Keep trace-layer candidate scoring consistent across repair stages

[seveibar]

The branch portfolio can discard a trace-layer repair that its inner solver accepted. After enabling inclusive via-span checks, SRJ18 sample 12 produces a trace-clean candidate with a remaining via-to-pad issue, but the outer acceptance guard counts that deferred issue as a hard via regression and restores the trace collisions. The same-machine autorouter benchmark changes from DRC pass to five reported errors.

Use the same staged scoring at both levels: exclude via-to-pad issues from the hard via regression count and pass the complete input snapshot to the existing candidate comparator. Remaining via-to-pad errors stay in the snapshot and are passed to the dedicated repair phase; via-to-trace and via-pair protections remain in place.

Add a two-route regression that fails on main. It checks acceptance of the deferred via-pad case, rejection of a new via-trace collision, visibility of the remaining error, and unchanged input routes.

Validation: all 89 tests pass (595 assertions), typecheck passes, and existing snapshots are unchanged. The consumer integration and full SRJ18 regression are in tscircuit/tscircuit-autorouter#2402; final same-machine benchmarks on consumer commit 239d57f restore sample 12 to zero DRC errors, with no pass/fail outcome regressions across dataset01 or SRJ18. The consumer’s full CI also passes.

[tscircuitbot]

Benchmark This PR

Run benchmarks by commenting on this PR:

/benchmark [benchmark.sh args...]
/benchmark [scenario-limit|all] --dataset <drc14|srj18> --concurrency <n> --effort <n> --max-iterations <n>

Everything after /benchmark is forwarded directly to ./benchmark.sh.

Examples:

• /benchmark -> run default SRJ18 benchmark settings
• /benchmark 10 -> run first 10 SRJ18 samples
• /benchmark --dataset srj18 -> run all SRJ18 samples
• /benchmark 1 --dataset srj18 -> run the first SRJ18 sample
• /benchmark --dataset drc14 -> run all DRC14 samples
• /benchmark all --concurrency 4 -> run all SRJ18 samples with 4 workers
• /benchmark --scenario-limit all --effort 2 -> run all samples with effort 2
• /benchmark --scenario-limit 20 --max-iterations 100 -> run 20 samples with a max iteration override

Any PR whose title contains [BENCHMARK TEST] will automatically run the benchmark workflow on PR updates.