Skip to content

feat(census): add canonical sharded Workshop feature census - #27

Merged
Teakowa merged 5 commits into
mainfrom
feat/workshop-rs-19-census
Aug 18, 2026
Merged

Teakowa merged 5 commits into
mainfrom
feat/workshop-rs-19-census

Conversation

@Teakowa

@Teakowa Teakowa commented Aug 18, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Implements #19 as a stacked Draft PR on #18 PR #26.

  • Adds workshop_rs::census with deterministic shards derived from the canonical catalog, reviewed settings table, and workshop-rs-owned WIR capabilities.
  • Publishes a deterministic public CensusIdentity with schema version, SHA-256 definition digest, and sorted shard IDs; every CensusReport carries it.
  • Every case carries stable Define the Workshop conformance result contract and stable feature identities #18 feature IDs and preserves explicit matched, unsupported, known-gap, unexpected-regression, or inconclusive states.
  • Every Inconclusive/generated probe still executes the offline parse, WIR validation, emission, semantic round-trip, and locale gates first; lack of independent expectation is classified as Inconclusive only after those gates pass.
  • Generated catalog, settings, content, and localization fixtures are inputs only. Newly authored en-US/zh-CN localization fixtures remain Inconclusive without independent evidence.
  • Adds workshop-rs-cli census [--json] and documents the ownership/evidence boundary in ADR-0003.

Evidence boundary

The census does not expand Workshop semantics to make generated probes pass. It does not claim live-client or runtime correctness, and it does not hide offline probe failures as inconclusive: unexpected regressions make the CLI exit 1 while the report preserves all states. Focused parser/catalog/emitter tests remain unchanged and complementary.

Checks

  • cargo fmt --all --check
  • cargo clippy --workspace --all-targets -- -D warnings
  • cargo test --workspace --all-targets
  • cargo run -p workshop-rs --bin workshop-catalog-gen -- check
  • cargo run -p workshop-rs-cli -- census (report includes explicit regression and inconclusive states; exit reflects unexpected regressions)
  • git diff --check

This branch is based on PR #26 head 58832ff. Fixes #19.

@Teakowa

Teakowa commented Aug 18, 2026

Copy link
Copy Markdown
Contributor Author

Follow-up after independent QA verification failed on the initial head:

  • CENSUS_CAPABILITIES is now the WIR-owned registry for WIR shards.
  • Generated probes with no independent expectation remain inconclusive; the implementation no longer supplies its own expected artifact.
  • en-US and zh-CN localization cases now use committed independent source fixtures in both directions.
  • CensusReport::validate_against binds catalog identity, sorted shards, and result-to-shard attribution.
  • CLI JSON success and invalid-argument paths are covered.

Fix commit: 6a23b3e (Refs #19). Please perform a fresh independent QA pass against this head.

Derive deterministic feature shards from the canonical catalog, settings table, and Workshop IR capabilities; run independent semantic/locale gates where evidence exists, preserve inconclusive and unsupported states, and expose deterministic JSON export through the CLI. Fixes #19
@Teakowa
Teakowa force-pushed the feat/workshop-rs-19-census branch from d172df7 to 098a623 Compare August 18, 2026 07:45
@Teakowa

Teakowa commented Aug 18, 2026

Copy link
Copy Markdown
Contributor Author

Fresh independent QA — VERIFIED / PASS

Exact head reviewed: f9a520b9735b57ceb4766edd76432f8551eb5d9a.

  • Public CensusIdentity is deterministic, SHA-256 based, sorted by shard, and carried by CensusReport.
  • Newly authored localization fixtures have no independent references and remain Inconclusive.
  • Inconclusive probes execute parse, WIR validation, emit/reparse, and both locale-conversion gates before final classification.
  • Failures remain visible: census --json exits 1 with 298 unexpected-regression, 169 inconclusive, and 18 unsupported; no failures are hidden.
  • WIR capabilities remain workshop-rs-owned; no provider inventory or OPY/DEL semantics were introduced.
  • Local fmt, Clippy, workspace/all-targets tests, catalog check, and diff check passed.
  • Stable and Rust 1.85 CI passed on this exact head; worktree is clean.

The previous QA on 6a23b3e is superseded by this exact-head review. No findings for the requested contract/evidence scope; surfaced census regressions remain explicit follow-up evidence.

Base automatically changed from feat/workshop-rs-18-conformance to main August 18, 2026 08:38
@Teakowa
Teakowa marked this pull request as ready for review August 18, 2026 08:45
@Teakowa
Teakowa merged commit 9b7b7a5 into main Aug 18, 2026
2 checks passed
@Teakowa
Teakowa deleted the feat/workshop-rs-19-census branch August 18, 2026 08:45
@github-project-automation github-project-automation Bot moved this from Backlog to Done in Make Workshop Great Again Aug 18, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Development

Successfully merging this pull request may close these issues.

Build the canonical sharded Workshop feature census

1 participant