Evidence

Every item below is tied to a named source state, corpus, environment, or test budget. Read its method and limitations before carrying an observation to a different machine or workload.

1. C and Python coverage baseline Informational line, function, and branch coverage from the complete local test suites. Coverage report · Informational baseline
2. CM13 parser fuzz smoke Reproducible libFuzzer smoke coverage for the canonical CM13 parser. Test report · Current evidence
3. Native and Z3 solver comparison smoke baseline Seven-sample native and Z3 baseline on the smallest shared SAT and UNSAT inputs. Benchmark report · Recorded evidence
4. Optimized solver MRV index Evidence for the optimized solver's private row-major MRV bucket index. Benchmark report · Accepted mechanism
5. Optimized solver SAT ownership transfer Evidence for transferring the verified SAT domain buffer without a redundant final copy. Benchmark report · Accepted mechanism
6. Optimized solver byte-wise support tables Evidence for aggregating Wang propagation support by domain byte. Benchmark report · Accepted mechanism
7. Optimized solver dynamic DFS stack Evidence for the optimized solver's geometrically growing DFS stack. Benchmark report · Accepted mechanism
8. Optimized solver initial-trail removal Evidence for omitting rollback entries during non-rollbackable initial propagation. Benchmark report · Accepted mechanism
9. Optimized solver queue and trail profile Queue, trail, Callgrind, and Cachegrind evidence used to select queue deduplication. Profiling report · Recorded evidence
10. Optimized solver queue deduplication Evidence for the optimized solver's packed pending-cell index. Benchmark report · Accepted mechanism
11. Serial solver reference profile Reproducible reference measurements for the serial Wang solver. Benchmark report · Recorded baseline