You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
End-of-day triage of the open queue in this repo (WXYC/wikidata-cache; GitHub redirects WXYC/wikidata-json-filter -> WXYC/wikidata-cache). Intended as the entry point for any contributor or agent picking up after the latest snapshot below: where to start, what blocks what, which issues are obsolete, and which proposals overlap.
Snapshot — 2026-06-08
Zero non-tracker open issues. Recommend treating this tracker as the index/landing page; any new triage findings should land as a comment here or as a freshly-filed issue.
WX-3.B consolidation landed (2026-06-08) — PR chore(deps): consolidate NUL-strip onto wxyc_etl::pg::to_pg_text_form #51 (chore(deps): consolidate NUL-strip onto wxyc_etl::pg::to_pg_text_form) merged 03:18Z. Bumps wxyc-etl pin to 0.8. Deletes src/wxyc_loader.rs::strip_pg_null_bytes + tests; delegates src/import.rs::escape_copy_text's NUL-strip arm to the upstream helper. Closes the wikidata-cache leg of WXYC/wxyc-etl#142 (closed 2026-06-09 03:22Z, org-wide sweep complete).
P0 sequencing — recommended order
None. Repo is at zero open non-tracker issues. The two near-term housekeeping items are open PRs (see below), not open issues.
Open PRs in flight
test: bump charset-torture corpus pin to v0.12.0 #29 — test: bump charset-torture corpus pin to v0.12.0. Opened 2026-05-05, 5+ weeks stale. All four checks green (lint, drift / drift-check, test, test-postgres). Drift guard now resolves since @wxyc/shared v0.12.0 is published. mergeable=UNKNOWN per the GH API — needs a rebase touch to refresh state, then it should be a clean merge. Not on any critical path; cheap to land.
Bump log from 0.4.29 to 0.4.32 #47 — Bump log from 0.4.29 to 0.4.32 (Dependabot, 2026-06-02). MERGEABLE / CLEAN, all checks green. Safe to merge.
Bump sha2 from 0.10.9 to 0.11.0 #49 — Bump sha2 from 0.10.9 to 0.11.0 (Dependabot, 2026-06-02). FAILING CI on lint/test/test-postgres: sha2 0.11.0 changes the digest output's GenericArray instantiation, breaking LowerHex formatting in tests/wxyc_identity_match_parity_test.rs. This is a semver-major upstream break, not a wikidata-cache bug. Do not merge as-is; either (a) hold the bump and let upstream consumers (wxyc-etl, hex formatting helper) adopt sha2 0.11 first, or (b) patch the test to use format!("{:x}", ...) via hex::encode on the byte slice. Either way, this is an upstream-coordinated decision — not a wikidata-cache-local task.
Cross-cluster conflicts (decide once before starting)
None — single open tracker, zero non-tracker open issues.
Obsolete / needs-revision (recommend close)
None.
Cross-repo dependencies (encoded both natively and in body prose)
Issue
Blocked by
Method
(none)
n/a
n/a
No active native blocked-by edges out of this repo; the previously-encoded edge on #41 (now closed) targeted WXYC/wxyc-etl#127 (also closed) — history-preserving and correct.
Cross-repo runtime risks (not actionable here; flagged for situational awareness)
WXYC/musicbrainz-cache#50 (bug, 2026-06-09) — schema::create_indexes uses naive split(";") parser; comments silently drop indexes. Sister-repo bug; wikidata-cache schema/migration shape is independent (we use sqlx-cli numbered migrations, not the split-on-semicolons helper). Cross-pollination risk if someone ports the MBC pattern here; today there's no shared code.
WXYC/musicbrainz-cache#42 — Consider regenerating MB cache (:5434) under wxyc-etl 0.2 to_match_form. Parallel to the wikidata-cache "regenerate after to_match_form contract changes" concern, but already-shipped here (wxyc-etl 0.8 in main; no equivalent open issue).
WXYC/library-metadata-lookup#216 — Wire up Wikidata leg of reconciliation pipeline (parent #207). LML downstream consumer for the wikidata-cache substrate; blocked by LML#211 (DiscogsReconciler trigram + name preprocessing). No work needed in wikidata-cache; the substrate is ready.
WXYC/library-metadata-lookup#212 — Railway CLI / Wikidata SPARQL: local network connectivity blocked from Rust HTTP clients. Local-dev connectivity issue with v6/dual-stack. Affects anyone running wikidata-cache locally if they need SPARQL fetches, but the production cron path (rebuild-cache.yml on GH-hosted runners) is unaffected.
.github/workflows/rebuild-cache.yml carries a TODO(runner-capacity) for the ~130GB Wikidata dump exceeding GitHub-hosted runner disk + 6h timeout. Smoke-test-only skeleton until runner-capacity is resolved (or until the workflow is moved to a self-hosted/larger runner). Acknowledged in source; not a triage finding.
Cluster map (0 non-tracker open)
n/a — repo is at zero open issues. Recent clusters (closed):
End-of-day triage of the open queue in this repo (WXYC/wikidata-cache; GitHub redirects WXYC/wikidata-json-filter -> WXYC/wikidata-cache). Intended as the entry point for any contributor or agent picking up after the latest snapshot below: where to start, what blocks what, which issues are obsolete, and which proposals overlap.
