Split the structural-change selector metadata into directional bits for
first/last-child and forward/backward positional selectors.
This gives sibling invalidation enough information to distinguish which
side of a mutation can affect an element, instead of treating all
structural selectors as bidirectional.
Introduce Document::update_layout_if_needed_for_node() which only calls
update_layout() when the node is connected. Use it at all call sites
that query layout metrics (offsets, client dimensions, image size, SVG
bounding box, etc.) so disconnected elements no longer trigger an
unnecessary layout.
With per-paintable display list command caching now in place, the
separate paint-only properties resolution phase is no longer needed.
Resolution now happens inline during painting and its cost is amortized
since it only runs on cache miss.
Move all property resolution to point of consumption:
- is_visible() and visible_for_hit_testing() compute on the fly
- Filter resolution moved to assign_accumulated_visual_contexts()
- Border radii, outlines computed on access
- Box shadows, backdrop filter resolved inline during painting
- Background resolution moved into paint_background()
- Mask resolution moved to StackingContext::paint()
- Text fragment and SVG stroke properties resolved during painting
Rename Document::set_needs_display() to set_needs_repaint() and make it
private. External callers must now go through Node/Paintable which
route the request to the document internally.
Fix one existing misuse in AnimationEffect that was calling
document-level set_needs_display() instead of routing through the
target element's paintable.
This is preparation for per-paintable display list command caching:
repaint requests must go through specific paintables so their cached
command lists can be invalidated.
* Don't publicly expose methods that are only used internally.
* Modernize comment style (wrap at 120 chars, use NB for our notes).
* Change the stringifier for RequestFullscreenError to return a UTF-16
string. We were allocating a String, then transcoding it to UTF-16.
* Give helper methods clearer names, e.g.:
fullscreen_has_error_check -> is_element_allowed_to_enter_fullscreen
This was arguably put in a worse place by #8162; we mostly need the
device pixel offsets from the scroll state so keep track of those and
convert back to CSS pixels when necessary (i.e. scrollbar data).
Add unsafe_layout_node(), unsafe_paintable(), and unsafe_paintable_box()
accessors that skip layout-staleness verification. These are for use in
contexts where accessing layout/paintable data is legitimate despite
layout not being up to date: tree construction, style recalculation,
painting, animation interpolation, DOM mutation, and invalidation
propagation.
Also add wrapper APIs on Node to centralize common patterns:
- set_needs_display() wraps if (unsafe_paintable()) ...set_needs_display
- set_needs_paint_only_properties_update() wraps similar
- set_needs_layout_update() wraps if (unsafe_layout_node()) ...
And add Document::layout_is_up_to_date() which checks whether layout
tree update flags are all clear.
Stop converting between CSS and device pixels as part of rendering - the
display list should be as simple as possible, so convert to DevicePixels
once when constructing the display list.
These properties only return the computed border width, which is a
style-level value that doesn't depend on layout geometry. Replace
the full update_layout() call with update_style_if_needed_for_element()
and resolve border widths directly from computed properties.
This avoids potentially expensive synchronous layout when only
CSS computed values are needed.
Very profitable on https://x.com/ where we avoid lots of layout work.
Eliminate O(n) tree-order sorted insert on every add() by simply
appending elements and deferring order resolution to lookup time. Cache
the first-in-tree-order element in get() so repeated getElementById()
calls avoid repeated subtree traversal.
Improves performance on YouTube where add() was hot in profiles while
scrolling through comments.
Elements in internal shadow trees that represent CSS pseudo-elements
(e.g. the DetailsContent slot in <details>) store their cascaded
properties on the host element's pseudo-element data, not on the
element itself. This meant that when slotted elements walked the
inheritance chain and encountered such a slot, they would dereference
null cascaded properties and crash.
Fix this by copying the cascaded properties onto the slot element
itself after computing its style, keeping both cascaded and computed
properties accessible in the same place.
Previously we computed font properties separately from other properties
for two reasons:
1) These font properties were computed using a different length
resolution context than the rest of the properties.
