Load commands and gdblib explicitly in __init__.py

[gsingh93]

For historical reasons, the modules in pwndbg/commands and pwndbg/gdblib rely on executing code on import. This isn't great design, but we're stuck with it for now. At the very least, instead of importing a bunch of things in pwndbg/__init__.py that aren't actually used in that file, we import them in the specific __init__.py for that module inside a function, and then we just call that function to trigger the side effects of importing them.

The remaining modules in pwndbg/__init__.py may also be able to be removed in the future, but for now I'm just looking gdblib and commands.

Some motivations for this:

• This may make it easier to add unit tests, as we can hopefully just patch these two functions out
• It makes it clear in __init_.py what is actually required to be imported for the code in that file, as opposed to side effects
• If you ever want to know which files have side effects, you can look at these two functions to get the full list
• It's easier to figure out if an import is missing. For example, with this PR I added the pwndbg.gdblib.ctypes import, which was easy to spot with the reduced amount of imports.
• It makes refactoring easier in the future, as we just need to move around two function calls instead of a ton of imports
• This may make it easier to reload modules during development, but this is something I still need to investigate.

[gsingh93]

Not sure what the best way to deal with this is. The above imports are there so that we can rename pwndbg.gdblib.arch to arch_mod and export the pwndbg.gdblib.arch object as arch. Maybe the right way to do it is to make them globals, and then assign to those globals inside load_gdblib?