[by-ae]

String mutations that case-flip characters were also changing the hex digits inside escape sequences like \u00a0, \xHH, \UHHHHHHHH, and the names inside \N{NAME}. Python treats those hex digits and unicode names case-insensitively, so mutants like "\u00A0" are equivalent to the original "\u00a0" and can never be killed by tests.

What changed

Modified the NON_ESCAPE_SEQUENCE regex in src/mutmut/mutation/mutators.py to match and preserve complete escape sequences with case-insensitive hex digits or names (\xHH, \uHHHH, \UHHHHHHHH, \N{NAME}), while still allowing case transformations for regular text and for raw strings (where escape sequences are literal and case changes are not equivalent).

Updated the existing test expectation for \N{ghost} \U11223344 to reflect that these are now preserved, and adjusted the e2e snapshot (4 mutants instead of 5, dropping the equivalent one).

How it was tested

• Added test cases covering \u, \U, \x, \N{...} escapes and mixed content
• Added a test for raw strings (r"..."), where case changes are not equivalent and still happen
• Full test suite passes: 467 passed, 4 skipped
• All pre-commit hooks pass (ruff, mypy, formatting)

An AI coding assistant helped with the implementation.

Fixes #609

[boxed]

Thanks!