Schema support for CSS3 transitions module

This adds schema support for "CSS Transitions" as speced at
http://dev.w3.org/csswg/css-transitions

OPEN QUESTION: How do we sanitize values of <single-transition-property>?
Check that they are allowed in the CSS white-list?
Should this be special cased?

> 2. Transitions

> Normally when the value of a CSS property changes, the rendered result
> is instantly updated, with the affected elements immediately changing
> from the old property value to the new property value. This section
> describes a way to specify transitions using new CSS properties.
> These properties are used to animate smoothly from the old state to
> the new state over time.

> For example, suppose that transitions of one second have been defined
> on the ‘left’ and ‘background-color’ properties. The following diagram
> illustrates the effect of updating those properties on an element, in
> this case moving it to the right and changing the background from red
> to blue. This assumes other transition parameters still have their
> default values.

----

Submitted @ r5512

[felix8a, 2013-07-26 23:00:18]

lgtm.

I'm ok with allowing <single-transition-property> to match /[\w\-]+/
I can't think of any particular security risk to that, since this can't affect
anything outside the guest pseudo-document.

[MikeSamuel, 2013-07-26 23:15:46]

On 2013/07/26 23:00:18, felix8a wrote:
> lgtm.
> 
> I'm ok with allowing <single-transition-property> to match /[\w\-]+/
> I can't think of any particular security risk to that, since this can't affect
> anything outside the guest pseudo-document.

Yeah.  I didn't want to run into a case where unreserved words can have multiple
possible meanings so I ended up doing a similar trick to quoted strings to make
sure they're unambiguous.

[MikeSamuel, 2013-07-26 23:16:01]

This adds schema support for "CSS Transitions" as speced at
http://dev.w3.org/csswg/css-transitions

OPEN QUESTION: How do we sanitize values of <single-transition-property>?
Check that they are allowed in the CSS white-list?
Should this be special cased?

> 2. Transitions

> Normally when the value of a CSS property changes, the rendered result
> is instantly updated, with the affected elements immediately changing
> from the old property value to the new property value. This section
> describes a way to specify transitions using new CSS properties.
> These properties are used to animate smoothly from the old state to
> the new state over time.

> For example, suppose that transitions of one second have been defined
> on the ‘left’ and ‘background-color’ properties. The following diagram
> illustrates the effect of updating those properties on an element, in
> this case moving it to the right and changing the background from red
> to blue. This assumes other transition parameters still have their
> default values.

[felix8a, 2013-07-26 23:35:27]

https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
File tests/com/google/caja/plugin/sanitizecss_test.js (right):

https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
tests/com/google/caja/plugin/sanitizecss_test.js:579:
'transition-property:opacity , , left;',  // bogus elided
eliding properties makes the value invalid, there needs to be at least a
placeholder item. the spec doesn't define any allowed null value, the keyword
'none' is invalid in this context. but if we really want to filter invalid
props, we can use some value like 'invalid'.

[MikeSamuel, 2013-07-27 02:19:21]

https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
File tests/com/google/caja/plugin/sanitizecss_test.js (right):

https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
tests/com/google/caja/plugin/sanitizecss_test.js:579:
'transition-property:opacity , , left;',  // bogus elided
On 2013/07/26 23:35:27, felix8a wrote:
> eliding properties makes the value invalid, there needs to be at least a
> placeholder item. the spec doesn't define any allowed null value, the keyword
> 'none' is invalid in this context. but if we really want to filter invalid
> props, we can use some value like 'invalid'.

We're doing token filtering now.  We could drop the property entirely, or have a
per-property substitution token, but that's a fairly significant change.

[felix8a, 2013-07-27 02:54:56]

On 2013/07/27 02:19:21, MikeSamuel wrote:
>
https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
> File tests/com/google/caja/plugin/sanitizecss_test.js (right):
> 
>
https://codereview.appspot.com/11949043/diff/5001/tests/com/google/caja/plugi...
> tests/com/google/caja/plugin/sanitizecss_test.js:579:
> 'transition-property:opacity , , left;',  // bogus elided
> On 2013/07/26 23:35:27, felix8a wrote:
> > eliding properties makes the value invalid, there needs to be at least a
> > placeholder item. the spec doesn't define any allowed null value, the
keyword
> > 'none' is invalid in this context. but if we really want to filter invalid
> > props, we can use some value like 'invalid'.
> 
> We're doing token filtering now.  We could drop the property entirely, or have
a
> per-property substitution token, but that's a fairly significant change.

ok, lgtm. if the elision becomes a problem we can fix that later.