DOM serialization: fix quoting of < and > in HMTL attribute values.

DOM serialization: proper quoting for < and > in HTML attribute values.

< and > should not be quoted in HTML attributes. Internet Explorer and
Firefox do not quote them. Safari (still) does. The DOM serialization
and parsing specification defers to the HTML specification, which says
that < and > should not be quoted.
http://www.w3.org/html/wg/drafts/html/master/syntax.html#attribute's-serialized-name

In XML attributes, < must be quoted. > does not need to be quoted, but
all browsers currently quote both < and > in XML, so we do the same. The
DOM serialization and parsing specification says neither should be
quoted, but there is an outstanding bug, because not quoting < produces
invalid XML.
https://dvcs.w3.org/hg/innerhtml/raw-file/tip/index.html#dfn-concept-serialize-xml-attributes
https://www.w3.org/Bugs/Public/show_bug.cgi?id=24174

This change makes us not quote < and > in HTML attributes, and slightly
expands the test that ensures we quote < and > in XML attributes.

BUG=330949
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=164489

[pwnall-personal, 2013-12-29 01:33:01]

Can you please take a look?

[abarth-chromium, 2014-01-04 05:30:49]

Can you add information to the description about whether this change aligns us
more or less with specifications and with other browsers?

[pwnall-personal, 2014-01-04 07:11:11]

On 2014/01/04 05:30:49, abarth wrote:
> Can you add information to the description about whether this change aligns us
> more or less with specifications and with other browsers?

Done. Can you please take another look?

For HTML attributes, this change gets us closer to the other browsers and the
spec.

For XML attributes, this change gets us further away from other browsers, but
closer to the spec. The spec is slightly broken (see bug), so I tried to follow
its spirit of quoting as little as possible, rather than its letter. Following
its letter would lead to invalid XML, and I filed a bug for W3C to address that.

Also, I can remove the XML change, if you'd rather not deviate from what other
browsers are doing. Un-quoted > is fine according to the XML spec though, it
seems to be desired by the DOM Parsing and Serialization spec, and it produces
smaller XML, which might be marginally faster to parse/send over the wire.

[abarth-chromium, 2014-01-04 17:56:34]

On 2014/01/04 07:11:11, pwnall wrote:
> For HTML attributes, this change gets us closer to the other browsers and the
> spec.

This part seems valuable.

> For XML attributes, this change gets us further away from other browsers, but
> closer to the spec. The spec is slightly broken (see bug), so I tried to
follow
> its spirit of quoting as little as possible, rather than its letter. Following
> its letter would lead to invalid XML, and I filed a bug for W3C to address
that.

It seems like we should keep matching other browsers and get the spec changed to
reflect how browsers behave.  I don't think it's that important to minimize what
we quote.

[pwnall-personal, 2014-01-04 19:28:11]

On 2014/01/04 17:56:34, abarth wrote:
> It seems like we should keep matching other browsers and get the spec changed
to
> reflect how browsers behave.  I don't think it's that important to minimize
what
> we quote.

Done on the first part. I'll add a comment to the W3C that I filed for the
second part.

Can you please take another look?

[abarth-chromium, 2014-01-04 19:33:58]

lgtm

Thanks!

[commit-bot: I haz the power, 2014-01-04 19:34:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/costan@gmail.com/117103003/200001

[commit-bot: I haz the power, 2014-01-06 04:34:50]

Change committed as 164489