Reject truncated RLP and trailing data when decoding

[sc0Vuai]

The decoder silently accepted malformed input:

• items shorter than their declared length: 0x83646f → "646f", 0xc3 → []
• a child longer than its parent list: 0xc2836400 → ["64"]
• extra bytes after the top-level item: 0x0102 → "01"

Each case now throws a RuntimeException, as ethers.js does. Every item's size is checked against the available data, and decode() rejects anything left after the top-level item.

Behavior change: input that used to decode silently now throws. That includes odd-length hex such as 0x123, which was padded and partly ignored. Correct RLP decodes the same as before.

Adds 19 regression vectors for truncated and trailing input.

🤖 Generated with Claude Code (https://claude.com/claude-code)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.27%. Comparing base (9d6e61b) to head (77ea176).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff              @@
##             master      #32      +/-   ##
============================================
+ Coverage     82.11%   84.27%   +2.15%     
- Complexity       55       59       +4     
============================================
  Files             3        3              
  Lines           151      159       +8     
============================================
+ Hits            124      134      +10     
+ Misses           27       25       -2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.