2) These properties were required to be computed before creating the
length resolution context for the rest of the properties.
The first issue was solved in the previous commit by introducing a
generic method to get the computation context for a property, and
the second is solved in this commit by computing properties in the
required order.
This simplifies the code a bit and opens up some opportunities for
optimization.
Replace per-element OrderedHashMap storage for custom properties with
a RefCounted chain (CustomPropertyData) that enables structural
sharing. Each chain node stores only the properties declared directly
on its element, with a parent pointer to the inherited chain.
Elements that don't override any custom properties share the parent's
data directly (just a RefPtr copy). During cascade, only entries that
actually differ from the parent are stored in own_values - the rest
are inherited through the chain. During var() resolution, resolved
values are compared against the parent's and matching entries are
dropped, enabling further sharing.
The chain uses a depth limit (max 32) with flattening, plus
absorption of small parent nodes (threshold 8) to keep lookups fast.
This reduces custom property memory from ~79 MB to ~5.7 MB on
cloudflare.com.
Two issues prevented slotted elements from correctly inheriting
styles from their assigned slot:
1. Element::element_to_inherit_style_from() was skipping the slot
element and returning the shadow host instead. This meant slotted
elements inherited from the host, completely ignoring any styles
on the slot itself.
2. When a slot element's style changed during the style tree walk,
its assigned (slotted) nodes were never marked for recomputation.
The tree walk follows the DOM tree, but slotted elements are DOM
children of the shadow host, not the slot, so they were missed.
Fix (1) by returning the slot directly as the inheritance parent.
Fix (2) by marking assigned nodes dirty in update_style_recursively
when a slot's style changes.
Remove unused/redundant includes from Element.h:
- AK/IterationDecision.h (redundant)
- ARIA/AttributeNames.h (redundant via ARIAMixin.h)
- CSS/CascadedProperties.h (redundant via PseudoElement.h)
- CSS/StylePropertyMapReadOnly.h (pointer types only)
- HTML/LazyLoadingElement.h (unused in header)
Extract IntersectionObserverRegistration struct from
IntersectionObserver.h into its own lightweight header.
This breaks the heavy transitive include chain through
IntersectionObserverEntry.h and Geometry/DOMRect.h that
was pulled into every file including Element.h.
Indirect recompilation impact reductions:
- IntersectionObserver.h: ~1387 -> ~27 files
- LazyLoadingElement.h: ~1387 -> ~1002 files
Remove includes from Node.h that are only needed for forward
declarations (AccessibilityTreeNode.h, XMLSerializer.h,
JsonObjectSerializer.h). Extract StyleInvalidationReason and
FragmentSerializationMode enums into standalone lightweight
headers so downstream headers (CSSStyleSheet.h, CSSStyleProperties.h,
HTMLParser.h) can include just the enum they need instead of all of
Node.h. Replace Node.h with forward declarations in headers that only
use Node by pointer/reference.
This breaks the circular dependency between Node.h and
AccessibilityTreeNode.h, reducing AccessibilityTreeNode.h's
recompilation footprint from ~1399 to ~25 files.
Remove 11 heavy includes from Document.h that were only needed for
pointer/reference types (already forward-declared in Forward.h), and
extract the nested ViewportClient interface to a standalone header.
This reduces Document.h's recompilation cascade from ~1228 files to
~717 files (42% reduction). Headers like BrowsingContext.h that were
previously transitively included see even larger improvements (from
~1228 down to ~73 dependents).
Previously only input elements were matched. Add placeholder_value()
to HTMLTextAreaElement mirroring the HTMLInputElement API and update
both selector matching code paths to handle textarea.
Replace the direct #include of StyleInvalidation.h in Element.h with a
forward declaration in Forward.h. Element.h only uses the type in
function declarations, so the complete type is not needed.
This reduces the recompilation impact of modifying StyleInvalidation.h
from ~1380 files to ~4 files, since Element.h is transitively included
by nearly every HTML and SVG element header.
