Regression: _WidgetBase#placeAt() can't be used on a DocumentFragment

[Bryan Forbes]

var frag = document.createDocumentFragment();
widget.placeAt(frag);

This will fail because reference.tagName is checked in placeAt which causes refWidget to be set to the document fragment. In the else condition, ref is set to undefined because refWidget is truthy and both containerNode and domNode are undefined on the fragment. This should be checking for appendChild or nodeType rather than tagName. This is also a regression because this used to work in 1.7.

[Bryan Forbes]

This broke as part of the change in [28251].

[bill]

Hmm, placeAt() isn't documented to be able to work on a DocumentFragment. The API doc says it takes a Widget, DOMNode, or id of widget or DOMNode. I guess you are saying you want _WidgetBase.placeAt(docFrag) it to call through to domConstruct.place(), but domConstruct.place() is also not documented to work with a DocumentFragment.

This will fail because reference.tagName is checked in placeAt which causes refWidget to be set to the document fragment.

I'm not following you there. The code checking tagName is:

var refWidget = !reference.tagName && registry.byId(reference);

Regardless of the first clause in the expression, refWidget will be null (or undefined), because registry.byId(frag) will return nothing.

[Bryan Forbes]

		byId: function(/*String|Widget*/ id){
			// summary:
			//		Find a widget by it's id.
			//		If passed a widget then just returns the widget.
			return typeof id == "string" ? hash[id] : id;	// dijit/_WidgetBase
		},

Since id is not a string (it's a DocumentFragment), it is returned. The same would happen with a DOMNode if not for the tagName check.

[bill]

Funny how registry.byId() can return something other than a widget.

Anyway, I talked to Bryan about this over IM. If he adds DOH tests for dojo.place() and _WidgetBase.placeAt() to work with a DocumentFragment target (in addition to an Element target), and updates the API docs to say that a DocumentFragment is also allowed, I'm OK with the suggested change.

Apparently, a DocumentFragment target is supposed to work exactly the same as an unattached Element target: the position parameter can be "first", "last"/undefined, or a number, but not "before", "after", or "replace".

[Colin Snover]

Moving incomplete issues to next minor release.

[Colin Snover]

1.8.5 is released; moving all outstanding tickets to next milestone.

[Colin Snover]

1.8.6 is released; moving all outstanding tickets to next milestone.

[Bill Keese <bill@…>]

In 2f1094bcf75cdc0ff58e3ce76bbedb8e9228bd41/dojo:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled? 

[Bill Keese <bill@…>]

In 38230b8c473fa1be0410366d9454332c5f38d44e/dijit:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled?