Split [StrictTypeChecking] into [TypeChecking=Interface|Nullable|String]

Split [StrictTypeChecking] into [TypeChecking=Interface|Nullable|String]

Prep for adding support for unrestricted double and float;
no changes to generated bindings.

In order to add proper type-checking to Blink code, we need to do this
on an opt-in basis, to avoid breaking the web interface.

There are lots of different type checking, so simplest is to put these
as values for a single extended attribute, namely [StrictTypeChecking].

Currently there are three types of type checking:
* Interface: if type is interface, does value implement interface?
* Nullable: if value is null, is type nullable?
* String: is a value stringable?

I've split these as 'Interface', 'Nullable', and 'String';
for *almost* all existing [TypeChecking] these are all turned on,
but now we can turn them on separately.

Also:
* Tests added

Docs updated:
http://www.chromium.org/blink/webidl/blink-idl-extended-attributes#TOC-TypeChecking-i-m-a-p-

I'll follow up with:
* [TypeChecking=Unrestricted]
  for floats and doubles, and then we can apply this to existing code,
  tearing out lots of boilerplate checks in Blink!

R=haraken
BUG=354298
BUG=321518
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=172485

[Nils Barth (inactive), 2014-04-24 07:03:55]

Prep for unrestricted; WDYT?

[haraken, 2014-04-24 07:32:49]

Just a clarifying question:

- StrictTypeChecking=Interface|Nullable (<-- This is a common case)
- StrictTypeChecking=Nullable (<-- Can this happen?)
- StrictTypeChecking=Interface (<-- Can this happen?)

Nit: [StrictTypeChecking] => [TypeChecking]

[Nils Barth (inactive), 2014-04-24 07:54:12]

On 2014/04/24 07:32:49, haraken wrote:
> Just a clarifying question:
> 
> - StrictTypeChecking=Interface|Nullable (<-- This is a common case)

Yes, currently this is the only case (b/c we don't distinguish), but it's
probably easier to turn on Interface checking (since interface usually correct)
than Nullability checking (since nullability often wrong in spec and libraries).

> - StrictTypeChecking=Nullable (<-- Can this happen?)

Yes, if you want to check nullablity, but not interface implementation:
[TypeChecking=Nullable] attribute Bar bar; // foo.bar = null throws, but foo.bar
= Foo() doesn't.

> - StrictTypeChecking=Interface (<-- Can this happen?)

Yes, if you want to check interface implementation but not nullablity:
[TypeChecking=Interface] attribute Bar bar; // foo.bar = Foo() throws, but
foo.bar = null doesn't.

These are more useful at the interface level: lets you turn on interface or
nullability checking for all members, without needing to get both correct
immediately.


> Nit: [StrictTypeChecking] => [TypeChecking]

Agreed! Will do.

[haraken, 2014-04-24 07:56:19]

ok, LGTM.

[Nils Barth (inactive), 2014-04-24 08:51:23]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-04-24 08:51:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nbarth@chromium.org/255553004/60001

[Nils Barth (inactive), 2014-04-24 08:53:08]

On 2014/04/24 07:56:19, haraken wrote:
> ok, LGTM.

Thanks!
Turns out there's 1 case of String type checking
(CanvasRenderingContext2D.color), and also we *do* have Interface but no
Nullable or Nullable but no Interface; showed up when I fixed the CG so it
checks these consistently (Element and CanvasRenderingContext2D).

Still no changes to generated code, but now CG is consistent and we do need all
the cases (>.<).

[Nils Barth (inactive), 2014-04-24 08:53:20]

The CQ bit was unchecked by nbarth@chromium.org

[Nils Barth (inactive), 2014-04-24 09:01:19]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-04-24 09:01:35]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/nbarth@chromium.org/255553004/60001

[commit-bot: I haz the power, 2014-04-24 11:15:48]

Change committed as 172485