[ADDRESS] Manage function and struct addresses via enum

[TheRedDaemon]

With the test in #9, it seems there are likely cases where we need to use addresses more than once. To reduce the chance of error and also to have a single source of address-truth, it seems like a good idea to create one, two or multiple files with the enums to addresses we will use (structs and functions).
For now, these should be the ones of Vanilla 1.41.
The discussion on how to handle the ones for Extreme (switchable file, etc.) can be postponed.

[gynt]

Why enum and not just #define ? Naming things is hard.
I was thinking of:

#define ADDR_SHC_SHASUM_0xDEADBEEF 0xDEADBEEF // Obivously it would be a different value for other builds.

A "database" of addresses and how they relate to other binary builds is being developed here: https://github.com/sourcehold/FingerprintDB (I should give access or make it public...)

[TheRedDaemon]

Usually because of the principle hammered into my head:
If there is a way to express what you want in the language itself, use this.
The other way around would be: I use Macros if there is no other way to do it, like generating the function variants for the function resolver (in C++03).

So my mind goes about it this way:

• Enum is a language construct to express constant values.
• Addresses are all numbers.
• A namespaced anonymous enum does not clutter the global namespace. Class enums do not exist yet, so it would likely be:

  namespace FunctionAddress {
      enum {
          ...
      }
  }

• They are inlined in the usage positions.

So, I kinda have to ask: Why would you use Macros?

[gynt]

I don't mind too much actually. I just like having as much control over codegen as possible. But if we know the 2005 compiler will behave like we expect then it is fine.

For the record:
ChatGPT

Aspect	Macro (#define)	Enum Constant
Inlining guarantee	✅ Always	✅ Practically yes, but not formally
Codegen predictability	✅ Perfect	✅ Usually good
Type safety	❌ None	⚠️ Partial (int only)
Debugger symbol	❌ No	✅ Yes
Scoping / namespacing	❌ Global only	✅ Scoped via enum
64-bit address safe	✅ Yes	❌ Not guaranteed (int width issue)

💡 Recommendation (for MSVC++ 2005)

Since you’re explicitly requiring inlined values for addresses in a low-level DLL hook, you want:

No runtime lookup

No symbol table dependencies

Literal substitution

So:

✅ Use macros (#define) for address constants.