`Layout::NodeWithStyle::computed_values()` actually holds used values
which may not be the same as computed values e.g. if they have been
modified by `Document::propagate_overflow_to_viewport()`
Instead of doing a full document style invalidation when a stylesheet is
dynamically added, we now analyze the new sheet's selectors to determine
which elements could potentially be affected, and only invalidate those.
This works by building an InvalidationSet from the rightmost compound
selector (the "subject") of each rule in the new stylesheet, extracting
class, ID, tag name, attribute, and pseudo-class features. We then walk
the DOM tree and only mark elements matching those features as needing a
style update.
If any selector has a rightmost compound that is purely universal (no
identifying features), or uses a pseudo-class not supported by the
invalidation set matching logic, we fall back to full invalidation.
In Element::set_scroll_offset(), when setting a pseudo-element's
scroll offset, the code was also incorrectly setting the generating
element's own m_scroll_offset. Added an else branch so only the
pseudo-element's offset is set.
Also adds a ref test for scrollable pseudo-elements to prevent
regression. The test scrolls a ::before pseudo-element via wheel
event and verifies the content scrolls correctly.
This ensures that we are explicitly declaring the allocator to use when
allocating a cell(-inheriting) type, instead of silently falling back
to size-based allocation.
Since this is done in allocate_cell, this will only be detected for
types that are actively being allocated. However, since that means
they're _not_ being allocated, that means it's safe to not declare
an allocator to use for those. For example, the base TypedArray<T>,
which is never directly allocated and only the defined specializations
are ever allocated.
The previous implementation had a bug: it composed all ancestor
transforms but applied them around only the innermost element's
transform origin. The correct behavior is to apply each transform
around its own origin.
AccumulatedVisualContext already tracks all visual transformations
(transforms, scroll offsets, perspective) correctly for hit testing.
This change adds a new transform_rect_to_viewport() method that performs
the forward transformation (element coordinates to viewport
coordinates), which is the inverse direction of
transform_point_for_hit_test().
This fixes getBoundingClientRect() returning incorrect coordinates for
elements inside transformed ancestors with non-default
transform-origins.
Previously, compute_required_invalidation() compared font lists using
cached_computed_font_list(), which returns the lazily-cached value.
Since newly computed styles haven't had their font list computed yet,
this compared a non-null cached value (old style) against null (new
style), causing unnecessary relayout even when fonts hadn't changed.
Fix by using computed_font_list() to ensure both styles have their
font lists computed before comparison.
This change is currently entirely undetectable because of what the
added FIXME talks about. Currently, the HTML element's overflow is
always set to visible in both axes, so it getting set to "clip" in
the imported test ends up not mattering at all.
This allows us to use these methods from `SVGAElement` without
inheriting `HTMLHyperlinkElementUtils`, which we can't do for
`SVGAElement` due to a naming conflict with the `href()` method in
`SVGURIReferenceMixin`.
A lot of our scrolling code is quite old, and doesn't match the spec,
but does use some similar names. This is quite confusing. In particular
`perform_scroll_of_viewport()` is not the same as the spec algorithm.
That algorithm is actually almost implemented in
`scroll_viewport_by_delta()`.
To clarify things, this commit makes a few changes:
- Rename perform_scroll_of_viewport() to
perform_scroll_of_viewport_scrolling_box(). This is a better match
for how we use this method, even if it's not actually a match for the
algorithm. (:yakbait:)
- Move `scroll_viewport_by_delta()`'s code into a new
`perform_a_scroll_of_the_viewport()` method, and make it take a
position like it should. `scroll_viewport_by_delta()` now calls it
with a calculated position.
I've avoided reusing the original `perform_scroll_of_viewport()` name to
avoid accidents.
An animation with an orphaned owning element should continue to be
ticked by the timeline.
Reverts c8b574e and instead avoids leaking animations by not visiting
`Animation`s from `AnimationTimeline`s.
Fixes a timeout in the imported test
The existing part_list() method used by the bindings lazily creates a
DOMTokenList, which we don't want to do just to check if an Element has
any parts defined.
