test(eth): pin the multi-byte chain_id EIP-1559 regression

[BitHighlander]

Firmware fix ed6db167 ("EIP-1559 chainId hashing wrong for multi-byte chain
IDs") shipped with no test at a chain id that reproduces it. Every EIP-1559
case in this file uses chain_id 1 or 3 — both single-byte — so the bug had zero
coverage in the file that tests the feature, and can silently return.

The bug: hash_rlp_field((uint8_t*)&chain_id, 1) fed only the least-significant
byte into keccak on little-endian ARM. Base (8453 = 0x2105) hashed 0x05, so the
signature recovered to an unrelated address with no funds. The RLP length was
computed correctly from the full value and the legacy EIP-155 path was always
right — only the EIP-1559 hash step was wrong.

A golden r/s would need a device run to produce, so this is a differential
instead. Sign one identical tx under two chain ids the BUGGY firmware cannot
distinguish:

8453 = 0x2105   low byte 0x05, two-byte value
4357 = 0x1105   low byte 0x05, two-byte value

Same low byte and same RLP length header, so the broken code hashes a
byte-identical pre-image for both. Signing is deterministic (RFC 6979), so
buggy firmware returns the same signature twice and the assertion fails.
Correct firmware hashes 0x21 0x05 vs 0x11 0x05 and they differ. No golden
value, no crypto deps — the repo has ecdsa but no keccak, so recovering the
signer address was not an option.

Version-gated to 7.15.0 so it SKIPs rather than fails on firmware predating
the fix.

Worth recording for whoever tests this by hand: Vault cannot reach this path at
all. It forces legacy transactions for chainId >= 256 (eip1559Ok = chainId <
256 in swap.ts), which is why ordinary Vault sends on Base were never broken.
The BEX builds type-2 with no such guard, so a dApp transaction through the
extension is the only manual route to it.

[BitHighlander]

Closing — wrong target. SOP for python-keepkey is to PR into the canonical branch of the single open PR into upstream master (currently keepkey:reconcile/upstream-sync), not the fork and not develop. I followed this clone's git config (origin = BitHighlander, origin/HEAD = develop) instead of the SOP.

Reopened with the full reconcile set as keepkey#212.