IDL: Support default values for dictionary members of union types

IDL: Support default values for dictionary members of union types

Fix in three parts:

1) For each union member, add a static "fromThatType(ThatType value)"
   function, that constructs and returns a union type container object.

2) Add IdlUnionType.literal_cpp_value() that returns an C++ expression
   using a static function (added in 1) that matches the literal value.

3) Use idl_type.literal_cpp_value(literal_value) in v8_dictionary.py
   instead of str(literal_value), thus using the union type override
   (added in 2) for union type members.

BUG=240176, 321462
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=190464

[Jens Widell, 2015-02-18 14:47:07]

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

[Jens Widell, 2015-02-18 14:47:08]

PTAL

[haraken, 2015-02-18 15:09:04]

https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
File Source/bindings/scripts/v8_types.py (right):

https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
Source/bindings/scripts/v8_types.py:900: def union_literal_cpp_value(idl_type,
idl_literal):

Can we merge literal_cpp_value and union_literal_cpp_value? (e.g., if idl_type
knows if it's a union type or not, we can switch the logic in
literal_cpp_value.)

[Jens Widell, 2015-02-18 15:14:39]

https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
File Source/bindings/scripts/v8_types.py (right):

https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
Source/bindings/scripts/v8_types.py:900: def union_literal_cpp_value(idl_type,
idl_literal):
On 2015/02/18 15:09:04, haraken wrote:
> 
> Can we merge literal_cpp_value and union_literal_cpp_value? (e.g., if idl_type
> knows if it's a union type or not, we can switch the logic in
> literal_cpp_value.)

We can implement a single function that handles both IdlType and IdlUnionType,
yes. (It would contain the same amount of code, only arranged slightly
differently.)

That single function would then need to be on IdlTypeBase, and then it would
apply to the other types as well.

Generally, though, I prefer to keep the handling in two separate function, like
this, unless there's some other gain, such as possibility of code sharing.

[haraken, 2015-02-18 15:19:55]

On 2015/02/18 15:14:39, Jens Widell wrote:
>
https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
> File Source/bindings/scripts/v8_types.py (right):
> 
>
https://codereview.chromium.org/937773002/diff/1/Source/bindings/scripts/v8_t...
> Source/bindings/scripts/v8_types.py:900: def union_literal_cpp_value(idl_type,
> idl_literal):
> On 2015/02/18 15:09:04, haraken wrote:
> > 
> > Can we merge literal_cpp_value and union_literal_cpp_value? (e.g., if
idl_type
> > knows if it's a union type or not, we can switch the logic in
> > literal_cpp_value.)
> 
> We can implement a single function that handles both IdlType and IdlUnionType,
> yes. (It would contain the same amount of code, only arranged slightly
> differently.)
> 
> That single function would then need to be on IdlTypeBase, and then it would
> apply to the other types as well.
> 
> Generally, though, I prefer to keep the handling in two separate function,
like
> this, unless there's some other gain, such as possibility of code sharing.

OK, LGTM.

[bashi, 2015-02-19 01:33:44]

LGTM

[Jens Widell, 2015-02-19 06:07:54]

The CQ bit was checked by jl@opera.com

[commit-bot: I haz the power, 2015-02-19 09:34:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/937773002/1

[commit-bot: I haz the power, 2015-02-19 13:12:04]

Committed patchset #1 (id:1) as
https://src.chromium.org/viewvc/blink?view=rev&revision=190464