fix(opy): converge parser source-form conformance - #163
Conversation
Align standalone lambda diagnostics with the pinned parse frontier and move Workshop-owned or lowering-only probes out of parser contracts. Fixes #160
Teakowa
left a comment
There was a problem hiding this comment.
Major — crates/opy-rs/src/lower.rs:1092
Changing the public diagnostic code from lambda-context to generic parse-error weakens an existing stable diagnostic solely to satisfy the conformance frontier. #93 explicitly requires stable source-attributed diagnostics and forbids weakening diagnostics to obtain a pass; the conformance runner already classifies lambda-context as the parse stage. Preserve the lambda-context public diagnostic and align the conformance frontier without regressing that diagnostic contract.
Keep the stable lambda-context diagnostic while normalizing only the conformance frontier to the pinned parse-error construct.
Teakowa
left a comment
There was a problem hiding this comment.
Major — crates/opy-rs/tests/differential.rs:325
The production diagnostic is correctly restored to lambda-context, and the new conformance normalization now maps that code to the reference parse-error frontier. This differential expectation must therefore remain lambda-context; changing it to parse-error contradicts the still-authoritative compiler diagnostic contract (compatibility/compiler-expectations.json) and weakens the native stable-code check. Revert this expectation to lambda-context.
Preserve lambda-context in the native diagnostic expectation while the conformance runner compares its canonical parse frontier.
Summary
Validation
The remaining reported divergences are outside this issue's lexical/parser scope or have explicit Workshop/lowering ownership evidence.
Fixes #160