[Vm-dev] Simulator Question

[Andreas Raab]

All you need to support this is in the Android VM. The Android VM calls 
interpret() in response to every input event and interpret() returns 
when the event is handled[*]. If you add an event that gets a selector 
(or some source code to be evaluated) you have just what you were 
looking for.

[*] For a certain definition of "handled", involving aspects such as 
whether delays are involved, or other processes are being scheduled. In 
effect, the VM runs as long as there are active processes, and returns 
when there is no more work to do. Whether or not that corresponds with 
the result of the input expression is a different question.

Cheers,
   - Andreas

On 1/6/2011 5:00 PM, Igor Stasenko wrote:
>
> On 6 January 2011 16:41, Gabriel Hernán Barbuto<gbarbuto at gmail.com>  wrote:
>>
>> Hi
>>
>> I am working with the InterpreterSimulator and I would like to send a
>> message to an object inside the simulated environment. For example, I
>> have the oop for true and I want to send it the message printString.
>>
>> My question is if this is possible. I don't know if there is a way to
>> send a message to an object inside the simulation. I have thought
>> about tinkering with the stack and calling an interpreter method. But
>> I am not sure if this will work.
>>
>
> Hi, Gabriel.
> The problem is, that even if you prepare a context and all arguments
> for message send,
> you should start interpreting the code and then leave the interpret
> loop once you received the answer.
> But if you look at #interpret method, there is an infinite loop,
> and no way how to escape it.
>
> I don't know if there is a way to tell simulator to leave the loop at
> certain point.
>
>
> P.S. In general, it would be cool to have something like:
>
> sqInt interpreter_call( sqInt receiver , sqInt selector , sqInt * arguments )
>
> means: send message to receiver and return the answer.
>
> Unfortunately, Squeak VM does not provides such things, in contrast to
> languages like Lua,
> which providing C interoperability layer by default from very
> beginning of it existence.
>
>> Any hint will be much appreciated. Thanks in advance.
>>
>> Gabriel
>>
>
>