Enforce zero-intersection geometric routing

[seveibar]

Summary

• validate greedy final hops with the same hard constraints as normal routing
• track actual segment endpoint coordinates instead of relying only on boundary-angle ordering
• add REQUIRE_ZERO_INTERSECTIONS for solves that must rerip every same-layer or cross-layer intersection
• accept immutable fixedRegionSegments so routes can avoid pre-routed fanout copper
• selectively rerip routes that cross another route or fixed geometry
• preserve the best partial state for bounded high-density solves

Why

The clad1 RP2040 route exposed three gaps in the existing intersection model. Interior ports can make angular interval ordering disagree with physical segment geometry, pre-routed fanout traces were invisible to the hypergraph cost model, and the timeout/greedy path could accept a crossing final hop. Together those paths could report a nominally routed graph while emitting intersecting copper.

The new strict mode uses real coordinates, includes immutable incoming geometry, and refuses to finish until both same-layer and crossing-layer intersection counts are zero.

Validation

• bun run typecheck
• 28 focused solver, selective-rerip, bus-routing, and region-path tests pass
• new regressions cover interior-port geometric crossings, fixed geometry surviving rerips, strict zero-intersection acceptance, and reripping real intersections
• clad1 RP2040 integration reaches 117/117 graph connections with zero reported geometric intersections

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
tiny-hypergraph	Ready	Preview, Comment	Jul 28, 2026 8:17am

[tscircuitbot]

Benchmark This PR

Run benchmarks by commenting on this PR:

/benchmark [benchmark.sh args...]
/benchmark --dataset <hg07|srj18> --limit <n> --solver <core|poly> --families <default|default+deep|all> --concurrency <n|auto>

Everything after /benchmark is forwarded directly to ./benchmark.sh.
The result comment shows the matching dataset from the latest stored main-branch benchmark artifact beside the PR run; main is not rerun for each PR benchmark request.

Examples:

• /benchmark -> hg07, all samples, core solver
• /benchmark --dataset srj18 --limit 20 -> first 20 srj18 samples
• /benchmark --dataset srj18 --concurrency auto -> srj18 samples with runner CPU-count concurrency
• /benchmark --sample sample002 -> one hg07 sample
• /benchmark --limit 20 --solver poly -> first 20 hg07 samples with the poly solver
• /benchmark --limit 40 --families default+deep -> first 40 hg07 samples with deeper candidate families

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