Make every service function shard-aware, without cutting over shard assignment yet - #125
Merged
brylie merged 2 commits intoAug 30, 2026
Conversation
…hard assignment yet (#120) Every service function that used to call resolveWorkspaceContext() bare now resolves its actual target shard via new catalog primitives (resolveShardForParent/resolveShardForRecord), falling back to the default context when untracked. createCollection still assigns shardId: 'default' deliberately — this proves the resolution mechanism correct for a genuinely separate shard (tests manually construct one, same pattern as the holds eviction-wiring fix) without changing where content actually lives yet, so production behavior is unchanged and no client/attach-ws changes are needed in this slice. - catalog.ts: reserveRecordLocator/releaseRecordLocator (row-level, closes the gap where write_record/delete_record/hold_records only ever receive a bare recordId), resolveShardForParent/resolveShardForRecord. - permissions.ts: resolveParentWorkspaceContext/resolveRecordWorkspaceContext/ groupRecordIdsByShard — shared resolution helpers; requireAccessibleRecord is now itself shard-aware, which every existing caller already goes through. - records.ts: createRecord reserves a row locator when its parent is a Collection; writeRecord/deleteRecord/getRecord resolve via the record's own locator. - collections.ts: queryCollection/updateCollectionTitle/deleteCollection resolve the collection's real shard. - holds.ts: hold_records/release_records group recordIds by resolved shard and operate against each shard's own Awareness (a cross-document agent batch is a stated acceptance criterion — see collaboration.md). - search.ts: Collections are enumerated via the catalog first (resolving each one's real shard, including its own meta entry — not just its rows), with a fallback pass over the default doc for uncataloged (direct-Yjs- written) Collections the catalog loop can't see. No MCP tool schema changes needed — every tool already carries enough of an id for server-side shard resolution. 661/661 tests passing (13 new). Refs #120. Branched off feat/workspace-catalog-113-phase-a (PR #119, not yet merged) since this depends on its catalog.ts.
Qodo reviews are paused for this user.Troubleshooting steps vary by plan Learn more → On a Teams plan? Using GitHub Enterprise Server, GitLab Self-Managed, or Bitbucket Data Center? |
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Comment |
brylie
added a commit
that referenced
this pull request
Aug 30, 2026
agentClocks/agentTtlTimers were keyed only by the synthetic clientId (deterministic per access token, regardless of which shard's Awareness it's operating against). A cross-shard agent hold batch under the same token now legitimately holds on two different Awareness instances at once (#125 made hold_records capable of this) — the second shard's scheduleTtl() call was silently cancelling the first shard's timer via the shared clientId key, so that hold never auto-expired. Both maps are now Map<Awareness, Map<number, T>>, with get-or-create helpers threaded through writeRemoteState/scheduleTtl/clearTtl. 635/635 tests passing (2 new, proving independent TTL expiry across two Awareness instances under the same clientId). Refs #120
…re-service-resolution-120 # Conflicts: # src/lib/server/catalog.test.ts # src/lib/server/db/schema.ts
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Stacks on #119 (Phase A catalog) — base branch is
feat/workspace-catalog-113-phase-a, notmain, so this diff only shows what's new here. Merge #119 first, then retarget/merge this one.Every service function that used to call
resolveWorkspaceContext()bare now resolves its actual target shard via new catalog primitives, falling back to the default context when untracked.createCollectionstill deliberately assignsshardId: 'default'— this PR proves the resolution mechanism correct for a genuinely separate shard (tests manually construct one, the same pattern as the holds eviction-wiring fix in #124) without changing where content actually lives yet. Production behavior is unchanged; no client orattach-ws.tschanges are needed in this slice, since the browser's live sync path is never affected.catalog.ts:reserveRecordLocator/releaseRecordLocator(row-level — closes the gap wherewrite_record/delete_record/hold_recordsonly ever receive a barerecordId, no parent hint),resolveShardForParent/resolveShardForRecord.permissions.ts:resolveParentWorkspaceContext/resolveRecordWorkspaceContext/groupRecordIdsByShard— shared resolution helpers.requireAccessibleRecordis now itself shard-aware, which every existing caller already goes through.records.ts:createRecordreserves a row locator when its parent is a Collection;writeRecord/deleteRecord/getRecordresolve via the record's own locator.collections.ts:queryCollection/updateCollectionTitle/deleteCollectionresolve the collection's real shard.holds.ts:hold_records/release_recordsgrouprecordIdsby resolved shard and operate against each shard's own Awareness (a cross-document agent batch is a stated acceptance criterion — seecollaboration.md).search.ts: Collections are enumerated via the catalog first (resolving each one's real shard, including its own meta entry — not just its rows), with a fallback pass over the default doc for uncataloged (direct-Yjs-written) Collections the catalog loop can't see.No MCP tool schema changes — every tool already carries enough of an id for server-side shard resolution.
Test plan
npm run test— 661/661 passing (13 new)npm run lint/npm run check— cleancreateCollection's always-'default'behavior) and assertqueryCollection,createRecord,writeRecord,getRecord,deleteRecord,holdRecords/releaseRecords, andsearchWorkspaceall correctly reach into that shard rather than the default docRefs #120 (not closing — the shard-assignment cutover itself and Document sharding remain)
🤖 Generated with Claude Code