coreui / coreui-pro / 31265888072

93%

main: 94%

push

github

feat(buttons): move .btn-check onto the label, with the input nested inside (#723)

* feat(buttons): move .btn-check onto the label, with the input nested inside

A toggle button was two elements held together by an `id`/`for` pair: a hidden
input carrying `.btn-check` and a sibling `<label>` carrying `.btn`. It is one
element now — the label carries both classes and the input sits inside it:

  <label class="btn btn-primary btn-check"><input type="checkbox">Toggle</label>

Nothing has to stay in sync, and the button group stops having to reason about
a hidden sibling that is not a button: the `:not(.btn-check)` rules that exist
purely to keep the input from counting as the visually-first child no longer
apply to the new spelling.

The states move from sibling selectors to `:has()` on the label. Disabled dims
rather than repaints, so a disabled toggle that is checked still reads as
checked — matching the deprecated spelling, where a `:disabled` input never
matched the label's own `.btn:disabled` rules.

Both spellings select themselves: the v6 label carries `.btn`, so the
deprecated one is the `.btn-check` that is not a `.btn`. That is
`$btn-check-deprecated`, used everywhere the old sibling logic lives.

Verified in Chromium: the deprecated spelling renders byte-identical to before
across four states, `.btn-group` and `.btn-group-vertical` at two and three
buttons, and a group mixing a toggle with a plain button; the new spelling
matches the old one on every measured property in the same cases.

* docs(buttons): show the nested-input toggle button as the canonical spelling

Button group, checkbox and radio pages carry the new markup; the deprecated
sibling spelling is noted next to the other v5 form-check leftovers rather than
demonstrated twice. The migration guide gets the diff and the reason the old
spelling keeps working.

4267 of 4819 branches covered (88.55%)

Branch coverage included in aggregate %.

8281 of 8675 relevant lines covered (95.46%)

363.57 hits per line