Better arity checks for overloads

Better arity checks for overloads

Overloaded methods have multiple valid argument lengths, not a single minium.
E.g., if valid lengths are 0 or 2, and there is 1 argument provided, this
should be a length error.
Now that we have a switch by length, we can easily handle this!

R=haraken

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=174168

[Nils Barth (inactive), 2014-05-15 06:08:50]

I noticed that we can improve arity checking!

[haraken, 2014-05-15 06:14:30]

+mkwst

LGTM.

[Nils Barth (inactive), 2014-05-16 05:37:52]

(Rebased, will commit after constructors land.)

[Mike West, 2014-05-16 06:35:54]

Code change LGTM. I'm a bit surprised that this didn't change any layout test
results, though. Do we not have coverage for arity exceptions (or did this CL
not change the exception message... it looks like it did?).

[Nils Barth (inactive), 2014-05-16 07:58:13]

On 2014/05/16 06:35:54, Mike West wrote:
> Code change LGTM. I'm a bit surprised that this didn't change any layout test
> results, though. Do we not have coverage for arity exceptions (or did this CL
> not change the exception message... it looks like it did?).

Thanks for review!
It'll change lots of test results (39 tests); haven't rebaselined yet
(this is dependent on earlier CLs).

You can see test results at:
[TEST ONLY] Run tests for 284163002
https://codereview.chromium.org/289063005

Actually, I can simplifying rebasing by using minimum number when that's
correct;
will update.

[Nils Barth (inactive), 2014-05-16 08:30:41]

Revised so now has simple error reporting if possible, which yields much more
legible error messages, and should significantly reduce the number of
rebaselines needed.

[Mike West, 2014-05-16 08:36:42]

Patchset #4 LGTM, thanks for the explanation!

-mike

[Nils Barth (inactive), 2014-05-16 11:19:36]

On 2014/05/16 08:30:41, Nils Barth wrote:
> Revised so now has simple error reporting if possible, which yields much more
> legible error messages, and should significantly reduce the number of
> rebaselines needed.

Turns out only 7 tests actually needed rebaselines:
fast/canvas/canvas-overloads-setStrokeColor.html
fast/canvas/canvas-overloads-setShadow.html
fast/canvas/canvas-overloads-setFillColor.html
fast/canvas/canvas-overloads-drawImage.html
fast/canvas/webgl/texImageTest.html
fast/js/select-options-remove.html
vibration/vibration-exceptions.html

6 of these (all but WebGL) actually have mistakes due to requesting exceptions
when there shouldn't be any!
(Due to too many arguments, which should instead be ignored.)

These will be fixed in the followup:
Overload resolution: ignore extra arguments
https://codereview.chromium.org/291563004/

[Nils Barth (inactive), 2014-05-16 11:19:44]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-05-16 11:19:49]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nbarth@chromium.org/284163002/80001

[Nils Barth (inactive), 2014-05-16 11:34:14]

The CQ bit was unchecked by nbarth@chromium.org

[Nils Barth (inactive), 2014-05-16 11:42:01]

Updated; hope you find the new error messages rather more legible!
(There may be a few more missing to rebaseline.)

[Nils Barth (inactive), 2014-05-16 11:42:06]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-05-16 11:42:11]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nbarth@chromium.org/284163002/100001

[Nils Barth (inactive), 2014-05-16 13:05:25]

The CQ bit was unchecked by nbarth@chromium.org

[Nils Barth (inactive), 2014-05-16 13:05:31]

The CQ bit was checked by nbarth@chromium.org

[commit-bot: I haz the power, 2014-05-16 13:05:43]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/nbarth@chromium.org/284163002/100001

[commit-bot: I haz the power, 2014-05-16 13:44:31]

Change committed as 174168