hardening(index_block): validate shared prefix against restart-key span

[polaz]

Context

Found during review of #182.

Problem

parse_truncated in src/table/index_block/block_handle.rs validates shared_prefix_len only against the current entry boundary (prefix_end <= offset). For restart intervals with multiple truncated entries, this can still allow a shared prefix that extends beyond the restart key bytes.

Suggested approach

• carry restart-key span (start + len) into truncated parse path
• reject shared_prefix_len > restart_key_len before constructing SliceIndexes
• add corruption regression test for malformed shared_prefix_len

Files involved

• src/table/index_block/block_handle.rs
• src/table/block/decoder.rs
• parser tests in src/table/index_block/block_handle.rs

Estimate

4h (implementation 2.5h, tests 1h, review/cleanup 0.5h)

Links

• PR: perf(table): support compressed index blocks #182

[coderabbitai]

⚠️ Possible Duplicate Issue(s)

• fix(index-block): validate shared prefix against restart-key span #183

🔗 Related PRs

#139 - fix: thread UserComparator through ingestion guards and range overlap [merged]
#146 - perf(compaction): merge input ranges before L2 overlap query [merged]
#163 - feat(config): per-level Fs routing for tiered storage [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!