👷 CI filter unrelevant tests from gas snapshot

[qd-qd]

No description provided.

[github-actions]

Changes to gas cost

Generated at commit: 650ea27b5cca82c824275b6137bb750fef6697a9, compared to commit: d940f6efadc8772f6699fa56eca6e52cc51f00d3

🧾 Summary (20% most significant diffs)

Contract	Method	Avg (+/-)	%
WebAuthnImplementation	_generateMessage	+1,714 ❌	+48.82%

---

Full diff report 👇

Contract	Deployment Cost (+/-)	Method	Min (+/-)	%	Avg (+/-)	%	Median (+/-)	%	Max (+/-)	%	# Calls (+/-)
WebAuthnImplementation	292,131 (0)	_generateMessage	5,194 (+4,112)	+380.04%	5,225 (+1,714)	+48.82%	5,194 (+872)	+20.18%	5,385 (0)	0.00%	6 (-8)

[qd-qd]

⚠️ Do not trust the gas diff report in this PR ⚠️

Gas consumption didn't change, we only blacklisted non-relevant tests that reverted. As those tests consumed less gas than the successful ones (as the execution is stopped before the end of the computation), the average consumption seems to raise. The logic didn't change, but the previous gas snapshots were misleading. This PR fixes that.