Instead of applying animations and transitions to styled elements,
include them in the cascade. This allows them to interact properly with
things like font-size and !important rules.
Depends on D80234
Differential Revision: https://phabricator.services.mozilla.com/D80235
Do the same we do for classes for now. We could be more precise and achieve a
bit more sharing with some more effort (left a comment there), but it seems
unlikely to matter in practice (and if we did that, we'd probably want to do the
same for classes).
Differential Revision: https://phabricator.services.mozilla.com/D58453
We already avoid putting styles in the shared style cache for an element that
has animations[1] but if we are doing the initial style of two siblings where
the _second_ has animations applied, there is nothing to stop us from trying to
share with the first (un-animated) element. This patch adds a check to prevent
sharing in that case since sharing style between animated elements is unsound
for the reasons described in [1].
A number of tests including:
testing/web-platform/tests/web-animations/animation-model/animation-types/visibility.html
will fail without this fix once we remove the style flush from
KeyframeEffect::SetKeyframes later in this patch series.
[1] https://searchfox.org/mozilla-central/rev/6e3cc153566f5f288ae768a2172385b8436d61dd/servo/components/style/sharing/mod.rs#597
Differential Revision: https://phabricator.services.mozilla.com/D18913
We were spuriously reframing the <shadow> because it initially shared style with
the <br>, which ended up being display: none, while the <shadow> should've been
display: contents from the beginning.
lookup_by_rules seems pretty prone to obscure bugs, and also it's pretty
complex... Probably we should try to get rid of it, I'm unconvinced that it's
worth it.
Even with that, in a normal restyle the <details> wouldn't have ended up with a
style. It of course never had it before the reframe because the <shadow> was
display: none, but that doesn't mean it shouldn't have gotten one, since we
detected we needed to go through kids in:
https://searchfox.org/mozilla-central/rev/6eea08365e7386a2b81c044e7cc8a3daa51d8754/servo/components/style/matching.rs#500
That code did happen, but since it's an animation-only restyle, we don't look at
unstyled stuff.
That looks somewhat fishy, but I guess for now it's fine as long as display
isn't animatable.
MozReview-Commit-ID: B6NMSTNOKgK
Currently, NAC always inherits from the closest non-NAC ancestor element,
regardless of whether it is for an element-backed pseudo or not.
This patch changes the inheritance so that for element-backed pseudos, we
inherit from the closest native anonymous root's parent, and for other NAC we
inherit from the parent.
This prevents the following two issues and allows us to remove the
NODE_IS_NATIVE_ANONYMOUS flag:
* Avoiding inheriting from the non-NAC ancestor in XBL bindings bound to NAC.
- This is no longer a problem since we apply the rule only if we're a
pseudo-element, and all pseudo-elements are in native anonymous subtrees.
- This also allows to remove the hack that propagates the
NODE_IS_NATIVE_ANONYMOUS flag from the ::cue pseudo-element from
BindToTree.
* Inheriting from the wrong thing if we're a nested NAC subtree.
- We no longer look past our NAC subtree, with the exception of
::-moz-number-text's pseudo-elements, for which we do want to propagate
::placeholder to.
A few rules from forms.css have been modified because they're useless or needed
to propagate stuff to the anonymous form control in input[type="number"] which
previously inherited from the input itself.
MozReview-Commit-ID: IDKYt3EJtSH
Now that rustfmt is getting close to stable, and work on the style system has died down a bit, it seemed like an opportune time to auto-format the style crates.
The first commit disables import reordering, since tidy and rustfmt don't currently agree on the correct ordering. The second commit does a bunch of manual fixups such that the output of rustfmt passes tidy. The third commit runs rustfmt on the three aforementioned crate.
There are a few dozen warnings in the style crate about lines longer than 100 characters. It would be good to fix these, but I don't have time for that now.
Source-Repo: https://github.com/servo/servo
Source-Revision: 9a900ef019cd643bff961d7b20db6da69f3edb29
Kinda tricky because :host only matches rules on the shadow root where the rules
come from. So we need to be careful during invalidation and style sharing.
