When attributes have no specified value, specify empty string

We've been rewriting attributes with no specified value to themselves:
   <div foo>bar</div>
currently has the same semantics as
   <div foo="foo">bar</div>
However, this is wrong according to the spec:
http://www.whatwg.org/specs/web-apps/current-work/multipage/tokenization.html#after-attribute-name-state
and current browser behavior.

Fixes http://code.google.com/p/google-caja/issues/detail?id=1602 .

[kpreid2, 2012-12-20 22:49:49]

LGTM

(I accidentally read the whole thing)

[felix8a, 2012-12-20 22:54:27]

relevant section in html5 spec is actually here
http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#attri...
"Empty attribute syntax: ... The value is implicitly the empty string."

[felix8a, 2012-12-20 23:01:04]

and the contradictory section in the html4 spec is the "Boolean attributes"
section here:
http://www.w3.org/TR/REC-html40/intro/sgmltut.html#h-3.3.4.2

[metaweta, 2012-12-20 23:30:44]

On 2012/12/20 23:01:04, felix8a wrote:
> and the contradictory section in the html4 spec is the "Boolean attributes"
> section here:
> http://www.w3.org/TR/REC-html40/intro/sgmltut.html#h-3.3.4.2

Both Chrome and FF just check for the existence of boolean attributes rather
than their value.  Anyone know about IE9+?

[felix8a, 2012-12-21 00:11:17]

IE9 seems to normalize <input checked> into <input checked="">

getAttribute("checked") returns
 "" in IE9 mode
 "checked" in IE8 mode
 true in IE7 mode and quirks mode

[metaweta, 2012-12-21 00:12:44]

On 2012/12/20 23:30:44, metaweta wrote:
> On 2012/12/20 23:01:04, felix8a wrote:
> > and the contradictory section in the html4 spec is the "Boolean attributes"
> > section here:
> > http://www.w3.org/TR/REC-html40/intro/sgmltut.html#h-3.3.4.2
> 
> Both Chrome and FF just check for the existence of boolean attributes rather
> than their value.  Anyone know about IE9+?

IE9 also behaves this way.  I don't have access to IE10, but I don't think it
will change.  I'm going to submit despite this part of the spec.

[felix8a, 2012-12-21 00:13:42]

On 2012/12/21 00:12:44, metaweta wrote:
> On 2012/12/20 23:30:44, metaweta wrote:
> > On 2012/12/20 23:01:04, felix8a wrote:
> > > and the contradictory section in the html4 spec is the "Boolean
attributes"
> > > section here:
> > > http://www.w3.org/TR/REC-html40/intro/sgmltut.html#h-3.3.4.2
> > 
> > Both Chrome and FF just check for the existence of boolean attributes rather
> > than their value.  Anyone know about IE9+?
> 
> IE9 also behaves this way.  I don't have access to IE10, but I don't think it
> will change.  I'm going to submit despite this part of the spec.

sgtm