Opened 9 years ago
Closed 9 years ago
#11618 closed defect (wontfix)
dijit._CssStateMixin should check for existence of nodes before tracking mouse state
| Reported by: | jasonwain | Owned by: | anonymous |
|---|---|---|---|
| Priority: | high | Milestone: | tbd |
| Component: | Dijit | Version: | 1.5 |
| Keywords: | Cc: | ||
| Blocked By: | Blocking: |
Description
if you override the template for a class that mixes in the _CssStateMixin and don't include one of the nodes that is declared in cssStateNodes, _CssStateMixin will still try to register event listeners for the undefined node (this[ap] on line 55).
its an easy workaround to also override cssStateNodes, but _CssStateMixin should probably have that check either way.
fyi, specifically, this occurred when extending dijit.Dialog and using a template without a closeButtonNode - after showing a dialog, there are dojo errors inside dojo.removeClass any time you click anywhere on the page.
Change History (1)
comment:1 Changed 9 years ago by
| Component: | General → Dijit |
|---|---|
| Resolution: | → wontfix |
| Status: | new → closed |
Note: See
TracTickets for help on using
tickets.
OTOH that would mask errors people make in their cssStateNodes array / templates.
In 2.0 most/all of the CssStateMixin code will go away, specifically the parts monitoring hover and focus, as that will be handled via CSS pseudo-selectors :focus and :hover (and IE6/7 will simply degrade gracefully).
Until then, I'd rather not change CssStateMixin, although I do see your point.