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
Deferred finding from the neutral triage of PR #954 (grep/t7-surfaces, LLP 0265 T7), triaged non-blocking at head 891f912b. One item, and it is sequenced work the maintainer has already decided, not an open question.
1. MEDIUM (integration, resolution decided): take #953 into grep/t7-surfaces and move the sidecarPathFor import
src/core/cache/maintenance.js:21 on grep/t7-surfaces imports sidecarPathFor from ../search/sidecar_build.js; PR #953 moves that function to src/core/search/searchable_columns.js with no re-export. The two branches merge with no git conflict and the merged tree then fails at module load (SyntaxError: The requested module '../search/sidecar_build.js' does not provide an export named 'sidecarPathFor'), taking down hyp cache status, hyp cache maintain, and the daemon maintenance tick. Reproduced, not inferred: see PR #954 review round 2 finding 1.
The maintainer answered the fork on the PR thread (option 1): #953 lands first, keeping the move with no re-export. Then this branch takes #953 and, as part of that update, applies the verified two-line fix:
It cannot be applied before #953 is in this branch's tree (this branch's searchable_columns.js does not export sidecarPathFor yet, so the switch alone would break the branch), and #953's head was still moving when this was triaged, so the work is deferred here rather than half-done now.
Rebase or merge it into grep/t7-surfaces and apply the import fix above as part of that update.
On the rebased head, re-run npm test, npm run typecheck, and npm run smoke -- query_grep_roundtrip to confirm the load failure is gone, rather than trusting the earlier merged-tree probe.
Re-confirm the surface text still matches the final SEARCHABLE_COLUMNS set. The applicable-now half is already done at 891f912b: the hypaware-query SKILL.md coverage bullet (both host copies) now states nine columns with tool_args in the not-searched list; the verb's coverage clause interpolates the constant so it follows Sidecar builds at maintenance: compaction finalizes a file, the index follows (LLP 0265 T6) #953 automatically; the query_grep_roundtrip smoke never seeds or asserts a tool_args hit (verified: its seeded schema has no tool_args column).
Backlink: PR #954 (triaged at 891f912b). The unstick decision is the maintainer comment on that thread; review records are rounds 1 and 2 there.
Deferred finding from the neutral triage of PR #954 (
grep/t7-surfaces, LLP 0265 T7), triaged non-blocking at head891f912b. One item, and it is sequenced work the maintainer has already decided, not an open question.1. MEDIUM (integration, resolution decided): take #953 into
grep/t7-surfacesand move thesidecarPathForimportsrc/core/cache/maintenance.js:21ongrep/t7-surfacesimportssidecarPathForfrom../search/sidecar_build.js; PR #953 moves that function tosrc/core/search/searchable_columns.jswith no re-export. The two branches merge with no git conflict and the merged tree then fails at module load (SyntaxError: The requested module '../search/sidecar_build.js' does not provide an export named 'sidecarPathFor'), taking downhyp cache status,hyp cache maintain, and the daemon maintenance tick. Reproduced, not inferred: see PR #954 review round 2 finding 1.The maintainer answered the fork on the PR thread (option 1): #953 lands first, keeping the move with no re-export. Then this branch takes #953 and, as part of that update, applies the verified two-line fix:
It cannot be applied before #953 is in this branch's tree (this branch's
searchable_columns.jsdoes not exportsidecarPathForyet, so the switch alone would break the branch), and #953's head was still moving when this was triaged, so the work is deferred here rather than half-done now.Steps, per the maintainer's instruction:
grep/t6-sidecar-build) to land.grep/t7-surfacesand apply the import fix above as part of that update.npm test,npm run typecheck, andnpm run smoke -- query_grep_roundtripto confirm the load failure is gone, rather than trusting the earlier merged-tree probe.SEARCHABLE_COLUMNSset. The applicable-now half is already done at891f912b: thehypaware-querySKILL.md coverage bullet (both host copies) now states nine columns withtool_argsin the not-searched list; the verb's coverage clause interpolates the constant so it follows Sidecar builds at maintenance: compaction finalizes a file, the index follows (LLP 0265 T6) #953 automatically; thequery_grep_roundtripsmoke never seeds or asserts atool_argshit (verified: its seeded schema has notool_argscolumn).Backlink: PR #954 (triaged at
891f912b). The unstick decision is the maintainer comment on that thread; review records are rounds 1 and 2 there.