[oilpan] Handlify childrenChanged.

[oilpan] Handlify childrenChanged.

R=haraken@chromium.org
BUG=

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=155585

[haraken, 2013-08-06 01:32:50]

LGTM! Thanks for the small CL.

https://codereview.chromium.org/22043003/diff/1/Source/core/html/HTMLElement.cpp
File Source/core/html/HTMLElement.cpp (right):

https://codereview.chromium.org/22043003/diff/1/Source/core/html/HTMLElement....
Source/core/html/HTMLElement.cpp:847: TextDirection
HTMLElement::directionality(Handle<Node>* strongDirectionalityTextNode) const

Could you change Handle<Node>* to Handle<Node>& ? It looks like a bit dangerous
programming pattern to expose Handle<Node>*.

[zerny-google, 2013-08-06 07:54:25]

On 2013/08/06 01:32:50, haraken wrote:
> LGTM! Thanks for the small CL.
> 
>
https://codereview.chromium.org/22043003/diff/1/Source/core/html/HTMLElement.cpp
> File Source/core/html/HTMLElement.cpp (right):
> 
>
https://codereview.chromium.org/22043003/diff/1/Source/core/html/HTMLElement....
> Source/core/html/HTMLElement.cpp:847: TextDirection
> HTMLElement::directionality(Handle<Node>* strongDirectionalityTextNode) const
> 
> Could you change Handle<Node>* to Handle<Node>& ? It looks like a bit
dangerous
> programming pattern to expose Handle<Node>*.

Not without changing the call sites too, since strongDirectionalityTextNode has
a default value which would need to be explicitly passed at each call site.
Also, using the non-const reference makes it less clear that
strongDirectionalityTextNode is an outgoing parameter.

[zerny-google, 2013-08-06 07:56:38]

Committed patchset #1 manually as r155585 (presubmit successful).

[haraken, 2013-08-06 07:58:09]

> Not without changing the call sites too, since strongDirectionalityTextNode
has
> a default value which would need to be explicitly passed at each call site.
> Also, using the non-const reference makes it less clear that
> strongDirectionalityTextNode is an outgoing parameter.

Sounds reasonable. The core problem would be we're mixing const& and * for
outgoing parameters :)

[haraken, 2013-08-07 03:25:31]

Interestingly, this CL fixed (or just hid?) the failure of
fast/forms/calendar-picker/week-picker-key-operations.html that was caused by
r155131. The test is now passing :) Will investigate when I have time.