Add FOLEAGE trace and AnyField OLE constructions

[ladnir]

Summary

• add trace-based binary OLE expansion for FOLEAGE and document the correlation layout and triple behavior
• add the AnyField OLE construction with F4, F9, and Goldilocks coefficient contexts
• support regular DPF and RevCuckoo setup independently of single-use and stationary expansion modes
• map regular-noise product regions by point multiplicity, expand RevCuckoo regions in parallel waves, and keep cached setup seeds independent of payloads
• add runtime secure parameter selection, exact construction cost accounting, bounded-memory expansion, and focused tests

#188 and #189 are merged. This PR now targets master.

AnyField setup and expansion tests use queued in-process socket continuations on the calling thread, preventing native-stack exhaustion during long synchronous protocol exchanges under GCC/ASan. Existing correlation and repeated-expansion checks remain in place.

Validation

• The full stack passed 318 Windows tests, with 24 disabled-configuration skips, including the previously failing Silent OT, AnyField F2 OLE, and RevCuckoo robustness cases.
• Expanded OT and VOLE size sweeps passed for padded requests and stationary noise.
• After the dependency fix, GCC 13/ASan passed 33 focused curve, base-OT, RingLPN, RevCuckoo, and AnyField tests (3 configuration skips). The 38,448-OT SimplestOT reproducer now has zero mismatches.
• Windows passed the added curve regression, existing hash-to-curve tests, and RingLPN base generation after the dependency update.
• The broader local GCC build has SoftSpoken disabled, so five VOLE tests requiring that backend cannot run successfully in that configuration. Silent OT is also disabled locally; its padding sweeps passed on Windows and in GitHub Ubuntu CI.
• Ubuntu/ASan and macOS CI passed on the preceding heads. Fresh CI is running for the final dependency update.

Final merge check

One Ubuntu/ASan run on 996b9ff crashed in CoeffCtxInteger::mask<FoleageF4x243> during FOLEAGE base generation; the duplicate run passed. The final fix gives the FOLEAGE coefficient context a direct, unrolled four-block mask. It removes the generic mask's temporary SIMD array from the ternary DPF coroutine without adding allocations or callback overhead.

The focused GCC 13/ASan runner enables FOLEAGE and ternary DPF and passes audit, tensor, F4 OLE, F2 OLE, triple, and 12 real-base-OT generation repetitions. Its local build lacks SoftSpoken, so these repetitions use Base mode; All six Windows FOLEAGE test groups passed, including base generation in both Base and BaseExtend modes. Fresh GitHub CI is still running on the final mask fix.