Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 | 4x | /**
* Per-component exclusion list for the regression-based compliance checks.
*
* A component's "legacy" CSS occasionally declares rules whose selector
* belongs to a *different* component (e.g. the legacy `icon.css` styles
* `.slds-input__icon`, which structurally owns by the `input` component).
* When the new CSS drops or relocates those rules, the regression audit
* correctly emits `hook-selector-moved` / `property-removed` flags — but
* they aren't regressions of the component under test. They're cleanup
* of cross-component baggage that never should have lived there.
*
* Entries here filter those flags out of the regression check
* (`no-unintended-visual-changes`) without touching the raw
* `regression.flags` array in the report; the
* underlying data stays queryable for anyone who wants to inspect it,
* only the row-level score/offender-list is affected.
*
* Match is performed against the flag's normalized selector key (see
* `selectorKey()` in `audit/extract.ts`), so whitespace and comma
* ordering between the entry and the source CSS don't matter.
*
* To add an entry: confirm the selector genuinely belongs to a different
* component (not a real regression of this one), append it with a short
* inline comment pointing to the other component. Keep the list minimal —
* anything here is "silent to the reviewer" and should survive a `git blame`.
*/
export const IGNORED_SELECTORS: Readonly<Record<string, readonly string[]>> = {
icon: [
// `.slds-input__icon` is an input-component selector. The legacy
// `icon.css` historically styled it; the new icon CSS nests it under
// `.slds-icon .slds-input__icon` (input still owns the selector).
'.slds-input__icon',
],
};
|