[hygorcluz]

get_json_schema() made shallow copies of json_schema_extra, then filled in nested items and additionalProperties dicts in place, so those writes ended up in the user's metadata. If two fields share one dict, e.g. {"items": {"description": "Tag."}} on a list[str] field and on a list[int] field of another class, the first schema adds "type": "string" to the shared dict and the second schema then reports string items too. This happened for dataclass, attrs, TypedDict, Pydantic and Scrapy items, and in one edge case for a class-level __json_schema_extra__ (a dataclass with no fields whose properties got attribute docstrings written into them).

This switches every copy() of user-provided schema data in _json_schema.py to deepcopy(), and adds tests for a shared field dict on each item type and for the class-level case. The new tests fail without the change. The py, extra, min-extra, docs, mypy, pylint and twinecheck tox envs and pre-commit pass.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.88%. Comparing base (4983811) to head (33dbea5).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #155   +/-   ##
=======================================
  Coverage   99.88%   99.88%           
=======================================
  Files           6        6           
  Lines         844      844           
  Branches      161      161           
=======================================
  Hits          843      843           
  Partials        1        1           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[codspeed]

Merging this PR will not alter performance

✅ 24 untouched benchmarks
⏩ 4 skipped benchmarks¹

---

_{Comparing hygorcluz:json-schema-extra-deepcopy (33dbea5) with master (4983811)}

Footnotes

1. 4 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[hygorcluz]

The Test: py311, macos-latest job was cancelled because no macOS runner picked it up, unrelated to this change. Could a maintainer re-run it?