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
{{ message }}
Repository navigation
Share typed-row parameter binding between analyzer validation and discovery rules #7465
The test-data analyzer computes the effective parameter types of a generic method inline, separately from discovery and invocation binding. Optional arguments, injected cancellation tokens, trailing arrays and params collections can therefore drift between validation and execution. Extract a pure binding helper and test its results against the discovery rules.
Proposal
Move the closed-signature row-to-parameter calculation in TestDataAnalyzer.CheckDataGenerator into a helper beside TypedDataSourceInference. Cover omitted optional and trailing array parameters, injected tokens selected by declared position, packed arrays, element conversions, loose trailing values and supported params collections. Assert agreement on inference success versus null, and compare effective expected types with invocation behavior; the existing inference-only parity table does not cover this calculation.
Acceptance criteria
The analyzer calls a pure helper for its effective closed parameter types without changing supported behavior.
Regression coverage checks the row-binding cases above and protects the supported collection set from drift.
Preserve analyzer diagnostics for invalid elements and constraints, and both discovery modes' execution behavior.
Follow-up requested by Claude's current-head review of #7460: #7460 (comment). The relevant code is TestDataAnalyzer. Related inference parity work: #7386; broader inference rules: #7371. Please consolidate this binding computation and add parity coverage.
Activity
Strengthen typed-row token identity assertions (#7465)
Extract typed-row binding and verify invocation parity (#7507)