Files
msd-core/tests/refactor-trigger-cli.test.cjs
Tom Boucher 4f32209f78 enhance(#3267): reduce handleEvaluate complexity below refactor-trigger's own threshold (#3978)
* fix(#3267): reduce handleEvaluate complexity below refactor-trigger's own threshold

handleEvaluate scored 26 (then 21 after later #3261 commits) against the
complexity-triggered-refactor feature's own default threshold of 15. Extracts
the read-and-analyze loop (analyzeTouchedFiles) and the artifact/baseline/
ledger write path (finalizeEvaluation) into named helpers, per the issue's
suggested direction. Behavior-preserving: every existing test in
tests/refactor-trigger-cli.test.cjs is unchanged, and every degrade-path
reason code (REFACTOR_INVALID_PHASE, REFACTOR_GIT_UNAVAILABLE,
REFACTOR_NO_TOUCHED_FILES, REFACTOR_FILE_UNREADABLE,
REFACTOR_ANALYZER_UNSUPPORTED, REFACTOR_ANALYZER_UNPARSEABLE,
REFACTOR_BASELINE_WRITE_FAILED, REFACTOR_STRICT_NOT_ENFORCING) keeps its
current value and emission path.

The four complexity-trigger.cts lexer functions (scanFunctions,
stripLiterals, skipTypeExpr, skipGenericParamList) are deliberately left
untouched, per ADR-1953 D5 — they are a hand-rolled lexer state machine,
densely branchy by construction, and refactoring them to lower the metric
would be exactly the "split a coherent function to satisfy a metric"
behavior D5 exists to prevent.

Closes #3267

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* docs: backfill changeset PR number for #3978

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

---------

Co-authored-by: sim <sim@local>
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-28 08:12:27 -04:00

44 KiB