coreui / coreui-pro / 32432557987

94%

main: 94%

push

github

fix(CloseButton): drop .btn-close-white and collapse the size surface (#811)

The icon has been a mask painted with the element's own color since the
SVGs moved off background-image, so .btn-close-white only ever restated
what the inherited color already gives you. Measured on a .text-bg-dark
toast, a .text-bg-primary surface and a .bg-body.text-body section under
data-coreui-theme="dark", the button renders identically with and without
it. Deprecated since v5.0.0; removed now.

The dark-mode rule behind it goes too. It forced the glyph white on every
descendant of [data-coreui-theme=dark], which made the close button the
one element in a bare color-mode section that looked right while the text
beside it stayed unreadable - a section flips the tokens but paints
nothing, so a plain <span> there keeps the outer page's ink. It also got
the mirror case wrong: a light section inside a dark page inherited the
forced white and drew a white icon on white. Without the rule every case
measures correctly, including that one.

Five size variables described one square: width, height and a padding per
axis, with height and padding-y already defined as copies of their
siblings. They collapse into --cui-btn-close-icon-size and
--cui-btn-close-padding. The rendered box is unchanged - 1em of glyph
inside .25em of padding, so the pointer target stays 24px at the default
font size, which is the WCAG 2.2 minimum. Upstream went further and
dropped the padding entirely; that leaves a 20px target, so we keep ours.

--cui-btn-close-bg becomes --cui-btn-close-icon: it stopped being a
background when the glyph moved to a mask, and the old name sent anyone
customizing it looking for the wrong property.

The docs page still described the filter-and-background-image
implementation, so it gets rewritten with a CSS variables section, and the
three pages that paired .btn-close with .btn-close-white drop the class.

4537 of 5095 branches covered (89.05%)

Branch coverage included in aggregate %.

8661 of 9055 relevant lines covered (95.65%)

399.87 hits per line