Refactor WebAuthn Implementation

[nlordell]

This PR refactors the WebAuthn implementation into a library instead of split into multiple contracts. This removes the amount of CALLs required for verifying a signature.

I did some analysis on the deployment vs. signature verification costs comparing this code, main and this change incorporating optimizations from #289 and found this:

setup	deployment	verification (dummy)	break even
main (viaIR = false)	373543	17789	0
main (viaIR = true)	393494	16443	15
this branch (viaIR = false)	612123	13835	61
this branch (viaIR = true)	594058	12551	43
base64 optimizations (viaIR = false)	442545	12965	15
base64 optimizations (viaIR = true)	414941	11102	7

Meaning that after 7 signatures, even with the slightly larger deployment costs, we would break even by avoiding the extra CALL (noting that ABI-encoding the call to the WebAuthnVerifier is not particularly efficient in the first place).

[remedcu]

Should we add the bounty email before this is audited? Or adding the email doesn't signify anything until we explcitly say somewhere that these files are also available for bounty hunting?

Above and small nit, rest LGTM 👍🏾

[remedcu]

Subjective nit: I resonate more with isValid than success in this case.

A similar change in SignatureValidator as well.

[nlordell]

Should we add the bounty email before this is audited?

Nice catch, yes we should.

[coveralls]

Pull Request Test Coverage Report for Build 8328648516

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+62.4%) to 76.923%

---

Totals
Change from base Build 8328447414:	62.4%
Covered Lines:	131
Relevant Lines:	155

---

💛 - Coveralls