This fixes common crash related to slottables, currently present on wpt.fyi.
Previously, the traversal parent of `Text` nodes was incorrectly
assumed to always be the parent or shadow host. That caused crashes
inside stylo's bloom filter. Now the traversal parent is the slot
that the node is assigned to, if any, and the parent/shadow host otherwise.
The slottable data for Text/Element nodes is now stored in NodeRareData.
This is very cheap, because NodeRareData will already be instantiated
for assigned slottables anyways, because the containing_shadow_root
field will be set (since assigned slottables are always in a shadow
tree). This change is necessary because we need to hand out references
to the assigned slot to stylo and that is not possible to do (without
unsafe code) if we need to downcast the node first.
As a side effect, this reduces the size of `Text` from 256 to 232 bytes,
because the slottable data is no longer stored there.
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Implement slot-related algorithms
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Hook up slot elements to DOM creation logic
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Set a slot assignment mode for servo-internal shadow roots
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Assign slots when a slottable's slot attribute changes
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Properly compute slot name
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* ./mach test-tidy
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Update <slot> name when name attribute changes
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Implement fast path for Node::assign_slottables_for_a_tree
assign_slottables_for_a_tree traverses all descendants of the node
and is potentially very expensive. If the node is not a shadow root
then assigning slottables to it won't have any effect, so we
take a fast path out.
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Move slottable data into ElementRareData
This shrinks all element descendants back to their
original size.
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Address review comments
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Update WPT expectations
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
---------
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
This change splits the style and layout data in DOM nodes that is
populated by style and layout passes. This makes Servo's data design
more like Gecko's. This allows:
1. Removing the various `StyleAndLayout` data structures used by layout.
2. Removing the `GetStyleAndLayoutData` and
`GetStyleAndOpaqueLayoutData` traits. Accessing style and layout data
are now just functions on the `LayoutNode` and `ThreadSafeLayoutNode`
traits.
3. Styling now doesn't populate layout data. This is is postponed until
layout itself.
4. Allows the DOM wrappers to no longer have to be generic over the
layout data. This data was already stored using `std::any::Any` and
the new code just makes layout responsible for downcasting. Cleaning
up the generic type parameter in the DOM wrappers can happen in a
followup change.
The main benefit to all of this is that we should be able to remove
unsafe creation of `ServoLayoutNode` in layout and
`TrustedLayoutNodeAddress` entirely, because `ServoLayoutNode` will be
able to be passed directly from script to layout. In addition, this
removes one more abstraction layer from the layout DOM wrappers, making
the code a lot more understandable.
Note: This increases the measured size of DOM types, but the same data
is stored. It's simply that before that data was stored behind a heap
pointer.
This includes a `size_of` regression for a few DOM types,
due to https://github.com/rust-lang/rust/pull/58623 which replaces the
implementation of `HashMap` in the standard library to Hashbrown.
Although `size_of<HashMap>` grows, it’s not obvious how total memory usage
is going to be impacted: Hashbrown only has one `u8` instead of one `usize`
of overhead per hash table bucket for storing (part of) a hash,
and so might allocate less memory itself.
Hashbrown also typically has better run time performance:
https://github.com/rust-lang/hashbrown#performance
Still, I’ve filed https://github.com/rust-lang/hashbrown/issues/69
about potentially reducing the `size_of<HashMap>` regression.
Implement custom element reactions
<!-- Please describe your changes on the following line: -->
Initial work for implementing custom element reactions: https://html.spec.whatwg.org/multipage/custom-elements.html#custom-element-reactions
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#17433 (github issue number if applicable).
<!-- Either: -->
- [X] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/17614)
<!-- Reviewable:end -->
This allows us to have ensure_data() and clear_data() functions on the TElement
trait, instead of hacking around it adding methods in random traits.
This also allows us to do some further cleanup, which I'd rather do in a
followup.
This will go back down soonish when NonZero stablizes and we can use it in
StyleArc, which I expect will be long before the memory overhead here matters.
This patch does a number of things, unfortunately all at once:
* Hoists a large subset of the layout wrapper functionality into the style system.
* Merges TElementAttributes into the newly-created TElement.
* Reorganizes LayoutData by style vs layout, and removes LayoutDataShared.
* Simplifies the API for borrowing style/layout data.
There's still more to do to make the style system usable standalone, but
this is a good start.
There is now an inclusive_descendants_version field of each node, which
increases each time the node, or any of its descendants, is dirtied.
This can be used for cache invalidation, by caching a version number
and comparting the current version number against the cached version number.
After rebasing, this suddenly became a problem again, even though there's no actual size increase here
(we're shrinking NodeFlags by 1 byte, and adding 1 byte of EventState). Moving the NodeFlags to the end
of Node and the EventState bits to the beginning of Element doesn't seem to helper either.
This is probably a padding issue that's worth investigating at some point, but given the level
of churn in this code it doesn't seem worth it to fuss to much over this right now.
