Backtestspydanticrun-20261003-150343#13687
populate_by_name (and implicit validate_by_name when validate_by_alias=False) is ignored by @validate_call
Better
Our patch was judged better than the merged fix.
In this judgment, Patch A = our agent, Patch B = the merged fix.
Both patches fix the root cause in `_arguments_schema` and `_arguments_v3_schema`. They stop embedding the raw `validate_by_name` key and use a value that accounts for `populate_by_name` and for `validate_by_alias=False` implying name validation. Patch A adds an `effective_validate_by_name` property that reproduces the `core_config()` logic. Patch B calls `core_config(title=None)` and reads `validate_by_name` from the result. B's approach reuses the existing logic and avoids duplicating it, which is a small craft advantage. Assuming `core_config` applies the `validate_by_alias=False` implication, both patches handle both scenarios in the issue. Patch A also applies the property to the model and dataclass signature generation, which is adjacent hardening and low risk. The deciding factor is tests. Patch A tests both scenarios from the issue (`populate_by_name=True` and `validate_by_alias=False`). It tests them through `validate_call`, through both `arguments` and `arguments-v3` schema types via `generate_arguments_schema`, and through model signatures. Its `validate_by_alias=False` test also asserts that the alias is rejected. Patch B tests only `populate_by_name`, and only for arguments-v3 plus one `validate_call` case. It leaves the issue's 'worse variant' (`validate_by_alias=False`) untested. Patch B also rewrites `dict(...)` to `{...}` in unrelated existing tests, which adds noise. Overall, A has stronger test coverage of the reported scenarios and no unrelated churn.
The judge saw only the two patches, labelled A and B in random order, with issue links and version numbers removed. The reasoning above is shown verbatim.
Solve
9m 38s
Wall-clock time in the container
Full test suite
No new failures
12,400 passed · 500 skipped · 12,900 total
Contamination audit
Clean
No sign the agent saw the real fix
Patch size
63vs 39 merged
Changed lines, ours vs merged