GAL-2 Y2038 Proof Bundle v0.3: Same-Binary Raw Edge vs API-Derived Continuity

This proof bundle documents a reproducible Y2038 boundary experiment using the same tested legacy-style binary in three modes: 1. Raw edge path 2. GAL-2 API-derived literal edge path 3. GAL-2 API-derived continuity path Results: Raw edge path: - SUMMARY order_fail=1 ttl_fail=9 sched_fail=9 GAL-2 API-derived literal edge path: - SUMMARY order_fail=1 ttl_fail=9 sched_fail=9 GAL-2 API-derived continuity path: - SUMMARY order_fail=0 ttl_fail=0 sched_fail=0 Interpretation: The same legacy-style binary collapses when driven through the raw 32-bit Y2038 edge path. The same binary also collapses when the literal Y2038 edge timeline is derived from the GAL-2 API and delivered directly to the 32-bit consumer. The same binary preserves event ordering, TTL validity, and scheduler coherence when time consumption is mediated through a GAL-2 API-derived continuity path. Defensible claim: For the tested same legacy-style binary, both the raw Y2038 edge path and the literal API-derived edge path collapse, while the GAL-2 API-derived continuity path preserves operational continuity. Scope note: This evidence is workload-scoped, not universal. However, it shows a critical distinction: GAL-2 does not eliminate the intrinsic 32-bit overflow of a literal Y2038 edge presented directly to a legacy consumer. Instead, GAL-2 can preserve operational continuity by mediating consumed time through an API-derived continuity path.