Sync the DOM* geometry interfaces with the spec

Sync the DOM* geometry interfaces with the spec

http://dev.w3.org/fxtf/geometry/

Note that DOMRect is missing the optional keyword in the spec:
https://www.w3.org/Bugs/Public/show_bug.cgi?id=27706

BUG=460722
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=190930

[philipj_slow, 2015-02-26 04:58:42]

philipj@opera.com changed reviewers:
+ jinho.bang@samsung.com

[philipj_slow, 2015-02-26 04:58:42]

PTAL

[zino, 2015-02-26 12:11:32]

Looks good. +haraken since he's the idl expert and owner.

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl
File Source/core/dom/DOMPoint.idl (right):

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl...
Source/core/dom/DOMPoint.idl:9: // FIXME: The x and y arguments should be
optional (See: crbug.com/462098)
I'm not sure but it seems to be spec[1] bug(?).

Constructor(optional DOMPointInit point)
 - Constructor() -> **conflict**
 - Constructor(DOMPointInit point)

Constructor(optional unrestricted double x, optional unrestricted double y, ...)
 - Constructor() -> **conflict**
 - Constructor(x)
 - Constructor(x, y)
 - ...

I didn't find the case of non-parameter like this situation in web-idl spec[2]
but I believe the two constructors are not distinguishable.

[1] http://dev.w3.org/fxtf/geometry/
[2] http://www.w3.org/TR/2012/CR-WebIDL-20120419/#distinguishable-table

[zino, 2015-02-26 12:12:18]

jinho.bang@samsung.com changed reviewers:
+ haraken@chromium.org

[zino, 2015-02-26 12:12:18]

+haraken since he's the idl expert and owner.

[haraken, 2015-02-26 12:40:58]

LGTM

[zino, 2015-02-26 13:51:15]

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl
File Source/core/dom/DOMPoint.idl (right):

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl...
Source/core/dom/DOMPoint.idl:9: // FIXME: The x and y arguments should be
optional (See: crbug.com/462098)
On 2015/02/26 12:11:32, zino wrote:
> I'm not sure but it seems to be spec[1] bug(?).
> 
> Constructor(optional DOMPointInit point)
>  - Constructor() -> **conflict**
>  - Constructor(DOMPointInit point)
> 
> Constructor(optional unrestricted double x, optional unrestricted double y,
...)
>  - Constructor() -> **conflict**
>  - Constructor(x)
>  - Constructor(x, y)
>  - ...
> 
> I didn't find the case of non-parameter like this situation in web-idl spec[2]
> but I believe the two constructors are not distinguishable.
> 
> [1] http://dev.w3.org/fxtf/geometry/
> [2] http://www.w3.org/TR/2012/CR-WebIDL-20120419/#distinguishable-table

Hi @haraken,

Could you please take a look?

[haraken, 2015-02-26 13:54:03]

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

[haraken, 2015-02-26 13:54:04]

+Jens

[Jens Widell, 2015-02-26 14:02:56]

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl
File Source/core/dom/DOMPoint.idl (right):

https://codereview.chromium.org/956203002/diff/1/Source/core/dom/DOMPoint.idl...
Source/core/dom/DOMPoint.idl:9: // FIXME: The x and y arguments should be
optional (See: crbug.com/462098)
On 2015/02/26 13:51:15, zino wrote:
> On 2015/02/26 12:11:32, zino wrote:
> > I'm not sure but it seems to be spec[1] bug(?).
> > 
> > Constructor(optional DOMPointInit point)
> >  - Constructor() -> **conflict**
> >  - Constructor(DOMPointInit point)
> > 
> > Constructor(optional unrestricted double x, optional unrestricted double y,
> ...)
> >  - Constructor() -> **conflict**
> >  - Constructor(x)
> >  - Constructor(x, y)
> >  - ...
> > 
> > I didn't find the case of non-parameter like this situation in web-idl
spec[2]
> > but I believe the two constructors are not distinguishable.
> > 
> > [1] http://dev.w3.org/fxtf/geometry/
> > [2] http://www.w3.org/TR/2012/CR-WebIDL-20120419/#distinguishable-table
> 
> Hi @haraken,
> 
> Could you please take a look?

It's a spec bug, alright. WebIDL doesn't allow (among other things) two
overloads that support being called with zero arguments.

The most natural fix IMO would be to make the DOMPointInit argument
non-optional, and leave all the coordinate arguments to the other constructor
optional (with default values.)

[philipj_slow, 2015-02-26 14:19:36]

Thanks for the feedback, I've submitted a spec bug:
https://www.w3.org/Bugs/Public/show_bug.cgi?id=28108

I'll update the FIXME to point to the spec bug instead.

[philipj_slow, 2015-02-26 15:00:48]

add inherit and tweak constructors

[philipj_slow, 2015-02-26 15:19:50]

Jens, can you take another look? The diff from PS1 is rather trivial, but better
safe than sorry.

[Jens Widell, 2015-02-26 15:24:31]

LGTM

No unexpected changes to generated code.

[philipj_slow, 2015-02-26 15:43:53]

The CQ bit was checked by philipj@opera.com

[philipj_slow, 2015-02-26 15:43:53]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/956203002/#ps20001 (title:
"add inherit and tweak constructors")

[commit-bot: I haz the power, 2015-02-26 15:44:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/956203002/20001

[philipj_slow, 2015-02-26 15:55:13]

By the way, is there a very quick way to compare the generated code? I create a
git repo in the gen folder, then recompile, then use git diff. It works, but
it's a bit of overhead every time.

[Jens Widell, 2015-02-26 15:56:40]

On 2015/02/26 15:55:13, philipj_UTC7 wrote:
> By the way, is there a very quick way to compare the generated code? I create
a
> git repo in the gen folder, then recompile, then use git diff. It works, but
> it's a bit of overhead every time.

I have some simple shell scripts that do exactly that.

[haraken, 2015-02-26 16:03:06]

On 2015/02/26 15:56:40, Jens Widell wrote:
> On 2015/02/26 15:55:13, philipj_UTC7 wrote:
> > By the way, is there a very quick way to compare the generated code? I
create
> a
> > git repo in the gen folder, then recompile, then use git diff. It works, but
> > it's a bit of overhead every time.
> 
> I have some simple shell scripts that do exactly that.

Me too...

[commit-bot: I haz the power, 2015-02-26 16:10:25]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=190930

[philipj_slow, 2015-02-26 18:18:09]

OK, I guess I'll write myself a shell script too :)