fix: patricia trie decoding

[jstinhw]

Solve #40

1. Ensure that the leading shared nibble is retained with odd number of shared nibbles.
2. Involves adjusting the slicing mechanism to account for the original offset of keyNibbles.

[seunlanlege]

I didn't implement the ethereum trie verifier, so i can't comment on if this fix is correct or not. But i'll tag the original author @ripa1995 to get his thoughts

[ripa1995]

Sorry for late reply, it was a busy week.
Anyway, LGTM
Just one minor comment, I would move the test you've added from test/MerklePatricia.t.sol to integration-tests/src/merkle_patricia.rs.

@seunlanlege, looks like integration tests are not run as part of the GHA, is that intentional?

[seunlanlege]

Sorry for late reply, it was a busy week.

Anyway, LGTM

Just one minor comment, I would move the test you've added from test/MerklePatricia.t.sol to integration-tests/src/merkle_patricia.rs.

@seunlanlege, looks like integration tests are not run as part of the GHA, is that intentional?

Yeah they're broken atm on CI haven't had the time to bring them back.