IDL: Support optional union type arguments with default values

IDL: Support optional union type arguments with default values

Since union container types have setters rather than assignment operators,
restructure the default value handling to generate a C++ expression that
sets the value instead of the value as a C++ literal expression.

And as a special case, since union containers are null initially, if the
default value is null, generate an empty expression then.

This also fixes a minor bug in the union handling: enum types were not
supported. They should be treated as string types, including conflicting
with any other string type in the union.

BUG=240176
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185046

[Jens Widell, 2014-11-10 09:36:59]

jl@opera.com changed reviewers:
+ bashi@chromium.org, haraken@chromium.org

[Jens Widell, 2014-11-10 09:37:00]

PTAL

This may be considered slightly overkill. I don't know if there is any union
type argument in any specification with a default value that isn't null. (I only
know of  HTMLSelectElement.add()'s second argument which defaults to null.)

I feel the implementation is simple enough, though.

[haraken, 2014-11-10 11:20:04]

Looks good.

Can we add unit tests for this? I want to have a test for an optional null, an
optional undefined etc.

[Jens Widell, 2014-11-10 12:51:11]

On 2014/11/10 11:20:04, haraken wrote:
> Can we add unit tests for this? I want to have a test for an optional null, an
> optional undefined etc.

Some unit tests added. Do they cover what you were after?

[haraken, 2014-11-10 12:55:42]

Thanks, LGTM.

https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
File Source/core/testing/UnionTypesTest.idl (right):

https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
Source/core/testing/UnionTypesTest.idl:14: [ImplementedAs=doubleOrStringArg]
DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg =
null);

Can we add:

  DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg =
undefined);

?

[Jens Widell, 2014-11-10 12:59:10]

https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
File Source/core/testing/UnionTypesTest.idl (right):

https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
Source/core/testing/UnionTypesTest.idl:14: [ImplementedAs=doubleOrStringArg]
DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg =
null);
On 2014/11/10 12:55:42, haraken wrote:
> 
> Can we add:
> 
>   DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg =
> undefined);
> 
> ?

Undefined is not a possible default value per spec (or in our implementation.)
The grammar only supports "null" as a special value.

[haraken, 2014-11-10 12:59:42]

On 2014/11/10 12:59:10, Jens Widell wrote:
>
https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
> File Source/core/testing/UnionTypesTest.idl (right):
> 
>
https://codereview.chromium.org/713683003/diff/20001/Source/core/testing/Unio...
> Source/core/testing/UnionTypesTest.idl:14: [ImplementedAs=doubleOrStringArg]
> DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg =
> null);
> On 2014/11/10 12:55:42, haraken wrote:
> > 
> > Can we add:
> > 
> >   DOMString doubleOrStringDefaultNullArg(optional (double or DOMString)? arg
=
> > undefined);
> > 
> > ?
> 
> Undefined is not a possible default value per spec (or in our implementation.)
> The grammar only supports "null" as a special value.

ah, makes sense.

[Jens Widell, 2014-11-10 13:35:03]

The CQ bit was checked by jl@opera.com

[commit-bot: I haz the power, 2014-11-10 13:36:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/713683003/20001

[commit-bot: I haz the power, 2014-11-10 15:31:13]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-11-10 15:31:14]

Try jobs failed on following builders:
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/35595)

[Jens Widell, 2014-11-10 15:31:58]

The CQ bit was checked by jl@opera.com

[commit-bot: I haz the power, 2014-11-10 15:32:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/713683003/20001

[commit-bot: I haz the power, 2014-11-10 17:37:08]

Committed patchset #2 (id:20001) as 185046

[bashi, 2014-11-10 23:21:39]

LGTM