Update from code changes: SymbolicVerifier DiagnosticResult (v5.3.0) - #249
Update from code changes: SymbolicVerifier DiagnosticResult (v5.3.0)#249mintlify[bot] wants to merge 2 commits into
Conversation
|
Skipping PR review because a bot author is detected. If you want to trigger CodeAnt AI, comment |
QWED Security Verification Report3 files scanned | 0 blocked | 0 warnings | 0 info | 0 suppressed | 3 verified Engines
Verified Files
Verified by QWED — deterministic security verification. No LLM used. |
|
Preview deployment for your docs. Learn more about Mintlify Previews.
|
Greptile SummaryRewrites the SymbolicVerifier documentation for the v5.3.0 DiagnosticResult contract.
Confidence Score: 3/5The release documentation should not merge until the canonical installation page is synchronized with v5.3.0; the symbolic-limits reference also needs two non-blocking consistency fixes. Announcing v5.3.0 while the Releases page still installs v5.2.0 can leave users running the legacy return contract against the new migration examples, and the symbolic-limits page contains contradictory and incomplete result guidance. Files Needing Attention: changelog.mdx, advanced/symbolic-limits.mdx
|
| Filename | Overview |
|---|---|
| advanced/symbolic-limits.mdx | Migrates examples and result documentation to DiagnosticResult, but retains contradictory Verified benchmark labels and omits bounds_transform_error from its reason tables. |
| changelog.mdx | Announces v5.3.0 and its breaking API migration, while the canonical releases page still directs users to install v5.2.0. |
| engines/code.mdx | Rewrites the Code engine reference around the six DiagnosticResult-returning SymbolicVerifier methods and removes fictional legacy options. |
Comments Outside Diff (1)
-
advanced/symbolic-limits.mdx, line 193-195 (link)Benchmarks claim impossible VERIFIED results
These benchmark rows label clean runs as “Verified,” contradicting the new contract that the engine always reports such runs as
UNVERIFIABLEwithsymbolic_verifier.no_counterexample_found; this gives readers conflicting expectations about an impossible status.
Reviews (1): Last reviewed commit: "docs: tighten SEO descriptions on code e..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| | Artifact | Previous | This release | | ||
| | --- | --- | --- | | ||
| | `qwed` (PyPI) | 5.2.0 | 5.3.0 | | ||
| | `qwed_sdk` (Python) | 5.2.0 | 5.3.0 | | ||
| | `@qwed-ai/sdk` (npm) | 5.2.0 | 5.3.0 | | ||
| | `qwed` (Rust crate) | 5.2.0 | 5.3.0 | | ||
| | API version marker | 5.2.0 | 5.3.0 | | ||
| | K8s deployment image | 5.2.0 | 5.3.0 | |
There was a problem hiding this comment.
When users follow the newly announced v5.3.0 release to the canonical Releases page, its commands still install v5.2.0, causing them to run the legacy return contract against the new DiagnosticResult migration examples and encounter missing attributes or keys.
| <ResponseField name="developer_fields.constraint_id" type="string"> | ||
| Structured reason code. One of: | ||
| - `symbolic_verifier.no_counterexample_found` — clean CrossHair run, still `UNVERIFIABLE` because completeness was not proven. | ||
| - `symbolic_verifier.counterexample_found` — CrossHair disproved a check. | ||
| - `symbolic_verifier.timeout` — verification did not converge within `timeout_seconds`. | ||
| - `symbolic_verifier.incomplete_coverage` — at least one function could not be checked. | ||
| - `symbolic_verifier.no_typed_functions` — no functions carried the type hints CrossHair requires. | ||
| - `symbolic_verifier.no_verifiable_functions` — the source contained no functions. | ||
| - `symbolic_verifier.syntax_error` — the code could not be parsed. | ||
| - `symbolic_verifier.crosshair_not_available` — the CrossHair engine is not installed. | ||
| - `symbolic_verifier.verification_error` — verification did not complete cleanly. |
There was a problem hiding this comment.
Both authoritative-looking constraint_id lists omit symbolic_verifier.bounds_transform_error, even though verify_bounded() is documented to return it, leaving consumers without a complete set of reasons to handle.
|
Preview deployment for your docs. Learn more about Mintlify Previews.
|
Summary
Rewrites the Code engine reference around the new
DiagnosticResultreturn contract shipped in QWED-Verification v5.3.0, and adds the corresponding changelog entry.Changes
engines/code.mdx— rewritten around theDiagnosticResultcontract. New sections cover the status/constraint_id cheat sheet,verification_mode(symbolicvsbounded_symbolic), whyVERIFIEDis intentionally never emitted, per-methoddeveloper_fieldsmaps, and a before/after migration snippet. Removes references to the never-wiredcheck_assertionsparameter and the fictionalcheck_security/check_typesarguments.advanced/symbolic-limits.mdx— updated result-field docs and examples to theDiagnosticResultshape, including the newconstraint_idreason table and the fail-closed matrix for typed/untyped/no-function inputs.changelog.mdx— new v5.3.0 entry noting the six migrated methods, theverification_modefield, the removal ofcheck_assertions, and that this is the first engine to conform to the unified diagnostic model. Includes a legacy-to-new field cheat sheet, before/after snippet, and version-propagation table.Context
Source PRs in QWED-Verification:
check_assertionsparameterverification_modeverify_boundedget_verification_budgetanalyze_complexityverify_safety_properties