Replace SledBackend by ckb-rocksdb

[eval-exec]

What problem does this PR solve?

This PR want to load/store Headers from/into ckb-rocksdb

What is changed and how it works?

Proposal: xxx
Design doc: TODO

What's Changed:

Related changes

• Remove SledBackend
• Load/Store packed::HeaderView from ckb-rocksdb
• Create a new COLUMN: COLUMN_HEADER_INDEX to store number->hash relation for Headers
• in processing headers, use insert headers by 1 batch insert transaction instead of 2000x sequencial insert transaction

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code ci-runs-only: [ quick_checks,linters ]

Side effects

• Performance regression
• Breaking backward compatibility

Release note

Title Only: Include only the PR title in the release note.

[chainTe]

Security Review

Result: 3 confirmed finding(s)
Reviewed range: 2f0cbbdacb484b69b2bbc7575f0fa19bf3a278a2..95c62175132b678df7bde3e13b890b4d7c5dd9c9

Validation:

• Existing broad tests were not run locally; CI is expected to cover them after the branch is buildable.
• Runtime validation was blocked because shared/src/types/header_map/memory.rs contains unresolved conflict markers.
• Static source tracing found security-relevant sync regressions.

Notes:

• Mainnet MIN_CHAIN_WORK_500K is changed to zero, which disables the intended minimum-work gate before block download/assume-valid progression.
• COLUMN_HEADER_INDEX is keyed only by block number, but get_ancestor_internal(base, number) trusts it before proving the returned hash is an ancestor of base; valid side-branch headers can poison branch-specific ancestor lookups.
• Block fetching can now proceed from height-only/zero-work best-known headers because the previous total-difficulty superiority check is commented out.