Pydantic stops relying on normalization for decimal validation

♥

FastAPI and SQLModel dependencies churned, while Pydantic hardened JSON mode and decimal checks.

20 commits 20 PRs merged 2 briefings covered

What shipped: Pydantic changed decimal validation so it no longer relies on normalization to catch invalid values [1]. It also hardened JSON mode serialization error handling by catching any exception when computing a JSON Schema default [2]. Meanwhile, dependency churn across FastAPI and SQLModel brought routine bumps to the python-packages groups [3][4], and SQLModel followed with sqlalchemy 2.0.52 [5].

Why it matters: Decimal validation now behaves more predictably in your models, catching edge cases normalization previously missed [1]. The JSON mode hardening closes a gap where schema default computation could raise unhandled exceptions that broke serialization [2]. Routine dependency bumps carry no immediate action.

Also: FastAPI updated fastar to 0.12.0 [6], pydantic bumped actions/checkout to 7.0.1 [7], and a docstring typo fix cleaned up fields documentation [8].

References

[1] Do not rely on normalization to validate decimals ↗ pydantic/pydantic
[2] Handle any exception when trying to compute JSON Schema default ↗ pydantic/pydantic
[3] ⬆ Bump the python-packages group across 1 directory with 15 updates ↗ fastapi/fastapi
[4] ⬆ Bump the python-packages group with 9 updates ↗ tiangolo/sqlmodel
[5] ⬆ Bump sqlalchemy from 2.0.51 to 2.0.52 ↗ tiangolo/sqlmodel
[6] ⬆ Bump fastar from 0.11.0 to 0.12.0 ↗ fastapi/fastapi
[7] Bump actions/checkout from 7.0.0 to 7.0.1 ↗ pydantic/pydantic
[8] Fix duplicate word 'it it' in fields docstring (#13756) ↗ pydantic/pydantic

$ ls fastapi/week/ # the briefings behind this review

Wed Sep 2 Dependency churn across FastAPI and SQLModel, decimal validation hardening in Pydantic Thu Sep 3 Pydantic hardens JSON mode serialization error handling

← previous week fastapi wire →