You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Today, color-contrast gives an incomplete result anytime it detects text with a background-image that involves any sort of gradient.
However, it's relatively common for gradient backgrounds to be "simple" cases like this:
<spanstyle="background-image: linear-gradient(rgb(240, 240, 240), rgb(255, 255, 255))"
>
Text
</span>
...where color contrast would pass if tested at any point along the interpolation between the 2 color stops.
It would be a nice feature for axe-core to be able to handle "simple" gradient cases like this by testing a foreground color against multiple background colors and only giving a bgGradient incomplete if the different tests give different results.
In color spaces where relative luminance changes monotonically during a gradient interpolation (eg, sRGB), this would just require testing against each color stop (and confirming that "which of fg and bg is lighter" doesn't swap between the stops). In other color spaces (probably including the default gradient interpolation color space (oklab)), it might be simplest to just pick some arbitrary number of points in the color space along the interpolation and test each.
The text was updated successfully, but these errors were encountered:
Product
axe-core
Feature Description
Today,
color-contrast
gives anincomplete
result anytime it detects text with abackground-image
that involves any sort ofgradient
.However, it's relatively common for gradient backgrounds to be "simple" cases like this:
...where color contrast would pass if tested at any point along the interpolation between the 2 color stops.
It would be a nice feature for axe-core to be able to handle "simple" gradient cases like this by testing a foreground color against multiple background colors and only giving a
bgGradient
incomplete if the different tests give different results.In color spaces where relative luminance changes monotonically during a gradient interpolation (eg, sRGB), this would just require testing against each color stop (and confirming that "which of fg and bg is lighter" doesn't swap between the stops). In other color spaces (probably including the default gradient interpolation color space (oklab)), it might be simplest to just pick some arbitrary number of points in the color space along the interpolation and test each.
The text was updated successfully, but these errors were encountered: