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
I have tested the issue with the latest version of the product
Issue Description
We received a customer report recently of a color-contrast false-negative where it reported a bgOverlapincomplete for a case that it ought to have been able to detect as a violation. This happens due to axe's grid machinery calculating an incorrect stacking order for the elements in question; it incorrectly determines that one of the ancestor container elements is being rendered on top of the leaf descendant element, which is inconsistent with the actual browser paint order (and the order reported by document.elementsFromPoint).
The specific customer repro is private, but the test cases below represent minimal repros of the same underlying issue. In the motivating example, this pattern was part of a more complicated interactive grid control.
Expectation
The suggested new test cases below should all be satisfied
Actual
The test cases are not satisfied; the integration tests incomplete with bgOverlap and the unit test shows an incorrect stack order calculation:
1) should correctly order deeply nested stacking contexts
dom.getElementStack stack order
AssertionError: expected [ '2', 'target', '3', '1', 'fixture' ] to deeply equal [ 'target', '3', '2', '1', 'fixture' ]
at Context.<anonymous> (test/commons/dom/get-element-stack.js:434:14)
How to Reproduce
Suggested new case for test/commons/dom/get-element-stack.js
We think this is probably the same root cause as #4350 (where it causes a target-size false positive instead of this color-contrast false negative). We started working on this in #4351 but didn't complete before hitting our timebox.
The text was updated successfully, but these errors were encountered:
Product
axe-core
Product Version
4.10.2
Latest Version
Issue Description
We received a customer report recently of a color-contrast false-negative where it reported a
bgOverlap
incomplete
for a case that it ought to have been able to detect as a violation. This happens due to axe's grid machinery calculating an incorrect stacking order for the elements in question; it incorrectly determines that one of the ancestor container elements is being rendered on top of the leaf descendant element, which is inconsistent with the actual browser paint order (and the order reported bydocument.elementsFromPoint
).The specific customer repro is private, but the test cases below represent minimal repros of the same underlying issue. In the motivating example, this pattern was part of a more complicated interactive grid control.
Expectation
The suggested new test cases below should all be satisfied
Actual
The test cases are not satisfied; the integration tests
incomplete
withbgOverlap
and the unit test shows an incorrect stack order calculation:How to Reproduce
Suggested new case for
test/commons/dom/get-element-stack.js
Suggested new cases for
test/integration/rules/color-contrast/color-contrast.html
Updating
color-contrast.json
with#fail13
,#fail14
,#pass24
,#pass25
:Additional context
We think this is probably the same root cause as #4350 (where it causes a target-size false positive instead of this color-contrast false negative). We started working on this in #4351 but didn't complete before hitting our timebox.
The text was updated successfully, but these errors were encountered: