fix(ol): select valid canonical epoch commitments

[voidash]

Summary

This PR fixes OL canonical block and epoch commitment selection so callers prefer terminal blocks with BlockStatus::Valid instead of blindly taking the first stored commitment.

It also includes the previously validated checkpoint DA/inbox-index fixes needed for the real bridge e2e path.

Bug Context

During the real bridge e2e run with staging-v2-like SP1 proving, the bridge deposit path succeeded, but the EE update path previously failed because the OL tracker/checkpoint prover could resolve an epoch terminal block from an invalid fork. That produced failures like No OL state found for terminal block ... even though the valid canonical block at the same slot existed.

The root cause is that epoch summaries/commitments can contain multiple candidates for the same epoch/slot, and the old lookup path selected the first stored entry. The first entry is not guaranteed to be valid.

Fix

• Prefer BlockStatus::Valid in OL canonical block lookup.
• Prefer valid terminal blocks when RPC/prover code resolves canonical epoch commitments.
• Re-check the valid canonical epoch commitment after checkpoint proof completion before advancing the checkpoint proof cursor.
• Keep fallback behavior to the first stored entry when no valid block is available.

Verification

• cargo fmt --check
• cargo check -F sequencer -F strata/sp1 -F debug-utils -F test-mode -F debug-asm --bin strata --bin alpen-client --bin strata-dbtool
• Local real bridge e2e in SP1 reserved-network mode: test_real_bridge_deposit_withdraw passed in datadir functional-tests/_dd/12-1-tbtyw. Deposits credited OL/EE, SAU landed, and withdrawal settled on regtest.

[github-actions]

Commit: 1024c50

SP1 Execution Results

program	cycles	success
EVM EE Chunk	553,055	✅
EVM EE Account	1,271,238	✅
Checkpoint	2,203,301	✅

[codecov]

Codecov Report

❌ Patch coverage is 89.23077% with 7 lines in your changes missing coverage. Please review.
✅ Project coverage is 78.88%. Comparing base (b819de3) to head (6d525a8).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/storage/src/managers/ol.rs	58.33%	5 Missing ⚠️
crates/ol/checkpoint/src/context.rs	95.23%	2 Missing ⚠️

@@            Coverage Diff             @@
##             main    #1806      +/-   ##
==========================================
- Coverage   78.95%   78.88%   -0.07%     
==========================================
  Files         655      655              
  Lines       71268    71328      +60     
==========================================
- Hits        56268    56266       -2     
- Misses      15000    15062      +62     

Flag	Coverage Δ
functional	58.13% <87.69%> (-0.28%)	⬇️
unit	64.94% <21.53%> (-0.07%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
bin/strata/src/rpc/provider.rs	73.91% <ø> (ø)
...s/alpen-ee/sequencer/src/ol_chain_tracker/state.rs	98.74% <100.00%> (-0.01%)	⬇️
crates/ol/checkpoint/src/context.rs	72.50% <95.23%> (+3.67%)	⬆️
crates/storage/src/managers/ol.rs	93.83% <58.33%> (-3.68%)	⬇️

... and 5 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6d525a8d43

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[delbonis]

nit, but significant

[voidash]

The checkpoint DA/preseal state-root mismatch fix belongs to STR-3412, which Bibek is already tracking separately. I’m going to remove that part from this PR so #1806 stays focused on valid canonical epoch commitment selection.