Apr 27, 2005, at 9:04 AM, Sean Schofield wrote: > > > >> Sureshbabu pointed out something interesting in an earlier post. The > >> TreeBacker class has the boolean values for leaf reversed. So all of > >> the branch nodes have setLeaf(true) and visa versa. > >> > >> I will fix this in CVS now, but there is another interesting point > >> here. It doesn't matter to any of the simple examples! This is > >> because the getType method is used to determine the facet (and hence > >> what type of icon to use). The navigation icons (the plus/minus > >> icons) are determined by the child count. > >> > >> IMO the getChildCount method and the getType methods could cover > >> pretty much every scenario that you would need isLeaf for. I have > >> opened a JIRA issue (MYFACES-215) to suggest we deprecate the is Leaf > >> method (and eventually eliminate it.) Please add comments there if > >> you have feelings on this one way or another. > >> > >> sean > >> > >> > > > > > > > > -- > Dr. Sean Radford, MBBS, MSc > [EMAIL PROTECTED] > http://bladesys.demon.co.uk/ > >
