IDL: Support runtime enabled overloads affecting Function.length

IDL: Support runtime enabled overloads affecting Function.length

Before this CL, a runtime enabled overload with fewer required arguments
than any non-runtime enabled overload resulted in an error from the IDL
compiler, since such an overload would change the correct Function.length
value (runtime) in a way that wasn't supported.

Since such overloads are now becoming a reality, add support for them by
generating a simple function that returns the correct Function.length
value depending on whether features are enabled.

BUG=437885
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=186393

[Jens Widell, 2014-12-02 11:29:47]

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

[Jens Widell, 2014-12-02 11:29:48]

haraken, bashi: PTAL
ajuma: FYI

This is obviously quite a lot of change to solve something rather trivial. An
alternative approach would be to just calculate the length statically assuming
the default set of enabled features. Function.length would be slightly off
sometimes then, but it's unlikely to actually cause any problems, I think.

[haraken, 2014-12-02 13:23:55]

On 2014/12/02 11:29:48, Jens Widell wrote:
> haraken, bashi: PTAL
> ajuma: FYI
> 
> This is obviously quite a lot of change to solve something rather trivial. An
> alternative approach would be to just calculate the length statically assuming
> the default set of enabled features. Function.length would be slightly off
> sometimes then, but it's unlikely to actually cause any problems, I think.

I'm fine with this approach, but let's wait for bashi-san's comment. LGTM.

[bashi, 2014-12-02 23:56:17]

Thank you for working on this! I think this is worth doing, given that one
actual need and such needs are likely to happen again as we are going to
implement web platform features.

LGTM.

https://codereview.chromium.org/751223003/diff/1/Source/bindings/tests/result...
File Source/bindings/tests/results/core/V8TestObject.cpp (right):

https://codereview.chromium.org/751223003/diff/1/Source/bindings/tests/result...
Source/bindings/tests/results/core/V8TestObject.cpp:8057: if (info.Length() < 0)
{
nit: It's nice to avoid generating this branch, but I guess it makes the
template more complicated. I prefer code simplicity.

[Jens Widell, 2014-12-03 07:19:31]

https://codereview.chromium.org/751223003/diff/1/Source/bindings/tests/result...
File Source/bindings/tests/results/core/V8TestObject.cpp (right):

https://codereview.chromium.org/751223003/diff/1/Source/bindings/tests/result...
Source/bindings/tests/results/core/V8TestObject.cpp:8057: if (info.Length() < 0)
{
On 2014/12/02 23:56:17, bashi1 wrote:
> nit: It's nice to avoid generating this branch, but I guess it makes the
> template more complicated. I prefer code simplicity.

This could be fixed by adding {% if overloads.length != 0 %}, which is simple
enough I think. Done in PS#3.

[bashi, 2014-12-03 07:29:39]

still LGTM.

[Jens Widell, 2014-12-03 08:10:42]

The CQ bit was checked by jl@opera.com

[commit-bot: I haz the power, 2014-12-03 08:11:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/751223003/40001

[commit-bot: I haz the power, 2014-12-03 08:48:08]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=186393