On 31/12/15 11:20 PM, Gan wrote:
I've written an AI to play a game in Objective-C on Mac. The AI bits are
mostly C++ code for performance. I want to make the AI cross platform
because the game has a huge windows user base. My hope is to get other
programmers interested in helping out because it is very time consuming
to make perfect screen detection and AI decisions.
My question is, to write the Mac and Windows specific code in their
respective languages(C# and Obj-C) for UI, screen capture and input
simulation; can I have that code be usable from D? I don't know much
about shared libraries or importing DLLs so I have no knowledge on the
possibility. I would like to write all the AI code in D but have all the
OS specific code written in their own languages that I can call from D.
I am very familiar with writing C#, Obj-C and even D code. I just don't
know how to make them all play nice. Maybe there's a way to compile the
OS specific code in libraries and D can call them depending on which OS
the D code gets compiled to?
Would anyone know how to do this? Or have basic examples?
To start I'd just like to write a D program that can call some specific
already compiled C# or Obj-C code that will create a UI window but
maintain full control in D such that D could issue further already
compiled commands. Also receive input, like a button is clicked and that
button sends an event notifier back to D.
Is this plausible?
I'll start out by saying, this is more appropriate for D.learn not the
main D Newsgroup.
What you have here is two separate issues.
- Interacting with Objective-C which over the next few years will become
very easy (WIP currently). Otherwise you can interact with it via extern(C).
- Interacting with C#. This will be a good deal harder, but still
do-able with the help of extern(C). You'll need to get good at
interacting with unmanaged code from there, which I cannot help with
since it is C# specific.
One thing to remember D is a native language meaning an exe cannot be
ran on Linux and vice versa.
It won't be as simple as compiling as a shared library once and using it
everywhere.
