chore: clean up unused code

[cpb8010]

Description

Removed the following:

• R1 validation and keys were unreachable after changing the init interface
• K1 validation was unused since it was validated directly without a module (simple signature instead of modular signature)
• Fallback modules were only implemented in storage, not in the account.
• The general module list wasn't used except for a general purpose auth check
• Exec modules could have been used, but haven't been tested and aren't supported by the factory to be installed at init.
• Call disable during validation module removal (matching hook removal)

Combined the following:

• IModule and IModuleValidator (duplicated interfaces)
• Init and addHook (duplicated)

Additional context

Less dead code means a faster audit, even if this code was mostly audited before it was incomplete in some places. Most of this was added in the direction 7579 compatibility or was leftover from Clave after the validation interface was made into a module.

This doesn't functionally change the factory interface, so we'll don't yet need to update the SDK

[cpb8010]

I realize this interacts more with the validation interface cleanup than I expected. I'm going to try to reformat these so the cleanup is more of a setup for removing the validation hook interface (which is what's really required to drop lots of this stuff)

[cpb8010]

I got some good feedback on this outside of this PR, so I'll post my responses here for better visibility:

• My previous remark about validation is because I'm entirely reworking the validation module interface to include the functionality of the validation hook (in this PR: Session Validation Module Interface Update #208. I have no intentions of adding functionally empty modules back.
• OnlySelfOrModule is poorly named! There's already an onlySelf modifier, even more to consolidate!
• disable() in WebAuthValidator should satisfy ERC165, but we don't actually have enough context to disable the module with just the sender address. I can make a note to either reject this or create a new storage layout so we can do this lookup. The code comment should include this context.
• The SsoAccount.sol initialize now calls _installHook instead of routing this via the generic module's init function. This removes the need for modules to have permissions to add other modules or hooks. This is part of the setup for the next PR that entirely removes validation hooks in favor of a single validation module interface.
• The tests pass the session key contract as a hook and as a validation module, so it only needs to be initialized once. We probably want to allow future hooks to have their init methods called, so can add that back so it at least gets called when doing the externally facing addHook

[ly0va]

looks good, but needs some fixes

currently if we want to uninstall the session key module, we have to uninstall it as both validator and as hook. if we do it in the wrong order, it will lock itself in a bad state (since we only check for isHook in isInitialized). So we have to fix isInitialized or even skip that check in uninstall entirely.

[ly0va]

We probably want to allow future hooks to have their init methods called, so can add that back so it at least gets called when doing the externally facing addHook

yes we should probably let hooks get inited, but with our current interface adding more hooks is problematic, especially if they expect hookData, since we have to properly pass it for every call and we might not know what data they expect.

so it should be fine for now, we might even consider disabling adding hooks temporarily, after initialization