Use B01 with regional reroute fallback in Pipeline9

[seveibar]

What changed

• Replaces Pipeline9's ordinary high-density stage with an immutable-first B01 adapter.
• Loads pre-routed copper into B01 as layer-aware route obstacles.
• When B01 fails, clips every pre-routed segment that enters that capacity region, makes those segments ordinary port pairs, and solves the combined region with the existing regular high-density portfolio.
• Splices only the regional replacements back into the original pre-routed traces, preserving PCB trace ids and copper outside the failed region.
• Treats an output trace with an existing pcb_trace_id as an intentional replacement during joint DRC.
• Lets same-net via simplification reuse an immutable pre-routed via without mutating or re-emitting that via.
• Keeps newly routed PCB trace ids distinct from preloaded ids.

Why

B01 should get the first opportunity to route around immutable pre-routed copper. Some valid SRJ23 regions still defeat B01 today, so Pipeline9 needs an explicit last-resort strategy that reroutes all copper inside only the failed region instead of failing the whole pipeline or silently ignoring preloaded-route intersections.

Validation

• bunx tsc --noEmit
• 12 focused tests across Pipeline9 fallback, joint DRC replacement semantics, immutable-route simplification, and same-net via reuse
• SRJ23 sample 10: solved with 2 fallback regions, 2 mutated preloaded traces, 0 joint relaxed-DRC errors
• SRJ23 representative visual snapshots (1, 10, 23): pass; pre-routed traces are included
• Full bun test was sampled through the long bugreport suite; change-specific tests remained green, while unrelated pre-existing 5-second timeouts and bugreport SVG mismatches occurred

Stack

This draft targets agent/joint-drc / #1807 because replacement semantics require DRC to evaluate preloaded and routed copper jointly. Retarget to main after #1807 merges.

The exact B01 failure input is isolated separately in high-density-b01#6.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
capacity-node-autorouter	Ready	Preview	Jul 31, 2026 6:36am

[tscircuitbot]

Benchmark This PR

Run benchmarks by commenting on this PR:

/benchmark [benchmark.sh args...]
/benchmark-long [benchmark.sh args...]
/benchmark-all

Comment /benchmark to run the default dataset, or append any arguments accepted by ./benchmark.sh.
Comment /benchmark-long for an 8-vCPU run that defaults to 8 workers and has an eight-hour timeout.
Comment exactly /benchmark-all to start separate workflow runs and result comments for the default dataset plus srj18, srj19, srj20, srj21, and srj23.

Everything after /benchmark or /benchmark-long is safely forwarded to ./benchmark.sh, except --profile-solvers, which enables profile comparison tables.
Examples: /benchmark --dataset 18 --sample-timeout 2000s, /benchmark --pipeline 7 --scenario-limit 20, /benchmark all 20 --concurrency 8, and /benchmark-long --dataset 18.

Use /update-snapshots (or /us) to run BUN_UPDATE_SNAPSHOTS=1 bun test --timeout 120_000 on the PR branch and auto-commit snapshot updates.
Use /usf to read recent failed test files, update and verify their exact CI test shards, and auto-commit only their snapshots. It uses the configured fast benchmark runner by default; use /usf --ubuntu-latest for GitHub-hosted x64 CI parity.

Any PR whose title contains [BENCHMARK TEST] will automatically run one default-dataset benchmark on PR updates; it does not post a PR result comment.