Revert 53892 - Initial scriptable object implementation....

Revert 53892 - Initial scriptable object implementation.

Broke ARM build, reverting.  Errors from log:
remoting/client/plugin/chromoting_scriptable_object.cc: In member function 'virtual bool remoting::ChromotingScriptableObject::HasProperty(const pp::Var&, pp::Var*)':
remoting/client/plugin/chromoting_scriptable_object.cc:48: error: NULL used in arithmetic
remoting/client/plugin/chromoting_scriptable_object.cc: In member function 'virtual bool remoting::ChromotingScriptableObject::HasMethod(const pp::Var&, pp::Var*)':
remoting/client/plugin/chromoting_scriptable_object.cc:63: error: NULL used in arithmetic

BUG=50248
TEST=write javascript to manually setup connection.

Review URL: http://codereview.chromium.org/3064009

TBR=ajwong@chromium.org

[kobi, 2010-07-28 18:10:18]

pp:Var has nothing to do with the error.
typedef pp::Var (ChromotingScriptableObject::*MethodHandler)(	
const std::vector<pp::Var>& args, pp::Var* exception);

Is just like 
typedef RET_VAL (CLASS::*type_name) (arg1, arg2, ...)

So the following should be legal:

type_name foo = 0;

Maybe arm does not like NULL. Basically, I always use 0 and not NULL to be cross
platform compatible.

http://codereview.chromium.org/2856067/diff/1/7
File remoting/client/plugin/chromoting_scriptable_object.cc (left):

http://codereview.chromium.org/2856067/diff/1/7#oldcode48
remoting/client/plugin/chromoting_scriptable_object.cc:48: return
properties_[iter->second].method == NULL;
That one should be a valid statement. I would use 0 instead of NULL and see if
it makes things better.

[bryeung1, 2010-07-28 18:12:27]

Thanks for the suggestion.

I'm just sheriff, so I'll let ajwong sort this out when he comes online.

Bryan

On Wed, Jul 28, 2010 at 11:10 AM,  <kobic@codeaurora.org> wrote:
> pp:Var has nothing to do with the error.
> typedef pp::Var (ChromotingScriptableObject::*MethodHandler)(
> const std::vector<pp::Var>& args, pp::Var* exception);
>
> Is just like
> typedef RET_VAL (CLASS::*type_name) (arg1, arg2, ...)
>
> So the following should be legal:
>
> type_name foo = 0;
>
> Maybe arm does not like NULL. Basically, I always use 0 and not NULL to be
> cross
> platform compatible.
>
>
> http://codereview.chromium.org/2856067/diff/1/7
> File remoting/client/plugin/chromoting_scriptable_object.cc (left):
>
> http://codereview.chromium.org/2856067/diff/1/7#oldcode48
> remoting/client/plugin/chromoting_scriptable_object.cc:48: return
> properties_[iter->second].method == NULL;
> That one should be a valid statement. I would use 0 instead of NULL and
> see if it makes things better.
>
> http://codereview.chromium.org/2856067/show
>

[awong (On leave), 2010-07-28 19:52:11]

I am out for a few days.  Gary...could I trouble to fix this for me and
resubmit?  I'm guessing it should be easy.  As for null versus 0, iirc c++
defines in the spec that NULL is exactly the same as the literal 0.  But
maybe I am wrong.

-Albert

On Jul 28, 2010 11:12 AM, "Bryan Yeung" <bryeung@google.com> wrote:
> Thanks for the suggestion.
>
> I'm just sheriff, so I'll let ajwong sort this out when he comes online.
>
> Bryan
>
> On Wed, Jul 28, 2010 at 11:10 AM, <kobic@codeaurora.org> wrote:
>> pp:Var has nothing to do with the error.
>> typedef pp::Var (ChromotingScriptableObject::*MethodHandler)(
>> const std::vector<pp::Var>& args, pp::Var* exception);
>>
>> Is just like
>> typedef RET_VAL (CLASS::*type_name) (arg1, arg2, ...)
>>
>> So the following should be legal:
>>
>> type_name foo = 0;
>>
>> Maybe arm does not like NULL. Basically, I always use 0 and not NULL to
be
>> cross
>> platform compatible.
>>
>>
>> http://codereview.chromium.org/2856067/diff/1/7
>> File remoting/client/plugin/chromoting_scriptable_object.cc (left):
>>
>> http://codereview.chromium.org/2856067/diff/1/7#oldcode48
>> remoting/client/plugin/chromoting_scriptable_object.cc:48: return
>> properties_[iter->second].method == NULL;
>> That one should be a valid statement. I would use 0 instead of NULL and
>> see if it makes things better.
>>
>> http://codereview.chromium.org/2856067/show
>>

[awong (On leave), 2010-07-29 00:33:18]

Just verified that Kobic is right in that comparing versus NULL does not
work here.  I think this sounds like a compiler error.  Added a TODO to
check later and submitting with 0 swapped in for NULL.

Thanks for the tip!

-Albert


On Wed, Jul 28, 2010 at 12:52 PM, Albert J. Wong (王重傑)
<ajwong@chromium.org>wrote:

> I am out for a few days.  Gary...could I trouble to fix this for me and
> resubmit?  I'm guessing it should be easy.  As for null versus 0, iirc c++
> defines in the spec that NULL is exactly the same as the literal 0.  But
> maybe I am wrong.
>
> -Albert
>
> On Jul 28, 2010 11:12 AM, "Bryan Yeung" <bryeung@google.com> wrote:
> > Thanks for the suggestion.
> >
> > I'm just sheriff, so I'll let ajwong sort this out when he comes online.
> >
> > Bryan
> >
> > On Wed, Jul 28, 2010 at 11:10 AM, <kobic@codeaurora.org> wrote:
> >> pp:Var has nothing to do with the error.
> >> typedef pp::Var (ChromotingScriptableObject::*MethodHandler)(
> >> const std::vector<pp::Var>& args, pp::Var* exception);
> >>
> >> Is just like
> >> typedef RET_VAL (CLASS::*type_name) (arg1, arg2, ...)
> >>
> >> So the following should be legal:
> >>
> >> type_name foo = 0;
> >>
> >> Maybe arm does not like NULL. Basically, I always use 0 and not NULL to
> be
> >> cross
> >> platform compatible.
> >>
> >>
> >> http://codereview.chromium.org/2856067/diff/1/7
> >> File remoting/client/plugin/chromoting_scriptable_object.cc (left):
> >>
> >> http://codereview.chromium.org/2856067/diff/1/7#oldcode48
> >> remoting/client/plugin/chromoting_scriptable_object.cc:48: return
> >> properties_[iter->second].method == NULL;
> >> That one should be a valid statement. I would use 0 instead of NULL and
> >> see if it makes things better.
> >>
> >> http://codereview.chromium.org/2856067/show
> >>
>