Analysis of selectors on modern websites shows that the `:hover`
pseudo-class is mostly used in the subject position within relatively
simple selectors like `.a:hover`. This suggests that we could greatly
benefit from segregating them by id/class/tag name, this way reducing
number of selectors tested during hover style invalidation.
With this change, hover invalidation on Discord goes down from 70ms to
3ms on my machine. I also tested GMail and GitHub where this change
shows nice 2x-3x speedup.
This is required to store Content Security Policies, as their
Directives are implemented as subclasses with overridden virtual
functions. Thus, they cannot be stored as generic Directive classes, as
it'll lose the ability to call overridden functions when they are
copied.
Regressed in 036327332f.
This commit moves the optimization a little later in replaceData(),
still avoiding relayout (the important part).
Recovers 480 points on WPT. :^)
If selector does not have any descendant combinators then we know for
sure it won't be filtered out by ancestor filter, which means there is
no need to check for it.
This change makes hover style invalidation go faster on Discord where
with this change we spend 4-5% in `should_reject_with_ancestor_filter()`
instead of 20%.
Allows us to avoid invalidating layout when CharacterData didn't change.
Results in visible improvement on Discord that continuously invokes
this function with the same data, which previously resulted in relayout
on every frame.
Before this change, we did the following:
1. Created a bitmap with the matching state for each rule containing
`:hover`.
2. Changed the actively hovered element in the document.
3. Created another bitmap with the matching state for each rule
containing `:hover`.
With this change, we iterate rule by rule and compare the matching
state. This allows us to break early once we find the first rule whose
matching state changes after the hovered element update. Additionally,
this removes the need to allocate a bitmap.
Instead of using `has_pseudo_elements()` that iterates over all pseudo
elements, only check if `::before` or `::after` are present.
Before this change, `has_pseudo_elements()` was 10% of profiles on
Discord while now it's 1-2%.
...in inherited style update. Instead of comparing old absolutized value
with new non-absolutized value, we should wait until
`absolutize_values()` and then compare old and new values, when both are
absolutized.
Improves performance on pages with GitHub action logs where previously
we had to invalidate layout after hover style recalculation, because
there was `margin-left: 1rem`.
MediaQueryList will now remember if a state change occurred when
evaluating its match state. This memory can then be used by the document
later on when it's updating all queries, to ensure that we don't forget
to fire at least one change event.
This also required plumbing the system visibility state to initial
about:blank documents, since otherwise they would be stuck in "hidden"
state indefinitely and never evaluate their media queries.
Instead of checking all elements in a document for containment in
`:has()` invalidation set, we could narrow this down to ancestors and
ancestor siblings, like we already do for subject `:has()` invalidation.
This change brings great improvement on GitHub that has selectors with
non-subject `:has()` and sibling combinators (e.g., `.a:has(.b) ~ .c`)
which prior to this change meant style invalidation for whole document.
This allows us to inspect its properties. To avoid wasted work, we only
compute and cache the properties if the originating element was, or is,
displaying as a list item.
This commit changes the strategy for updating inherited styles. Instead
of marking all potentially affected nodes during style invalidation, the
decision is now made on-the-fly during style recalculation. Child nodes
will only have their inherited styles recalculated if their parent's
properties have changed.
On Discord this allows to 1000x reduce number of nodes with recalculated
inherited style.
`invalidate_style()` already tries to avoid scheduling invalidation for
`:has()` by checking result of `may_have_has_selectors()`, but it might
still result in unnecessary work because `may_have_has_selectors()`
does not force building of rules cache. This change adds
`have_has_selectors()` that forces building of rules cache and is
invoked in `update_style()` to double-check whether we actually need to
process scheduled `:has()` invalidations.
This allows to skip ~100000 ancestor traversals on this WPT test:
https://wpt.live/html/select/options-length-too-large.html
Moves pseudo class matching helpers into Element methods, so they don't
have to be duplicated between SelectorEngine and function that checks if
element is included in invalidation set.
The current implementation of `:has()` style invalidation is divided
into two cases:
- When used in subject position (e.g., `.a:has(.b)`).
- When in a non-subject position (e.g., `.a > .b:has(.c)`).
This change focuses on improving the first case. For non-subject usage,
we still perform a full tree traversal and invalidate all elements
affected by the `:has()` pseudo-class invalidation set.
We already optimize subject `:has()` invalidations by limiting
invalidated elements to ones that were tested against `has()` selectors
during selector matching. However, selectors like `div:has(.a)`
currently cause every div element in the document to be invalidated.
By modifying the invalidation traversal to consider only ancestor nodes
(and, optionally, their siblings), we can drastically reduce the number
of invalidated elements for broad selectors like the example above.
On Discord, when scrolling through message history, this change allows
to reduce number of invalidated elements from ~1k to ~5.
Before this change, tasks associated with a destroyed document would get
stuck in the task queue forever, since document-associated tasks are not
allowed to run when their document isn't fully active (and destroyed
documents never become fully active again). This caused everything
captured by task callbacks to leak.
We now treat tasks for destroyed documents as runnable immediately,
which gets them out of the queue.
This fixes another massive GC leak on Speedometer.
Before this change, Agent held on to all of the live MutationObserver
objects via GC::Root. This prevented them from ever getting
garbage-collected.
Instead of roots, we now use a simple IntrusiveList and remove them
from it in the finalizer for MutationObserver.
This fixes a massive GC leak on Speedometer.
f7a3f78 made the layout tree invalidate only the inserted nodes
themselves, but it turned out that CSS containment invalidation relies
on the parent being invalidated as well.
There is no need for this invalidation because taking care of siblings
is already done by invalidation with `NodeInsertBefore` reason. Parent
element itself (without subtree) is always invalidated by
`Node::children_changed()` hook, so `:empty` pseudo-class invalidation
is already covered.
When checking whether an early return is possible because some ancestor
already has the whole subtree invalidation flag set, the check should
begin with the current node's parent rather than with the node itself.
Otherwise, if a node already has the whole subtree invalidation flag
set and is subsequently invalidated for the reason `NodeInsertBefore`
or `NodeRemove`, we will skip the sibling invalidation required for
these operations
This fix is required for optimizations in subsequent commits.
With this change, siblings of an inserted node are no longer invalidated
unless the insertion could potentially affect their style. By
"potentially affected," we mean elements that are evaluated against the
following selectors during matching:
- Sibling combinators (+ or ~)
- Pseudo-classes :first-child and :last-child
- Pseudo-classes :nth-child, :nth-last-child, :nth-of-type, and
:nth-last-of-type
This is not really a context, but more of a set of parameters for
creating a Parser. So, treat it as such: Rename it to ParsingParams,
and store its values and methods directly in the Parser instead of
keeping the ParsingContext around.
This has a nice side-effect of not including DOM/Document.h everywhere
that needs a Parser.
While keyword_to_foo() does return Optional<Foo>, in practice the
invalid keywords get rejected at parse-time, so we don't have to worry
about them here. This simplifies the user code quite a bit.
