Hello Leonardo, I think we should fix it like my suggestion in 1.2.x. Your suggestion didn't work with the "data:1:command" expression. You can see it with my last commit on the trunk. Regards Bernd
On Fri, Oct 21, 2011 at 6:35 PM, Leonardo Uribe <lu4...@gmail.com> wrote: > Hi Bernd > > I have checked the problem against the algorithm implemented in 2.0.x > and I see the problem. The spec does not define if this method can be > overriden, so that details was ignored on MYFACES-3268, and all known > tests doesn't check that part. It seems we need to sync the algorithm > in 1.2.x with the code in 2.0.x, which was enhanced. > > What I see is the case that fails is: > > :data:1:command > > The base when a findComponent expression starts with ':' is > UIViewRoot, the algorithm found it, then it found data, but the code > does not delegate to data, instead start looking from data. > > This is the patch agains 2.0.x branch: > > https://issues.apache.org/jira/secure/attachment/12500202/MYFACES-3268-3-fix-delegation.patch > > Note the overriden method on the test case has a problem, because the > call for findComponent does not assume the id of the map should be > attached. > > I have seen other changes on 1.2.x, but I think we should copy the > code on 2.0.x/2.1.x to 1.2.x (changing the separator stuff). Do you > agree with this fix? > > regards, > > Leonardo Uribe > > 2011/10/21 Bernd Bohmann <bernd.bohm...@atanion.com>: >> Hello, >> >> just commited a Test that shows the wrong behavior of >> UIComponent.findComponent since 2.0.8 and 2.1.2. >> I would like to discuss this with leonardo. >> >> Sorry for any inconvenience. >> >> Regards >> >> Bernd >> >