Gecko's GetContext has: if (aFlags & eIsLink) { // If this node is a link, we want its visited's style context's // parent to be the regular style context of its parent, because // only the visitedness of the relevant link should influence style. parentIfVisited = aParentContext; } Stylo doesn't have anything equivalent at the moment. This is causing the inherit-keyword-1.xhtml subtest of test_visited_reftests.html to fail.

Looks like bug 1381635 will re-enable this test.

Never mind, looks like bug 1381635 is unrelated to this.

Yeah, this should be relatively easy to fix... Let me just do that.

Comment on attachment 8888257 [details] Bug 1377469: Update reftest expectations. https://reviewboard.mozilla.org/r/159210/#review164660 ::: layout/style/test/stylo-failures.md:68 (Diff revision 1) > * test_css_supports.html: issues around @supports syntax servo/servo#15482 [2] > * test_author_specified_style.html: support serializing color as author specified bug 1348165 [27] > * browser_newtab_share_rule_processors.js: agent style sheet sharing [1] > * :visited support bug 1381635 > - * test_visited_reftests.html `inherit-keyword-1.xhtml` [2] > * ... `mathml-links.html` [2] Please update this line to replace `...` with `test_visited_reftests.html`. (You could also mention bug 1371030 since that's filed for it.)

Comment on attachment 8888156 [details] Bug 1377469: Don't inherit from the parent visited style if we're a link. https://reviewboard.mozilla.org/r/159070/#review164666 Thanks for working on this! :) ::: servo/components/style/style_resolver.rs:477 (Diff revision 2) > > if self.element.skip_root_and_item_based_display_fixup() { > cascade_flags.insert(SKIP_ROOT_AND_ITEM_BASED_DISPLAY_FIXUP); > } > if cascade_visited.visited_dependent_only() { > + if pseudo.is_some() || !self.element.is_link() { A comment to explain this condition would be great to have! Gecko's `nsStyleSet::GetContext` attempts to explain their version of similar logic, and I think it's helpful for future readers.

(In reply to J. Ryan Stinnett [:jryans] (use ni?) from comment #9) > A comment to explain this condition would be great to have! Gecko's > `nsStyleSet::GetContext` attempts to explain their version of similar logic, > and I think it's helpful for future readers. Err, whoops, the PR is in the queue already, but I'll land the comment.

Pushed by ecoal95@gmail.com: https://hg.mozilla.org/integration/autoland/rev/7707a729e9c8 Update reftest expectations. r=jryans