[Development] QtV8 and the unittests

[Holger Hans Peter Freyther]

On 11/22/2011 10:57 PM, Holger Hans Peter Freyther wrote:
> On 11/22/2011 06:49 PM, Holger Hans Peter Freyther wrote:

Hi again,

the previous points are not answered but I have one more question. Why do we
need to have the QML_GLOBAL_INDEX slot and carry it around? Why can't we just
patch src/contexts.cc with the current code but resolve the qml_global and
qml_global_global through the Global object? What would break/not work if it
is done like this? It would save us from touching 2*Number(ARCHs) jit?

sorry if this question appears stupid.

	holger


>> Hi all,
>>
>> I want to run QtDeclarative on MIPS and I will need to apply the QML changes
>> for this backend. I wonder about the pro/cons of having implementation and
>> tests separated in two different repositories.
> 
> Hi Aaron,
> 
> I think there is no Gerrit for V8 changes itself. I have patched the MIPS
> FullCodeGenerator like it is patched in ARM for the QmlGlobalObjectOperand.
> The test case in QtV8 are passing now. It would be great if the initial commit
> message would go into some more details when and when not to use
> QmlGlobalObjectOperand() and when and when not to use RelocInfo::CODE_TARGET.
> 
> E.g. why does this not use CODE_TARGET?
> 
> FullCodeGenerator::EmitVariableAssignment
>     __ ldr(r1,
> var->is_qml_global()?QmlGlobalObjectOperand():GlobalObjectOperand());
>     Handle<Code> ic = is_strict_mode()
> 
>         ? isolate()->builtins()->StoreIC_Initialize_Strict()
> 
>         : isolate()->builtins()->StoreIC_Initialize();
> 
>     __ Call(ic, RelocInfo::CODE_TARGET_CONTEXT);
> 
> thanks
> 	holger
> 
> 
> PS: Will work on the global compare tomorrow.
> 
> 
> 
> _______________________________________________
> Development mailing list
> Development at qt-project.org
> http://lists.qt-project.org/mailman/listinfo/development