Hi Sven, The problem is in 00279f1d8df057bbf434f3ea1eb47c7ef7de17d2. I'm not sure what I thought to do it this way.
I have used https://github.com/wicketstuff/core/tree/master/jdk-1.6-parent/stateless-parent/stateless-examples to debug the problem and to fix it. And since the stateless ajax behaviors there return 'true' in #getStatelessHint() I did this wrong fix. The real fix should be in ajax stateless behaviors - they should register the needed id themselves. On Tue, Oct 29, 2013 at 5:43 PM, Sven Meier <[email protected]> wrote: > BTW I stumbled over this, because we have stress tests clicking behaviors, > which were never enabled. > Since an id is never requested for the behavior, our tests fail with > 'Cannot find behavior with id..." instead of "behavior not enabled ...". > > Sven > > > On 10/29/2013 04:34 PM, Sven Meier wrote: > >> Hi, >> >> I don't understand Behaviors#internalAdd(**Behavior): >> >> private void internalAdd(final Behavior behavior) >> { >> component.data_add(behavior); >> if (behavior.getStatelessHint(**component)) >> { >> getBehaviorId(behavior); >> } >> } >> >> Why does it register the behavior for an id if its stateless? Shouldn't >> that be done in the opposite? Why isn't it done lazily when needed only? >> >> Note that the condition has flipped back and forth a few times in the >> commit history. Now each simple AttributeModifier is put into the >> BehaviorIdList, although the id is never used for it. >> >> @marting + igor: Any thoughts? >> >> Sven >> > >
