[Dlang-internal] A Proof of Concept for the baremetal D runtime plans

[Mike Franklin]

On Friday, 3 May 2019 at 23:20:13 UTC, mbalantz wrote:

> * The linker errors complain about undefined symbols and 
> mention alphanumeric strings as reference, containing the SDL 
> binding for D keywords and SDL keywords as per the functions 
> that used them in the pongd.d file.

I'm sorry, but I don't fully understand.  Please copy and paste a 
few of the errors here.

> * I downloaded and installed the dmd compiler on x86_64 Windows 
> 10 Education on an AMD Ryzen 1950X. Therefore:

I think DMD is fine if you want to do bare-metal programming on 
x86_64.

> * ..it is not truly baremetal. I intended to share it to show a 
> demo, and then maybe we could look into reverse engineering it 
> to be baremetal. Though I don't know if I picked a good demo 
> for running baremetal (i.e. a pong game).

What I'm able to ascertain from this is that you've ported a pong 
game to SDL, not to bare metal.  If SDL were ported to bare 
metal, then it would probably work because SDL abstracts the 
architecture and platform for you.

> * Yes, it is quite dependent on SDL. Porting SDL to D would be 
> no small task, I would think? I tried using the bindbc bindings 
> provided in D's library to provide the SDL. Though that is not 
> sufficient? The linker errors complain of undefined symbols for 
> a lot of SDL functions.

Yes, porting SDL to D would be a big task.  I don't think it's 
necessary though because D has good support for binding to C 
APIs.  You can try compiling with -betterC to see if that gets 
rid of some of the linker errors you're encountering.  Otherwise, 
I'd need to see the linker output to understand what's going on.

Mike