Snapshot — 2026-06-08
Zero non-tracker open issues. Recommend treating this tracker as the index/landing page; any new triage findings should land as a comment here or as a freshly-filed issue.
Switch PG to wxyc-postgres image; add F0000 precheck on 0003merged ~7h after this tracker was filed. Verified in main:migrations/0003_wxyc_identity_match_functions.sqlcarries the plpgsqlF0000wrapper,docker-compose.ymlpoints atghcr.io/wxyc/wxyc-postgres:pg16. Production fresh-DB scenario now safe against the previously-latent migration 0003 break.chore(deps): consolidate NUL-strip onto wxyc_etl::pg::to_pg_text_form) merged 03:18Z. Bumpswxyc-etlpin to0.8. Deletessrc/wxyc_loader.rs::strip_pg_null_bytes+ tests; delegatessrc/import.rs::escape_copy_text's NUL-strip arm to the upstream helper. Closes the wikidata-cache leg of WXYC/wxyc-etl#142 (closed 2026-06-09 03:22Z, org-wide sweep complete).P0 sequencing — recommended order
None. Repo is at zero open non-tracker issues. The two near-term housekeeping items are open PRs (see below), not open issues.
Open PRs in flight
test: bump charset-torture corpus pin to v0.12.0. Opened 2026-05-05, 5+ weeks stale. All four checks green (lint,drift / drift-check,test,test-postgres). Drift guard now resolves since@wxyc/sharedv0.12.0 is published. mergeable=UNKNOWN per the GH API — needs a rebase touch to refresh state, then it should be a clean merge. Not on any critical path; cheap to land.Bump log from 0.4.29 to 0.4.32(Dependabot, 2026-06-02). MERGEABLE / CLEAN, all checks green. Safe to merge.Bump serde_json from 1.0.149 to 1.0.150(Dependabot, 2026-06-02). MERGEABLE / CLEAN, all checks green. Safe to merge.Bump sha2 from 0.10.9 to 0.11.0(Dependabot, 2026-06-02). FAILING CI onlint/test/test-postgres: sha2 0.11.0 changes the digest output'sGenericArrayinstantiation, breakingLowerHexformatting intests/wxyc_identity_match_parity_test.rs. This is a semver-major upstream break, not a wikidata-cache bug. Do not merge as-is; either (a) hold the bump and let upstream consumers (wxyc-etl, hex formatting helper) adopt sha2 0.11 first, or (b) patch the test to useformat!("{:x}", ...)viahex::encodeon the byte slice. Either way, this is an upstream-coordinated decision — not a wikidata-cache-local task.Bump assert_cmd from 2.2.1 to 2.2.2(Dependabot, 2026-06-02). MERGEABLE / CLEAN, all checks green. Safe to merge.Critical / high severity
Cross-cluster conflicts (decide once before starting)
None — single open tracker, zero non-tracker open issues.
Obsolete / needs-revision (recommend close)
None.
Cross-repo dependencies (encoded both natively and in body prose)
No active native blocked-by edges out of this repo; the previously-encoded edge on #41 (now closed) targeted WXYC/wxyc-etl#127 (also closed) — history-preserving and correct.
Cross-repo runtime risks (not actionable here; flagged for situational awareness)
schema::create_indexes uses naive split(";") parser; comments silently drop indexes. Sister-repo bug; wikidata-cache schema/migration shape is independent (we use sqlx-cli numbered migrations, not the split-on-semicolons helper). Cross-pollination risk if someone ports the MBC pattern here; today there's no shared code.Consider regenerating MB cache (:5434) under wxyc-etl 0.2 to_match_form. Parallel to the wikidata-cache "regenerate after to_match_form contract changes" concern, but already-shipped here (wxyc-etl 0.8 in main; no equivalent open issue).Wire up Wikidata leg of reconciliation pipeline (parent #207). LML downstream consumer for the wikidata-cache substrate; blocked by LML#211 (DiscogsReconciler trigram + name preprocessing). No work needed in wikidata-cache; the substrate is ready.Railway CLI / Wikidata SPARQL: local network connectivity blocked from Rust HTTP clients. Local-dev connectivity issue with v6/dual-stack. Affects anyone running wikidata-cache locally if they need SPARQL fetches, but the production cron path (rebuild-cache.yml on GH-hosted runners) is unaffected..github/workflows/rebuild-cache.ymlcarries aTODO(runner-capacity)for the ~130GB Wikidata dump exceeding GitHub-hosted runner disk + 6h timeout. Smoke-test-only skeleton until runner-capacity is resolved (or until the workflow is moved to a self-hosted/larger runner). Acknowledged in source; not a triage finding.Cluster map (0 non-tracker open)
n/a — repo is at zero open issues. Recent clusters (closed):
How to find work
blockedBy) and history-preserving in body prose under each issue's## Relatedsection.What's already done (recent closures referenced in triage)