[Microdata] itemtype attribute should be space-separated list to allow multiple types

[Arko Saha]

Incorporating W3C the spec changes:

1. Microdata itemtype attribute is now a space-separated list of types rather than a single type, so that it can allow multiple types if they share the same vocabulary.
This is added in the spec in the revision number : http://html5.org/r/6668.
Spec: http://www.whatwg.org/specs/web-apps/current-work/#attr-itemtype

2. getItems() should return all the elements in the document that are each top-level microdata items whose types include all the types specified in the method's argument.
This is added in the spec in the revision number : http://html5.org/r/6680.
Spec: http://www.whatwg.org/specs/web-apps/current-work/#dom-document-getitems

[Arko Saha]

Created attachment 111776 [details]
Patch

[Ryosuke Niwa]

Comment on attachment 111776 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=111776&action=review

> LayoutTests/fast/dom/MicroData/006-expected.txt:5
>  PASS document.getItems(' http://example.com/foo http://example.com/bar ').length == 2 is true
> -PASS document.getItems(' http://example.com/foo data:text/plain ').length == 2 is true
> -PASS document.getItems(' http://example.com/foo data:text/plain http://example.com/foo').length == 2 is true
> +PASS document.getItems('  http://example.com/foo  http://example.com/bar  http://example.com/foo').length == 2 is true

Why the change?

> LayoutTests/fast/dom/MicroData/006.html:11
> -<div itemscope itemtype="http://example.com/foo"></div>
> -<div itemscope itemtype="http://example.com/bar">
> -<div itemscope itemtype="data:text/plain"></div>
> +<div itemscope itemtype="http://example.com/foo http://example.com/bar"></div>
> +<div itemscope itemtype="http://example.com/bar http://example.com/foo">

Ditto.

> LayoutTests/fast/dom/MicroData/006.html:16
> -shouldBeTrue("document.getItems(' http://example.com/foo  data:text/plain  ').length == 2");
> -shouldBeTrue("document.getItems('  http://example.com/foo  data:text/plain  http://example.com/foo').length == 2");
> +shouldBeTrue("document.getItems('  http://example.com/foo  http://example.com/bar  http://example.com/foo').length == 2");

Ditto.

> LayoutTests/fast/dom/MicroData/itemtype-add-remove-tokens.html:35
> +debug("<br>itemType.remove must not make any changes if a non-existing token is removed.");

"non-existing token is removed" sounds odd. How about "remove is called for a token that doesn't exist"?

[Arko Saha]

(In reply to comment #2)
> (From update of attachment 111776 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=111776&action=review
> 

> > LayoutTests/fast/dom/MicroData/006.html:11
> > -<div itemscope itemtype="http://example.com/foo"></div>
> > -<div itemscope itemtype="http://example.com/bar">
> > -<div itemscope itemtype="data:text/plain"></div>
> > +<div itemscope itemtype="http://example.com/foo http://example.com/bar"></div>
> > +<div itemscope itemtype="http://example.com/bar http://example.com/foo">
> 
> Why the change?

Accroding to the latest spec getItems() should return all the elements that are each top-level microdata items whose types include "all the types specified" in the method's argument.
Earlier the spec specified that getItems() should return all the elements that are each top-level microdata items whose types include "any one of the types specified" in the method's argument.

So if we do not modify the below test-case,

-<div itemscope itemtype="http://example.com/foo"></div>
-<div itemscope itemtype="http://example.com/bar">
+<div itemscope itemtype="http://example.com/foo http://example.com/bar"></div>
+<div itemscope itemtype="http://example.com/bar http://example.com/foo">

document.getItems(' http://example.com/foo http://example.com/bar ').length == 2 will always be false. Hence I made this changes.
And I have modified this test-case more perfect in my new patch.

> 
> > LayoutTests/fast/dom/MicroData/itemtype-add-remove-tokens.html:35
> > +debug("<br>itemType.remove must not make any changes if a non-existing token is removed.");
> 
> "non-existing token is removed" sounds odd. How about "remove is called for a token that doesn't exist"?
Done!

[Arko Saha]

Created attachment 112305 [details]
Updated patch

[WebKit Review Bot]

Comment on attachment 112305 [details]
Updated patch

Clearing flags on attachment: 112305

Committed r98353: <http://trac.webkit.org/changeset/98353>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.