Bound rules-editor inference work with explicit incomplete status
Adversarial nested constraints consume excessive work. Add operation and depth budgets with honest termination status.
- Focused work estimate
- 3h 30m + prerequisites
- Priority in the scenario
- Medium
- Engineering practice
- Resource limits · Static analysis
Estimated field mix
- Compiler and language tooling60%
- Performance engineering40%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
Fictional rules editor Willow supports numbers, strings, booleans, records and functions in a deliberately small language. Define syntax and type rules locally; no arbitrary code execution is needed.
Setup prerequisites
- Create synthetic AST fixtures and a written type-rule table.
- Define source spans and structured diagnostic output.
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
- CTYPE-101 · Assign explicit types to rules-editor literals
- CTYPE-102 · Resolve expression identifiers through lexical scope
- CTYPE-103 · Explain rules-editor operator operand mismatches
- CTYPE-104 · Check function calls for arity before argument inference
- CTYPE-105 · Unify rules-editor type variables with an occurs check
- CTYPE-106 · Report missing rules-editor record fields without losing known fields
- CTYPE-107 · Merge rules-editor branch types through explicit compatibility rules
- CTYPE-108 · Make rules-editor diagnostics deterministic across map insertion order
- CTYPE-109 · Cancel obsolete rules-editor checks without publishing stale results
Acceptance criteria
- Supported fixtures complete
- Budget exhaustion is explicit
- Incomplete checks cannot claim type safety
Implementation constraints
- Use deterministic counters rather than wall time alone.
Verification to include
- Check normal corpus
- Exceed constraint budget
Deliverables
- Inference budget and stress cases
Rollout and recovery
Lower limits and mark affected files incomplete until improved.
Value of the work
For the engineer: Practice type environments, unification and explainable rejection.
For the team: Inspect whether developer tooling catches mistakes without hiding uncertainty.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.