design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036

Activity

1. dezren39

   added

   area/infraInfrastructure, deployment, and operationsInfrastructure, deployment, and operations

   kind/designA design proposal (design-architect): multi-step, phased, decision-bearingA design proposal (design-architect): multi-step, phased, decision-bearing

   on Nov 5, 2026
2. developing-today commented on Nov 5, 2026

   

   developing-todayBot

   on Nov 5, 2026 – with developing.today

   Contributor

   More actions

   CODE-1701
3. dezren39

   added

   queue/queuedHas a queued task in the Supabase agent queueHas a queued task in the Supabase agent queue

   task/triageQueue packet kind: verify, reproduce, and route this itemQueue packet kind: verify, reproduce, and route this item

   work/assignedQueue worker lifecycle stateQueue worker lifecycle state

   queue/claimedQueue worker lifecycle stateQueue worker lifecycle state

   status/in-progressQueue worker lifecycle stateQueue worker lifecycle state

   and removed

   queue/queuedHas a queued task in the Supabase agent queueHas a queued task in the Supabase agent queue

   on Nov 5, 2026
4. dezren39 commented on Nov 5, 2026

   

   dezren39

   on Nov 5, 2026

   MemberAuthor

   More actions

   🤖 Agent note · picked up

   • Agent: queue worker (repo-worker-code), lane code-n
   • Session / ID: task 10402
   • Found via: queue claim
   • Task: Triage issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 (design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment) — verify state at current main, scope the design path, record blockers/ordering.
   • Expectation: effort class s; done = triage findings posted on issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 and task 10402 completed as triaged, or closed with evidence if already handled/superseded.
   • Links: parent issue design(atlas): delta-only ingestion + offset embeds — land only changes; span-anchored chunks with non-doubling joins #6028; sibling issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035
5. dezren39 commented on Nov 5, 2026

   

   dezren39

   on Nov 5, 2026

   MemberAuthor

   More actions

   🤖 Agent note · done

   • Agent: queue worker (repo-worker-code), lane code-n
   • Session / ID: task 10402
   • Found via: queue claim
   • Task: Triage of issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 complete — outcome: triaged; issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 stays OPEN as a design ticket.
   • Expectation: done = findings below recorded; no implementation task spawned (premature, see ordering).
   • Links: parent issue design(atlas): delta-only ingestion + offset embeds — land only changes; span-anchored chunks with non-doubling joins #6028; sibling issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035

   Triage findings (verified fresh 2026-11-05)

   • Still live: issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 is OPEN with no hard-stop labels (no held, user/owned, do-not-merge, do-not-auto-edit). No PR implements or owns this design, and there is no duplicate issue — parent issue design(atlas): delta-only ingestion + offset embeds — land only changes; span-anchored chunks with non-doubling joins #6028 and sibling issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035 are the distinct neighbouring tickets in the same Atlas ingestion cluster.
   • Nature: research-grounded algorithm design ticket, not a bug or code fix. The finding is already settled in substance (the matcher is the rsync sliding-window algorithm, Tridgell & Mackerras TR-CS-96-05); the open work is the design pass: rolling-hash choice and window size at embed scale (~1–2 KB spans), strong-verify hash, minimum match run, the token-stream output contract, a per-source-type cost model, failure cases (reflow/re-wrap phantom diffs), and a worked run over the 17–241 example. "Done" per the issue body is that algorithm design, recorded.
   • Ordering / blockers: parent issue design(atlas): delta-only ingestion + offset embeds — land only changes; span-anchored chunks with non-doubling joins #6028 (delta-only ingestion + offset embeds) is still OPEN with no decision record yet — the matcher design must fit its span-anchored ingestion shape and must not fork it. The matcher's output contract (token stream → span references + literal regions) is defined against the span-table schema of sibling issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035, so the two designs are a set: issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036's output contract and issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035's schema should be settled together or explicitly cross-referenced, in either order, before either is implemented.
   • Route: issue design(atlas): offset-matching algorithm — rsync-style sliding-window matcher for span alignment #6036 stays OPEN for a design pass producing the algorithm design with parameters, failure cases, and the worked example. No implementation task is enqueued now: there is nothing to implement until the design (and the sibling issue design(atlas): span-table schema — spans as (file_version_id, start, end) with embedded-text snapshots #6035 schema it emits into) is recorded.
6. dezren39

   removed

   status/in-progressQueue worker lifecycle stateQueue worker lifecycle state

   queue/claimedQueue worker lifecycle stateQueue worker lifecycle state

   work/assignedQueue worker lifecycle stateQueue worker lifecycle state

   on Nov 5, 2026
7. github-actions

   assigned

   dezren39

   on Nov 5, 2026