Handle (numeric) indexed access of DOMStringMap.

Handle (numeric) indexed access of DOMStringMap.

As the set of supported property names over DOMStringMap reflects the
attributes of an Element prefixed with 'data-', these naturally also
include numeric indexes.

To allow such index values to be used over DOMStringMap, extend its
IDL definition with indexed accessors.

Also address an infelicity in comparing attribute names to property
names if the property name uses a "-" but isn't followed by an ASCII
uppercase letter (e.g. "a-1".)

Extended test coverage to check for these variations also.

R=tkent@chromium.org
BUG=171175
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=161864

[sof, 2013-11-06 21:34:40]

Please take a look

[tkent, 2013-11-06 23:18:34]

This CL addresses number-only names cases (data-123), and not names including
numbers cases (data-foo1), right?
https://code.google.com/p/chromium/issues/detail?id=171175#c2

[sof, 2013-11-07 06:48:56]

On 2013/11/06 23:18:34, tkent wrote:
> This CL addresses number-only names cases (data-123), and not names including
> numbers cases (data-foo1), right?
> https://code.google.com/p/chromium/issues/detail?id=171175#c2

Great question; "data-foo1" (as an attribute) is de/encoded fine (or
"data-1foo"), but "data-foo-1" isn't.

Will address.

[sof, 2013-11-12 11:36:47]

On 2013/11/07 06:48:56, sof wrote:
> On 2013/11/06 23:18:34, tkent wrote:
> > This CL addresses number-only names cases (data-123), and not names
including
> > numbers cases (data-foo1), right?
> > https://code.google.com/p/chromium/issues/detail?id=171175#c2
> 
> Great question; "data-foo1" (as an attribute) is de/encoded fine (or
> "data-1foo"), but "data-foo-1" isn't.
> 
> Will address.

Does patchset #2 take care of the issue?

[tkent, 2013-11-12 20:49:34]

Thanks.

Does this work correctly with data-123 and data-0123?
I'd like to ask binding gurus if this approach is the best.

[sof, 2013-11-12 21:27:12]

On 2013/11/12 20:49:34, tkent wrote:
> Thanks.
> 
> Does this work correctly with data-123 and data-0123?

Yes, it ought to as an ES implementation is only allowed to
consider a property name P an array index if
  
   ToString(ToUInt32(P)) == P

(Well, this is an indexed property we're dealing with here, so
an engine like v8 could consider applying a slightly different
rule, but somewhat unlikely and potentially complex.)

I've added some tests to cover this.

> I'd like to ask binding gurus if this approach is the best.

The Storage object also defines indexed accessors this way.

[tkent, 2013-11-12 22:02:42]

lgtm 

> The Storage object also defines indexed accessors this way.

Sounds reasonable.  Thank you!

[commit-bot: I haz the power, 2013-11-12 22:03:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sigbjornf@opera.com/63223002/100001

[sof, 2013-11-12 22:08:41]

On 2013/11/12 22:02:42, tkent wrote:
> lgtm 
> 
> > The Storage object also defines indexed accessors this way.
> 
> Sounds reasonable.  Thank you!

There might be alternate ways to express it, but not a unique approach.

Thanks very much for the review & (slightly devious) test cases :)

[commit-bot: I haz the power, 2013-11-13 00:10:19]

Change committed as 161864