[3.12] orbital: Added custom entry in PyInterpreterState

Hijacked the interpreter state to get my boys on the inside. Currently,
there doesn't seem to be an officially supported way to get a piece of
pre-initialized per-interpreter memory over to an embedded module (more
specifically, the embedded module's functions).
For example:
    [init'd mem python#1] ----> [subinterp python#1] ----> {module fn call}
    [init'd mem python#2] ----> [subinterp python#2] ----> {module fn call}
{module fn call} has a single implementation with access to its module
state via `PyModule_GetState`. The initialization of a subinterpreter
populates a custom entry for the pre-initialized memory in its interpreter
state. On initialization of the embedded module (`Py_mod_exec`), the
module state is populated with the custom entry
(`PyThreadState_Get()->interp`). The module function now has access to
the pre-initialized memory via its module state.

Author: bytemarx

Include/internal/pycore_interp.h

@@ -103,6 +103,11 @@ struct _is {
         size_t stacksize;
     } threads;
 
+
+    // ORBITAL MODIFICATIONS (hey guys, don't mind me)
+    const void* orb_passthough;
+
+
     /* Reference to the _PyRuntime global variable. This field exists
        to not have to pass runtime in addition to tstate to a function.
        Get runtime from tstate: tstate->interp->runtime. */