Backtestspydanticrun-20260930-144322#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 B = our agent, Patch A = the merged fix.
Both patches fix the root cause the issue identifies: the raw `self._config_wrapper.validate_by_name` value was embedded into the arguments and arguments-v3 schemas. Patch A replaces it with `core_config(title=None).get('validate_by_name')`, which reuses the existing backwards-compatibility mapping. That is the more elegant single source of truth, but it builds a full core config just to read one key. Patch B adds `_effective_validate_by_name` to `ConfigWrapper`. This property duplicates the precedence rules from `core_config()`: explicit `validate_by_name`, then `populate_by_name`, then `validate_by_alias is False`. It covers both scenarios in the issue, at the cost of a small risk of the two copies drifting apart later. On correctness they are roughly equal, since both handle `populate_by_name=True` and the implicit `validate_by_alias=False` case.
Tests are the deciding factor. Patch B tests the main `populate_by_name` scenario thoroughly: every combination of name and alias for two parameters. It also tests the 'worse variant' the issue highlights, `validate_by_alias=False`, checking that the name works and the alias is rejected. It adds an arguments-v3 test that checks both alias and name input. Patch A tests only `populate_by_name` with a single `foo(a=1)` call, never tests the `validate_by_alias=False` case, and leaves `<link>` placeholder docstrings.
Patch A also adds unrelated churn by rewriting `dict(...)` to `{...}` in existing tests. That is not harmful but obscures the diff. Overall, Patch B's coverage of both reported scenarios outweighs Patch A's slightly cleaner reuse of `core_config()`.
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
3m 30s11 turns
Wall-clock time in the container
Full test suite
No new failures
12,395 passed · 500 skipped · 12,895 total
Contamination audit
Clean
No sign the agent saw the real fix
Patch size
52vs 39 merged
Changed lines, ours vs merged