Outcome
Demonstrate that the toolkit safely owns the complete supported authored corpus and produces identical results from clean and incremental builds.
Scope and invariants
- Pin a full content revision and resource/sound inputs; enumerate formats/files/records/references/licenses and expected exclusions.
- Run parse/catalog/validate/compile/package from clean state and from randomized incremental edit/revert sequences.
- Compare artifact bytes/digests, diagnostics, source mappings, caches, and consumer results across supported platforms.
- Exercise large/adversarial files, cache corruption/eviction, concurrent read/build, cancellation, external changes, disk/permission failures, and recovery.
- Publish machine-readable coverage, timing/memory/cache, diagnostic, and provenance reports.
Acceptance criteria
- All supported corpus files are owned by a format/schema/workflow; zero silently skipped paths.
- Clean and equivalent incremental builds produce identical diagnostics/artifacts/manifests.
- Go server and Rust editor/renderer adapters pass the released corpus conformance package.
- No old Python tool/parser is required for production content build/validation.
- Exceptions are explicit issues with owner/milestone and block affected cutover scope.
Dependencies and parallelization
Requires M2 toolkit and converted workflow groups; can expand continuously as content migrations land.
Licensing
Toolkit code is MIT. Original past contributions by a person listed in the merged approved MIT provenance-grantor registry in atrinik/atrinik#275 may be copied, migrated, translated, or relicensed under MIT only after a complete, non-shallow Git-history audit, including renames and moves, proves the selected material is that grantor's solely authored original work and contains no embedded third-party or conflicting-licensed material. Record the exact evidence, applicable grantor, and destination. Mixed authored data and fixtures retain their actual licenses.
Program equivalence output
Export whole-corpus ownership, conversion, determinism, diagnostic, consumer-compatibility, timing/bounds, and license/provenance results using the join keys defined by atrinik#279. This issue is a native blocker of atrinik#280; zero silently skipped paths and an immutable content/toolkit/resource coordinate set are required before the aggregate gate runs.
Outcome
Demonstrate that the toolkit safely owns the complete supported authored corpus and produces identical results from clean and incremental builds.
Scope and invariants
Acceptance criteria
Dependencies and parallelization
Requires M2 toolkit and converted workflow groups; can expand continuously as content migrations land.
Licensing
Toolkit code is MIT. Original past contributions by a person listed in the merged approved MIT provenance-grantor registry in atrinik/atrinik#275 may be copied, migrated, translated, or relicensed under MIT only after a complete, non-shallow Git-history audit, including renames and moves, proves the selected material is that grantor's solely authored original work and contains no embedded third-party or conflicting-licensed material. Record the exact evidence, applicable grantor, and destination. Mixed authored data and fixtures retain their actual licenses.
Program equivalence output
Export whole-corpus ownership, conversion, determinism, diagnostic, consumer-compatibility, timing/bounds, and license/provenance results using the join keys defined by atrinik#279. This issue is a native blocker of atrinik#280; zero silently skipped paths and an immutable content/toolkit/resource coordinate set are required before the aggregate gate runs.