Oilpan: fix MSVC compilation after r180556.

Oilpan: fix MSVC compilation after r180556.

Have the "same-size" class inherit from the same classes as CSSRule, so
as to have the MSVC-specific alignment constraints of the parent
ScriptWrappableBase class apply in equal measure.

TBR=oilpan-reviews@chromium.org,yukishiino@chromium.org
BUG=235436
NOTRY=true

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180601

[sof, 2014-08-19 21:15:50]

Please take a look.

[sof, 2014-08-19 21:15:56]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-08-19 21:16:42]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sigbjornf@opera.com/487923002/1

[commit-bot: I haz the power, 2014-08-19 21:17:13]

Committed patchset #1 (1) as 180601

[haraken, 2014-08-20 00:17:02]

Thanks for the fix, but this wouldn't be an intended fix... The original CL was
intended to not increase the size of CSSRule.

shiino-san: would you take a look?

[Yuki, 2014-08-20 03:23:45]

On 2014/08/20 00:17:02, haraken wrote:
> Thanks for the fix, but this wouldn't be an intended fix... The original CL
was
> intended to not increase the size of CSSRule.
> 
> shiino-san: would you take a look?

At a glance, I guess that the cause is that CSSRule is too small which is 8 byte
without vtbl when oilpan is enabled.  Without oilpan, CSSRule is 12 byte without
vtbl and the alignment change of ScriptWrappableBase doesn't increase the size
of CSSRule.

May I apply the same temporary fix as
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
?

By the way, I don't understand why trybots (or CQ) didn't catch this issue.  I
think we have oilpan builds at trybots and CQ.

[haraken, 2014-08-20 03:44:03]

On 2014/08/20 03:23:45, Yuki wrote:
> On 2014/08/20 00:17:02, haraken wrote:
> > Thanks for the fix, but this wouldn't be an intended fix... The original CL
> was
> > intended to not increase the size of CSSRule.
> > 
> > shiino-san: would you take a look?
> 
> At a glance, I guess that the cause is that CSSRule is too small which is 8
byte
> without vtbl when oilpan is enabled.  Without oilpan, CSSRule is 12 byte
without
> vtbl and the alignment change of ScriptWrappableBase doesn't increase the size
> of CSSRule.
> 
> May I apply the same temporary fix as
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> ?

Sounds reasonable to me.

> By the way, I don't understand why trybots (or CQ) didn't catch this issue.  I
> think we have oilpan builds at trybots and CQ.

A default set of trybots and CQ don't try oilpan builds. You need to explicitly
choose oilpan trybots to see the build results.

[sof, 2014-08-20 05:41:39]

On 2014/08/20 03:23:45, Yuki wrote:
> On 2014/08/20 00:17:02, haraken wrote:
> > Thanks for the fix, but this wouldn't be an intended fix... The original CL
> was
> > intended to not increase the size of CSSRule.
> > 
> > shiino-san: would you take a look?
> 
> At a glance, I guess that the cause is that CSSRule is too small which is 8
byte
> without vtbl when oilpan is enabled.  Without oilpan, CSSRule is 12 byte
without
> vtbl and the alignment change of ScriptWrappableBase doesn't increase the size
> of CSSRule.
> 
> May I apply the same temporary fix as
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> ?
> 

A comment would do just as fine a job? Having platform specific types/classes
here doesn't sound ideal.

[Yuki, 2014-08-20 08:14:23]

On 2014/08/20 05:41:39, sof wrote:
> On 2014/08/20 03:23:45, Yuki wrote:
> > On 2014/08/20 00:17:02, haraken wrote:
> > > Thanks for the fix, but this wouldn't be an intended fix... The original
CL
> > was
> > > intended to not increase the size of CSSRule.
> > > 
> > > shiino-san: would you take a look?
> > 
> > At a glance, I guess that the cause is that CSSRule is too small which is 8
> byte
> > without vtbl when oilpan is enabled.  Without oilpan, CSSRule is 12 byte
> without
> > vtbl and the alignment change of ScriptWrappableBase doesn't increase the
size
> > of CSSRule.
> > 
> > May I apply the same temporary fix as
> >
>
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...
> > ?
> > 
> 
> A comment would do just as fine a job? Having platform specific types/classes
> here doesn't sound ideal.

Okay, then, I'll just add a comment.