This change contains three semi-related clean ups:
1. the `to_webrender()` and `from_webrender()` functions on Pipeline are
turned into more-idiomatic `From` and `Into` implementations.
2. `combine_id_with_fragment_type` now returns a `u64` as that is what is
expected for all callers and not a `usize`.
3. The `query_scroll_id` query is removed entirely. The
`ExternalScrollId` that this queries is easily generated directly
from the node's opaque id. Querying into layout isn't necessary at
all.
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.
Remove the use of unsafe code in the layout wrappers of the DOM. The
main change here is that `unsafe_get()` no longer needs to be an unsafe
method, which allows us to transitively remove or reduce unsafe blocks
from callers. The function itself is not renamed, because it's still
a bit dangerous to start removing the layers of abstraction from actual
DOM nodes.
In addition `init_style_and_opaque_layout_data` can be merged into
`initialize_data`, which removes one more unsafe method.
Finally, a "Safety" section is added to some unsafe methods.
Instead of the tricky `LayoutRPC` interface, query layout using the
`Layout` trait. This means that now queries will requires calling layout
and then running the query. During layout an enum is used to indicate
what kind of layout is necessary.
This change also removes the mutex-locked `rw_data` from both layout
threads. It's no longer necessary since layout runs synchronously. The
one downside here is that for resolved style queries, we now have to
create two StyleContexts. One for layout and one for the query itself.
The creation of this context should not be very expensive though.
`LayoutRPC` used to be necessary because layout used to run
asynchronously from script, but that no longer happens. With this
change, it becomes possible to safely pass nodes to layout from script
-- a cleanup that can happen in a followup change.
This brings the version of WebRender used in Servo up-to-date with Gecko
upstream. The big change here is that HiDPI is no longer handled via
WebRender. Instead this happens via a scale applied to the root layer in
the compositor. In addition to this change, various changes are made to
Servo to adapt to the new WebRender API.
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
Instead of using a simple `Atom` to identify a local font, use a data
structure. This allows us to carry more information necessary to
identify a local font (such as a path on MacOS). We need this for the
new version of WebRender, as fonts on MacOS now require a path.
This has a lot of benefits:
1. We can avoid loading fonts without paths on MacOS, which should
avoid a lot of problems with flakiness and ensure we always load the
same font for a given identifier.
2. This clarifies the difference between web fonts and local fonts,
though there is more work to do here.
3. This avoid a *lot* of font shenanigans, such as trying to work
backwards from the name of the font to the path of the font we
actually matched. In general, we can remove a lot of code trying to
accomplish these shenanigans.
4. Getting the font bytes always returns an `Arc` now avoiding an extra
full font copy in the case of Canvas.
The specification gives instructions for how these values should be
propagated. The other big changs here is that they aren't applied to the
`<body>`.
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
This also ignores a clippy warning for a new function (and a similar
existing one), until this code can be refactored to use temporary Rust
strutures to carry display list building state.
There are a few new test failures here:
- FAIL [expected PASS] /css/css-images/image-set/image-set-conic-gradient-rendering.html
- FAIL [expected PASS] /css/css-images/image-set/image-set-repeating-conic-gradient-rendering.html
These fail because Servo does not yet support `image-set()`.
- FAIL [expected PASS] /css/filter-effects/filter-function/filter-function-conic-gradient.html
- FAIL [expected PASS] /css/filter-effects/filter-function/filter-function-repeating-conic-gradient.html
These fail because Servo does not support the very early filter effects
specification.
- FAIL [expected PASS] /html/canvas/element/manual/fill-and-stroke-styles/conic-gradient-rotation.html
- FAIL [expected PASS] /html/canvas/element/manual/fill-and-stroke-styles/conic-gradient.html
These fail because this change only adds support for CSS conical
gradients. Another set of changes will be necessary to support this for
Canvas.
Synthetic small caps is supported by the font subsystem, but this is
disabled in Layout 2020. We can turn this on to bring support to parity
with the old layout system.
In addition to turning on synthetic small-caps this change also improves
the way that they work. Before, synthetic small caps meant that every
character was a small version of capitalized character. After this
change, capital letters are larger than small caps versions of small
letters -- matching other browsers and the common expectation of how
small caps works.
* Remove packages that were moved to external repo
* Add workspace dependencies pointing to 2023-06-14 branch
* Fix servo-tidy.toml errors
* Update commit to include #31346
* Update commit to include servo/stylo#2
* Move css-properties.json lookup to target/doc/stylo
* Remove dependency on vendored mako in favour of pypi dependency
This also removes etc/ci/generate_workflow.py, which has been unused
since at least 9e71bd6a70.
* Add temporary code to debug Windows test failures
* Fix failures on Windows due to custom target dir
* Update commit to include servo/stylo#3
* Fix license in tests/unit/style/build.rs
* Document how to build with local Stylo in Cargo.toml
* script: Do not run layout in a thread
Instead of spawning a thread for layout that almost always runs
synchronously with script, simply run layout in the script thread.
This is a resurrection of #28708, taking just the bits that remove the
layout thread. It's a complex change and thus is just a first step
toward cleaning up the interface between script and layout. Messages are
still passed from script to layout via a `process()` method and script
proxies some messages to layout from other threads as well.
Big changes:
1. Layout is created in the script thread on Document load, thus every
live document is guaranteed to have a layout. This isn't completely
hidden in the interface, but we can safely `unwrap()` on a Document's
layout.
