[helo060228]

Describe your changes

A frozen benedict rejects d["a.b"] = 2, but del d["a.b"] and d.pop("a.b") still remove the nested value. Keypath deletion resolves a nested parent and operates on it without reaching the root frozen check.

Check the root state in benedict.__delitem__ and benedict.pop, alongside the existing benedict.__setitem__ check. The production change adds five lines. Directly obtained nested dictionaries and lists remain mutable, as documented for shallow freeze. The lower-level keylist/keypath classes retain their existing behavior.

Five test methods cover dotted paths, keylists, list indexes, custom separators, pop with omitted/None/explicit defaults, missing paths, unfrozen and unfrozen-again instances, and direct child mutation. Add one Unreleased changelog entry.

Related issue

Fixes #634

Validation

• Base commit c4c8a58d517b4ceaa21b6d7529afe1527446100e.
• Unmodified source: python -m unittest — 963 tests ran.
• New tests against unmodified source: 25 freeze test methods ran, with 38 failing subcases across three new methods (24 pop keypath, 8 del keypath, 6 missing keypath with default). Mutable and direct-child controls passed.
• Patched source: python -m unittest tests.dicts.test_benedict_freeze — 25 tests passed.
• Patched source: python -m unittest — 968 tests ran.
• All applicable pinned pre-commit hooks passed for the changed files, including Mypy 2.4.0. check-yaml had no matching files.
• python -m mypy benedict (Mypy 2.3.1) — no issues in 65 source files. git diff --check passed.

Environment: Windows CPython 3.13.5; an earlier run on Linux CPython 3.12.14 produced the same test counts and the same failure breakdown. A few IO tests that fetch fixtures over the network were skipped by transport errors in both runs and are unrelated to this change.

Checklist before requesting a review

• I have performed a self-review of my code.
• I have added tests for the proposed changes.
• I have run the tests and there are not errors.

Disclosure: developed, tested, and self-reviewed with Codex. No human or independent agent review is claimed.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.69%. Comparing base (c4c8a58) to head (edd4218).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #635   +/-   ##
=======================================
  Coverage   98.69%   98.69%           
=======================================
  Files          65       65           
  Lines        2446     2450    +4     
=======================================
+ Hits         2414     2418    +4     
  Misses         32       32           

Flag	Coverage Δ
unittests	98.69% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.