Repository navigation
feat: Execute fixed recurrence and create relative schedules (#20) - #22
Conversation
|
To preview the documentation for this pull request, visit the following URL:
|
Bugbot: review incomplete
Incomplete coverage
|
There was a problem hiding this comment.
Current status: Overall coverage is partial; this snapshot does not prove that all of its findings are resolved.
Last reconciled oned7cd14. See aggregate Bugbot status.
🤖 Bugbot review snapshot
Bugbot reported 1 potential problem when commit ec6a719 was analyzed. This snapshot is historical; use the status block above for current state. 1 finding is linked to changed code.
Findings
- medium: Accepts slots that the schedule never generates —
backend/src/domus/domain/recurrence_records.py:99
There was a problem hiding this comment.
Current status: Overall coverage is partial; this snapshot does not prove that all of its findings are resolved.
Last reconciled oned7cd14. See aggregate Bugbot status.
🤖 Bugbot review snapshot
Bugbot reported 1 potential problem when commit 06b3b9f was analyzed. This snapshot is historical; use the status block above for current state. 1 finding is linked to changed code.
Findings
- medium: Historical slots are not verified against their original pattern —
backend/src/domus/infrastructure/sqlite_recurrence_integrity.py:24
There was a problem hiding this comment.
Current status: Overall coverage is partial; this snapshot does not prove that all of its findings are resolved.
Last reconciled oned7cd14. See aggregate Bugbot status.
🤖 Bugbot review snapshot
Bugbot reported 1 potential problem when commit 19ff35c was analyzed. This snapshot is historical; use the status block above for current state. 1 finding is linked to changed code.
Findings
- medium: Relative source cycle cannot be decoded from stored slots —
backend/src/domus/infrastructure/sqlite_relative_recurrence_records.py:181
Refs #20. Continues retained recurrence execution after the planning foundation. Fixed schedules create actual recurring work; relative schedules now support strict persistence and current-authorized creation with an explicit first obligation. The complete issue and all 52 companion + 37 owner cases remain in progress. The owner now requests merging verified available work; the full unfinished scope is preserved in continuation #23 before this checkpoint merge. This merge does not establish complete recurrence or product acceptance.
Relative creation fingerprints the initial nominal local date/time, interval, snapshotted policy and successor visibility with existing task content, overrides and material selections. Pinned time preparation runs outside writers, while current task-create/configuration/material-link grants precede original lookup and are rechecked at commit. One definition, rule and pristine generation-bound progress commits with the original receipt/audit/outbox. The first obligation remains unconsumed for actual materialization, and ordinary manual creation keeps its existing behavior.
Fixed execution retains actual atomic materialization, original replay, restored absent-effect reconciliation, fair bounded budgets and both owned ASGI lifetimes. Fixed and relative records share global typed identities, kind-aware CAS, one nonretired timing owner per definition and complete original timing/source evidence.
Validation at ed7cd14:
Bugbot completes 15/15 code partitions, 133 fragments and 105 files with zero candidate, published, open or reopened findings. Six valid prior fixes remain confirmed; one reproduced false positive is dismissed. Its exact dict-versus-text reproduction and harmful-suggestion control are in discussion_r4442303467 and vypdev/copilot#855. The three relevant codec/document modules are byte-identical to19ff35c; no production workaround was applied.
Overall native review remains partial/neutral because human preflight retains 7 of 9 items, omitting 2 without truncation. Separate issue6/6, PR14/14, threads7/7 and diff105/105 dimensions are complete. All nine known agent-authored technical inputs are manually inspected with provenance preserved; no new owner answers, omitted-item identity or whole-prompt absence is inferred. Progress still rejects the complete105-path/4-part comparison with validation.evidence-too-large, reference2e271fe7-5bf7-4341-bb09-55e8bbe633b9, recovery usage null and no concrete capacity dimension/count/limit. Reports851/843/855 remain distinct. No Copilot resource, configuration, credential or bound changed.
Outstanding: actual relative initial/successor materialization, reviewed attachment/adoption, bounded current/historical source observations, durable inbox/cycle closure/eligibility/invalidation and restore reconciliation; one shared mixed-rule budget/frontier and both runtime compositions; protected calendar/operation APIs, reviewed schedules/exceptions/classification and configurable household timing defaults; complete retained-case, UI/HA/Stable-Edge/all-domain acceptance. All unfinished scope is explicitly retained in #23. Administration #21 owns schema9 after this actual schema8 merge; any later recurrence DDL must use the next genuinely merged schema version. Action lifecycle closure must not be interpreted as full feature completion.
Schema8 remains unreleased. This creation increment changes no layout, dependency, timeout or owner installation. Supported released1..7 inputs and owner Edge.5/schema7/rollback remain unchanged. No HA update, release or deployment is claimed. See docs/architecture-recurrence.md and docs/evidence/recurrence-relative-creation.json for exact source, evidence and limits.