2. Layout configuration is abstracted away into a LayoutConfig struct
and the LayoutFactory is a struct passed around by the Constellation.
This is to avoid having to monomorphize the entire script thread
for each layout.
3. Instead of having the Constellation block on the layout thread to
figure out the current epoch and whether there are pending web fonts
loading, updates are sent synchronously to the Constellation when
rendering to a screenshot. This practically only used by the WPT.
A couple tests start to fail, which is probably inevitable since removing
the layout thread has introduced timing changes in "exit after load" and
screenshot behavior.
Co-authored-by: Josh Matthews <josh@joshmatthews.net>
* Update test expectations
* Fix some issues found during review
* Clarify some comments
* Address review comments
---------
Co-authored-by: Josh Matthews <josh@joshmatthews.net>
This adds support for table rows, columns, rowgroups and colgroups.
There are few additions here:
1. The createion of fragments, which allows script queries and hit
testing to work properly. These fragments are empty as all cells are
still direct descendants of the table fragment.
2. Properly handling size information from tracks and track groups as
well as frustrating rules about reordering rowgroups.
3. Painting a background seemlessly across track groups and groups. This
is a thing that isn't done in legacy layout (nor WebKit)!
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
In order for stylo to be a separate crate, it needs to depend on less
things from Servo. This change makes it so that stylo no longer depends
on servo_url.
This adds support for fixing up tables so that internal table elements
that are not properly parented in the DOM have the correct box tree
structure according to the CSS Table specification [1]. Note that this
only comes into play when building the DOM via script, as HTML 5 has its
own table fixups that mean that the box tree construction fixups here
are not necessary.
There are no tests for this change. In general, it's hard to write tests
against the shape of the box tree, because it depends on the DOM. We
plan to test this via WPT tests once layout is complete.
1. https://drafts.csswg.org/css-tables/#table-internal-element
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
* Add initial support for sticky positioning for non-legacy layout
Many tests still fail for a variety of reasons. One of the primary ones
is that CSSOM currently does not return correct values for elements
positioned by sticky nodes. This requires changes to WebRender to work
properly.
* Fix an assertion failure in the legacy layout sticky code
Layout asserts that it never creates stacking contexts that have a zero
scale, yet it doesn't prevent the creation of those stacking contexts.
This change stops their creation at an earlier stage.
Fixes#30118.
This is the start of the organization of types that are in their own
crates in order to break dependency cycles between other crates. The
idea here is that putting these packages into their own directory is the
first step toward cleaning them up. They have grown organically and it
is difficult to explain to new folks where to put new shared types. Many
of these crates contain more than traits or don't contain traits at all.
Notably, `script_traits` isn't touched because it is vendored from
Gecko. Eventually this will move to `third_party`.
In legacy layout, anonymous text wrappers were inheriting the `overflow`
and `text-overflow` properties. This results in the creation of extra
clipping for these anonymous wrappers which could clip away floats. We
will likely implement `text-overflow` differently in non-legacy layout.
This change marks all legacy layout pseudo elements as "legacy" and also
adds a new pseudo element for non-legacy layout that does not inherit
`overflow`.
Fixes#30562.
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
When display lists update quickly, a hit test result might be returned
for a previous display list / list of hit test items. When that happens,
ignore the hit test result.
This fixes a crash, but there might be situations where we can do
something better, such as wait for display list processing to finish
before performing the hit test. A future change might do this for events
like mouse clicks and touch events that should never be thrown away.
Ultimately, the best thing is likely moving hit testing back to layout
or script so a valid hit test can always be performed against the
current DOM.
Fixes#29796.
* Implement support for `drop-shadow`
* Clean up remnant from early attempts
* Fix misleading comments on GenericSimpleShadow
If Servo-specific `style` changes will need to be upstreamed anyway, I might as well fix a thing that had thrown me off!
* Revert "Fix misleading comments on GenericSimpleShadow"
This reverts commit cdc810b826ac082041adc212c24649ee3b86ca0a.
* Clean up an import
* Update test expectations
* Fix missing expectation on Layout 2013
Refactor the scrolling and scrollable area calculation on the window
object, to make it better match the specification. This has some mild
changes to behavior, but in general things work the same as they did
before. This is mainly preparation for properly handling viewport
propagation of the `overflow` property but seems to fix a few issues as
well.
There is one new failure in Layout 2020 regarding `position: sticky`,
but this isn't a big deal because there is no support for `position:
sticky` in Layout 2020 yet.
Co-authored-by: Rakhi Sharma <atbrakhi@igalia.com>
* Upgrade vendored version of WebRender
* Patch WebRender: upgrade version of gleam
* Restore hit testing implementation
* Fix WebRender warnings
* Adapt Servo to new WebRender
* Update results
* Add a workaround for #30313
This slightly expands text boundaries in order to take into account the
fact that layout isn't measuring glyph boundaries.
Before WebRender would ignore these, but newer version of WebRender have
issues with them. This change simply prevents legacy layout from
creating display items for these types of gradients. This is already the
behavior of non-legacy layout.
The border image outset support in WebRender is going to be removed and
even in versions of WebRender where it still exists, it fails to render
properly.
A border image is a type of border composed of slices of images. The
"outset" of this kind of border is a property in CSS that makes the
border boundaries expand. Previously, the value was passed to WebRender
which would expand the border by this amount and render the images into
the expanded rectangle.
Since this is going to be removed, we handle this property outside of
WebRender. The change is simply to expand the border area by the outset
before calculating the rest of the border values.
This is necessary for the WebRender upgrade.
