From kde-bindings  Tue Jan 15 18:47:59 2013
From: Arno Rehn <arno () arnorehn ! de>
Date: Tue, 15 Jan 2013 18:47:59 +0000
To: kde-bindings
Subject: Re: [Kde-bindings] Get a pointer to SmokeBinding
Message-Id: <50F5A45F.3080108 () arnorehn ! de>
X-MARC-Message: https://marc.info/?l=kde-bindings&m=135827569914530

On 15/01/13 17:54, j j wrote:
>> I see. This will be supported in smokev4, but not the current Smoke.
>
> Is it already supported in current versions of smokev4? If so, how do I use
> it?
Yep. When you use SmokeMethod::call to call a method you can pass along 
a flag which decides whether to use dynamic dispatching or a direct 
method call.

>> What are you doing with Smoke? I'm interested in every project using it :)
>
> My goal is to write a program that will automatically generate D (
> http://dlang.org) bindings for Smoke libraries such as smokeqt. I'm
> currently only learning to use Smoke and trying to come up with a good
> design for generating class hierarchies, etc. I can't promise anything will
> actually come out of it :D.
Interesting :)
So if you're starting from scratch anyway, I'd suggest you use smokev4 
right from the beginning. The API might not be stable as of now, but for 
a new project it should be fine :)

> On Tue, Jan 15, 2013 at 3:28 PM, Arno Rehn <arno@arnorehn.de> wrote:
>
>> On 15/01/13 00:19, j j wrote:
>>
>>> But if there's a valid use-case I wouldn't object to implementing this in
>>>>
>>> smokev4.
>>>
>>> I'm only learning to use Smoke, so maybe what I want to do can be achieved
>>> in some other way. I wan't to be able to call the most derived method
>>> given
>>> a Smoke object.
>>>
>>> Say I have a Smoke object foo of type Foo, on which I have set a
>>> SmokeBinding bind which overrides bar(). Now say I call foo's method bar()
>>> thorugh the Smoke api. It seems that this results in Foo.bar being called
>>> directly, without bind->callMethod() being called first. I could call the
>>> most derived method if I could get to the binding instance, but if there's
>>> a better way to do it, I'd love to hear about it.
>>>
>> I see. This will be supported in smokev4, but not the current Smoke.
>> Currently we assume that if you create a smoke object, the target language
>> takes care of virtual method resulution and if you call a virtual method
>> through the Smoke API, you really want to call the original C++ method.
>>
>> Now you could do some nasty pointer arithmetic for the time being. The
>> SmokeBinding* pointer is simply appended to the original class and we have
>> size information in Smoke. So something like this could work:
>>
>> char *ptr = reinterpret_cast<char*>(**smokeObject) + smokeClass->size;
>> SmokeBinding *bind = *reinterpret_cast<**SmokeBinding**>(ptr);
>>
>> Note that this is probably undefined behaviour per se ;)
>>
>> What are you doing with Smoke? I'm interested in every project using it :)
>>
>>
>> --
>> Arno Rehn
>> ______________________________**_________________
>> Kde-bindings mailing list
>> Kde-bindings@kde.org
>> https://mail.kde.org/mailman/**listinfo/kde-bindings<https://mail.kde.org/mailman/listinfo/kde-bindings>
>>
>
>
>
> _______________________________________________
> Kde-bindings mailing list
> Kde-bindings@kde.org
> https://mail.kde.org/mailman/listinfo/kde-bindings
>


-- 
Arno Rehn
_______________________________________________
Kde-bindings mailing list
Kde-bindings@kde.org
https://mail.kde.org/mailman/listinfo/kde-bindings