I didn't use the non_ts_pseudo_class_list bits because as soon as we implement
the :host(..) bits we're going to need to special-case it anyway.
The general schema is the following:
* Rightmost featureless :host selectors are handled inserting them in the
host_rules hashmap. Note that we only insert featureless stuff there. We
could insert all of them and just filter during matching, but that's slightly
annoying.
* The other selectors, like non-featureless :host or what not, are added to the
normal cascade data. This is harmless, since the shadow host rules are never
matched against the host, so we know they'll just never match, and avoids
adding more special-cases.
* Featureless :host selectors to the left of a combinator are handled during
matching, in the special-case of next_element_for_combinator in selectors.
This prevents this from being more invasive, and keeps the usual fast path
slim, but it's a bit hard to match the spec and the implementation.
We could keep a copy of the SelectorIter instead in the matching context to
make the handling of featureless-ness explicit in match_non_ts_pseudo_class,
but we'd still need the special-case anyway, so I'm not fond of it.
* We take advantage of one thing that makes this sound. As you may have
noticed, if you had `root` element which is a ShadowRoot, and you matched
something like `div:host` against it, using a MatchingContext with
current_host == root, we'd incorrectly report a match. But this is impossible
due to the following constraints:
* Shadow root rules aren't matched against the host during styling (except
these featureless selectors).
* DOM APIs' current_host needs to be the _containing_ host, not the element
itself if you're a Shadow host.
Bug: 992245
Reviewed-by: xidorn
MozReview-Commit-ID: KayYNfTXb5h
Source-Repo: https://github.com/servo/servo
Source-Revision: cb754b262747e7cab794411df55588f0f0b30b5e
Also move the checks so that the tag name checks (presumably cheaper) come earlier.
Source-Repo: https://github.com/servo/servo
Source-Revision: 087cf21d79cb090c2a8b55b441e6004921b540aa
This was reviewed by Xidorn in #20223, but that needs to wait for a few days so let's land this separately.
Source-Repo: https://github.com/servo/servo
Source-Revision: 21290276c2c85cf5dd55474ba7a641fb13f127a6
This allows to selector-match ::slotted, though we still don't parse it.
Bug: 1425834, 1424607, 1425755
Reviewed-by: heycam
MozReview-Commit-ID: ItELHkf2PMl
Source-Repo: https://github.com/servo/servo
Source-Revision: bb34b7f54aa85cde600c6d675360d03714cf1f8b
This code used to insert duplicate entries to avoid expensive shuffling of the LRU cache. With uluru this is no longer necessary, because reordering the cache is cheap.
Now it uses the `LRUCache::find` method from uluru 0.2 to update entries in-place. This should increase cache hits by eliminating unnecessary evictions.
This PR also updates `arrayvec` because `uluru` depends on it, and `immeta` because it depends on `arrayvec`.
---
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes do not require tests because they don't change behavior
Source-Repo: https://github.com/servo/servo
Source-Revision: 0c13f14aaa151764e1b3dd2d4503cb6fc3c9595b
We check that a few lines before calling test_candidate.
Source-Repo: https://github.com/servo/servo
Source-Revision: ec794c82978a60dd0d5fb65b7712138c3535b714
Reland of #19045, because it was backed out for an unexpected pass.
Bug: 1412251
Reviewed-by: bz
MozReview-Commit-ID: II6lk6OmSZU
Source-Repo: https://github.com/servo/servo
Source-Revision: 52f962c387456308b68f062bff1ad72947da79b9
It's easy to construct examples where not inserting in those cases causes performance
to get worse (for example, any long list of siblings that match the same selectors
while having some non-effectual differences in LocalName/Class/Id/etc). And the LRU
nature of the cache already does the right thing of pruning non-useful entries.
Fixing this causes several hundred more sharing hits on wikipedia.
Source-Repo: https://github.com/servo/servo
Source-Revision: da64abc4298549362debc96e5843bd36e6d23b7f
This avoids doing wasted work, at least in the recascade case, and pretty likely
in the other as well.
Source-Repo: https://github.com/servo/servo
Source-Revision: 988728e9d5c66e6bf2f9e0e03a50a814d0e1f2ab
