Please consider this code:

Program program = new Program;
RegisterDScriptTypes(program);
program.compile("test", "var d=new String('bobef');", null);
Value ret=*new Value;
Value[] args;
args.length=1;
args[0]=*new Value;
args[0].putVstring("Main(d);");
Dglobal_eval(null,program.callcontext,null,&ret,args);


//end

RegisterDScriptTypes() registers global function Main which displays a message box. Executing the code above produces a message box saying "undefined" (main works ok - tested). So it is clear to me that the call to eval() doesn't know about 'var d...'. And I am not passing any objets to it. I tried with program.callcontext.global - it does the same. What I need to do to make it working?

After two days experimenting I was able to answer myself...

bobef wrote:
> Please consider this code:
> 
> Program program = new Program;
> RegisterDScriptTypes(program);
> program.compile("test", "var d=new String('bobef');", null);
> Value ret=*new Value;
> Value[] args;
> args.length=1;
> args[0]=*new Value;
> args[0].putVstring("Main(d);");
> Dglobal_eval(null,program.callcontext,null,&ret,args);
> 
> 
> //end
> 
> RegisterDScriptTypes() registers global function Main which displays a message
> box. Executing the code above produces a message box saying "undefined" (main
> works ok - tested). So it is clear to me that the call to eval() doesn't know
> about 'var d...'. And I am not passing any objets to it. I tried with
> program.callcontext.global - it does the same. What I need to do to make it
> working?
> 
> 

Please post the answer. You never know who will find this post in the future and thank you for it (even if you never hear about it).

Regan

On Fri, 24 Jun 2005 14:31:50 +0300, bobef <bobef@lessequal.com> wrote:
> After two days experimenting I was able to answer myself...
>
> bobef wrote:
>> Please consider this code:
>>  Program program = new Program;
>> RegisterDScriptTypes(program);
>> program.compile("test", "var d=new String('bobef');", null);
>> Value ret=*new Value;
>> Value[] args;
>> args.length=1;
>> args[0]=*new Value;
>> args[0].putVstring("Main(d);");
>> Dglobal_eval(null,program.callcontext,null,&ret,args);
>>   //end
>>  RegisterDScriptTypes() registers global function Main which displays a message
>> box. Executing the code above produces a message box saying "undefined" (main
>> works ok - tested). So it is clear to me that the call to eval() doesn't know
>> about 'var d...'. And I am not passing any objets to it. I tried with
>> program.callcontext.global - it does the same. What I need to do to make it
>> working?
>>

Yes you are right. I didn't post it because my solution is incomplete and inexact but perfect for my case. But I should have posted it anyway.
What I needed is the following thing. I extend the script as described in the dmdscript page and execute it. At some point I need to call a script callback function. So I do it trough eval(). It is perfect because I don't need to check if function exists or something else. So what I do is compile and execute as usual and at some point do this to call eval():

Value ret=*new Value;
Value[] args;
args.length=1;
args[0]=*new Value;
args[0].putVstring("something();");
Dglobal_eval(null,program.callcontext,program.callcontext.global,&ret,args);

In fact I don't pass program.callcontext.global but instead save the othis argument in one of my custom global functions. But I checkeed othis.toString() and it says Global something so I guess it is the same...
The important thing I found is that you can't call dglobal_eval before the program.execute() call, because it do some initialization there. Not all the initialization is done in program.compile(). I compared the source of dglobal_eval() and program.compile/execute and they do the same. I didn't need to go there but if you don't want to call execute I guess you should make your own function with the stuff from program.compile except the actual execution (which is ir.call or something like that if I understood correctly)...
So this is what I learned. Hopefuly it will be useful.

Regan Heath wrote:
> Please post the answer. You never know who will find this post in the  future and thank you for it (even if you never hear about it).
> 
> Regan
> 

Here is one more thing. There was a site called 'making dmd script suit your needs' which is no more. I was able to pull out this examle out of google's cache. It is exapmle of interaction between D and DMDScript. There is also slightly modified version of this example at walnut project's svn on dsource. I attach them both here.

bobef escribió:
> Here is one more thing. There was a site called 'making dmd script suit your needs' which is no more. I was able to pull out this examle out of google's cache. It is exapmle of interaction between D and DMDScript. There is also slightly modified version of this example at walnut project's svn on dsource. I attach them both here.
> 

Yeah, the site was mine. I think my uni's server is down. I don't think it's worthy to have it any longer because I think Walnut is more complete.

BTW, Walnut is kinda on hold now because I'm busy. If you (bobef) want to help a bit with it, it'd be truly appreciated.

-- 
Carlos Santander Bernal

Carlos Santander wrote:

> BTW, Walnut is kinda on hold now because I'm busy. If you (bobef) want to help a bit with it, it'd be truly appreciated.
> 

Well, sorry, I don't have any interest in Walnut because I think is too general and I don't think so 'wide' scripting capabilities are in any use because if you need to create full application you'd better use D. Don't get me wrong and please no offense because often my words sound offensive. Walnut is nice project but not in my sphere of interests.

Anyway I would support and help a project documenting the DMDScript extending/integration. In fact I am planning such project. After I finish my work with DMD (integration to akIDE, which I am currently working on) and respectively gather some more knowledge about it probably I will try to document it.

P.S. recently i amused myself by playing witg d/opengl/physics so this is kind of offtopic but I'm thinking about something like dmdscriptable physical simulation enviroment...

bobef escribió:
> 
> Well, sorry, I don't have any interest in Walnut because I think is too general and I don't think so 'wide' scripting capabilities are in any use because if you need to create full application you'd better use D. Don't get me wrong and please no offense because often my words sound offensive. Walnut is nice project but not in my sphere of interests.
> 

You don't need to worry. I was just looking for more hands. It's always worth a try.

-- 
Carlos Santander Bernal