Add a public Blink function to canonicalize and restrict the type of CSS selectors.

Add a public Blink function to canonicalize CSS selectors and restrict their type.

This adds the ability for Blink embedders to check that a selector is
appropriate in the <content select> attribute or the planned
WebDocument::watchSelectors() API, which is currently impossible in
Javascript.

This also adds a C++ equivalent of writing a rule into a stylesheet
and then retrieving its selector from CSSStyleRule.selectorText, which
is also useful for the WebDocument::watchSelectors() API, which
reports results using the selectorText string.

BUG=172011
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=156880

[Jeffrey Yasskin, 2013-08-28 01:18:33]

Please forward to whoever's best to look at the API expansion.  I'm hoping to
use this to replace the JS code in https://codereview.chromium.org/23478003/.
Thanks.

[abarth-chromium, 2013-08-28 06:29:37]

Seems fine.  LGTM but I think we should go one more round on the API.

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h
File public/web/WebSelector.h (right):

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:38: struct WebSelectorRestriction {
What's the point of this class?  Why not just declare the enum in the WebKit
namespace and pass that directly to canonicalizeSelector?

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:40: WebSelectorRestriction();
WEBKIT_EXPORT  <--- needed for the component build

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:44: Complex,
Complex -> SelectorTypeComplex

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:45: Compound,
Compound -> SelectorTypeCompound

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:57: WebString canonicalizeSelector(WebString selector,
WebSelectorRestriction = WebSelectorRestriction());
EXPORT_WEBKIT

[Jeffrey Yasskin, 2013-08-28 17:55:14]

Thanks!

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h
File public/web/WebSelector.h (right):

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:38: struct WebSelectorRestriction {
On 2013/08/28 06:29:37, abarth wrote:
> What's the point of this class?  Why not just declare the enum in the WebKit
> namespace and pass that directly to canonicalizeSelector?

I was thinking that we may want more restrictions in the future (e.g.
single-selector vs list, only certain kinds of simple selectors), and it'll be
better to express those in a struct than in a mask type or in multiple
parameters, but I guess we can change the canonicalizeSelector() signature when
that happens. I've switched to a raw enum.

Should the enum's name be WebKit::SelectorType or WebKit::WebSelectorType? And
if WebSelectorType, should the enumerators be WebSelectorTypeComplex, etc.?

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:40: WebSelectorRestriction();
On 2013/08/28 06:29:37, abarth wrote:
> WEBKIT_EXPORT  <--- needed for the component build

Argh, thanks. I wish the tests lived outside libwebkit.so so they'd catch things
like this.

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:44: Complex,
On 2013/08/28 06:29:37, abarth wrote:
> Complex -> SelectorTypeComplex

Done.

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:45: Compound,
On 2013/08/28 06:29:37, abarth wrote:
> Compound -> SelectorTypeCompound

Done.

https://codereview.chromium.org/23659002/diff/7001/public/web/WebSelector.h#n...
public/web/WebSelector.h:57: WebString canonicalizeSelector(WebString selector,
WebSelectorRestriction = WebSelectorRestriction());
On 2013/08/28 06:29:37, abarth wrote:
> EXPORT_WEBKIT

Done.

[abarth-chromium, 2013-08-28 19:11:47]

On 2013/08/28 17:55:14, Jeffrey Yasskin wrote:
> Should the enum's name be WebKit::SelectorType or WebKit::WebSelectorType? And
> if WebSelectorType, should the enumerators be WebSelectorTypeComplex, etc.?

Yes, the style we use in the API so repeat the name of the enum in its values.

[abarth-chromium, 2013-08-28 19:12:41]

lgtm

https://codereview.chromium.org/23659002/diff/9001/public/web/WebSelector.h
File public/web/WebSelector.h (right):

https://codereview.chromium.org/23659002/diff/9001/public/web/WebSelector.h#n...
public/web/WebSelector.h:41: enum SelectorType {
Now that this is in the global WebKit scope, it should have a "Web" prefix
(i.e., WebSelectorType).

[Jeffrey Yasskin, 2013-08-28 19:19:26]

Thanks again.

https://codereview.chromium.org/23659002/diff/9001/public/web/WebSelector.h
File public/web/WebSelector.h (right):

https://codereview.chromium.org/23659002/diff/9001/public/web/WebSelector.h#n...
public/web/WebSelector.h:41: enum SelectorType {
On 2013/08/28 19:12:41, abarth wrote:
> Now that this is in the global WebKit scope, it should have a "Web" prefix
> (i.e., WebSelectorType).

Done.

[commit-bot: I haz the power, 2013-08-28 19:28:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jyasskin@chromium.org/23659002/20001

[commit-bot: I haz the power, 2013-08-28 21:33:57]

Change committed as 156880