Currently ceval.c contains the interpreter and a lot of helper functions.

We should factor the interpreter into a separate file from everything else, because:

• The interpreter is special. It is more of a template for various different interpreters than a normal function, having it in its own file makes it a lot easier to follow the various macros and #if conditions
• It means that all helpers must have symbols (not be static) helping any JITs using PEP 523, such as PyTorch and CinderX
• Having the interpreter in its own file makes it easier to build it differently for JIT, free-threading and other custom builds that might need to modify the interpreter, but don't want to have to mess with the rest of the VM
• ceval.c is big and breaking it up makes it a bit easier to read the code