The patch provides FFI access to Gecko's SeenPtrs type from Rust, in
order to record what has already been measured when measuring Arcs. (The
SeenPtrs must be initialized on the Gecko side because the same table is
reused for measuring all Elements within a window, because Elements can
share ComputedValues.) I have confirmed with DMD that this is working
correctly.
The patch also introduces MallocSizeOfRepeats, which is like
MallocSizeOf but takes a SizeOfState, which holds a SeenPtrs table.
The alias is left there temporarilly and will be removed completely in a later commit where
also components/style/gecko/generated/structs_{debug|release}.rs are re-generated (they still
use the old alias).
This is unfortunate, but for now it complicates things, I would like not needing
a ElementData to get the style of an Element in order to fix all the
getDefaultComputedStyle bugs.
MozReview-Commit-ID: LZvsdFEqrDE
`ElementStyles` holds an optional list of values for each eager pseudo-element.
However, the type was declared as a slice instead of a fixed size array, so an
extra 8 bytes were being allocated to hold the size, even though it never
changes.
Moving the constant size into the type reduces `ElementStyles` and `ElementData`
by 8 bytes.
MozReview-Commit-ID: GaO6DKFxUMo
Before this change, the `ComputedStyle` struct that is part of permanent style
data per element holds 2 `StrongRuleNode`s (unvisited and visited) and 2
`Arc<ComputedValues>` (unvisited and visited).
Both rule nodes and the visited values don't actually need to be here. This
patch moves these 3 to new temporary storage in `CascadeInputs` on
`CurrentElementInfo` during the match and cascade process. Rule nodes are
pushed down inside the `ComputedValues` for later access after the cascade.
(Visited values were already available there.)
The permanent style data per element now has just the `Arc<ComputedValues>` for
itself and eager pseudo-elements (plus the `RestyleHint`).
MozReview-Commit-ID: 3wq52ERMpdi
Otherwise, tests like the following fail, given we always match as unvisited,
and we'd never mark the link as needing invalidation.
<!doctype html>
<style>
a {
color: red !important;
}
.foo :visited {
color: green !important;
}
</style>
<div>
<a href="https://google.es">visit me</a>
<button onclick="this.parentNode.className = 'foo'">Then click me</button>
</div>
Bug: 1368240
MozReview-Commit-ID: LDv6S28c4ju
This commit also removes the old restyle_hints module and splits it into
multiple modules under components/style/invalidation/element/.
The basic approach is to walk down the tree using compound selectors as needed,
in order to do as little selector-matching as possible.
Bug: 1368240
MozReview-Commit-ID: 2YO8fKFygZI
To support visited styles, we match and cascade a separate set of styles any
time we notice that an element has a relevant link.
The visited rules and values are held in `ComputedStyle` alongside the
regular rules and values, which simplifies supporting various APIs like
`cascade_primary_and_pseudos` which expect easy access to previously matched
rules.
To simplify passing the additional values around, an additional reference to the
visited `ComputedValues` is placed inside the regular `ComputedValues`.
MozReview-Commit-ID: 2ebbjcfkfWf
In the case where we process an element which has SMIL animations in normal travesal
the SMIL styles must have been computed in animation-only restyles. So we
have only to pick the computed styles instead of recomputing it.
If we add/remove important rules, we may need to update a list of all important
rules (in Gecko) which overrides animation properties. Therefore, we need to
set a flag if we update the primary rules which includes important ones.
If we have animations on this element, we update its effect properties, and
also send a task to update cascade results.
Calling get_properties_overriding_animations() might cases some impact
on performance because we need to walk the rule tree, so if possible, we could
just store this set into TNode to avoid finding the properties for both old
and new rules each time. This could be a future work if necessary.
I've chosen this approach mainly because there's no other good way to guarantee
the model is correct than holding the snapshots alive until a style refresh.
What I tried before this (storing them in a sort of "immutable element data") is
a pain, since we call into style from the frame constructor and other content
notifications, which makes keeping track of which snapshots should be cleared an
which shouldn't an insane task.
Ideally we'd have a single entry-point for style, but that's not the case right
now, and changing that requires pretty non-trivial changes to the frame
constructor.
MozReview-Commit-ID: FF1KWZv2iBM
Signed-off-by: Emilio Cobos Álvarez <emilio@crisal.io>
