In the past, all scroll offsets were reset when navigating between
pages,
and layouts probably didn't preserve their scroll positions between back
and forward operations. Whatever the reason, the compositor was still
trying to
reset these positions right after a load.
This is an issue, because a page can request a scroll during a load, and
this
scroll might be requested from `<iframe>` content. In that case, the
scroll
positions should be the ones that layout reflects and not cleared from
the
Compositor. This reset operation had the potential to stomp on scroll
positions
set during load.
It's quite likely that the Compositor shouldn't be trying to reset them
at all,
so just stop doing that. This removes a source of intermittency when
running
tests that set scroll positions.
Testing: Once #39475 lands after this, its test should always pass.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
This change removes the `DebugOption` (`-Z`) for touch event simulation
and moves the implementation of the feature to servoshell. The resaoning
for this is:
- This is really a servoshell feature and can be implemented on top of
the API. This moves more code out of the already too-complicated
renderer.
- I would like to consolidate `DebugOptions` into a `ServoLogOptions`
to collect all options for configuring Servo logging. This requires
moving away all of the non-logging options.
- Eventually touch event simulation will be able to reuse the fling
implementation from servoshell as we are actually simulating touch
events sent to the `WebView`.
Testing: This changes a conditional feature that's used for manual
debugging.
It is difficult to write tests for this as there are no servoshell tests
that
verify input handling.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Changes function signatures to accept `ClipId`, `ExternalScrollId` and
`ScrollTreeNodeId` instead of `&ClipId`, `&ExternalScrollId` and
`&ScrollTreeNodeId`. This avoids several `&` and `*`.
Testing: not needed, no behavior change.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
This should be the final PR for the Hash Function series that is
trivial.
Of note: I decided to transform `HashMapTracedValues<Atom,..>` to use
FxBuildHasher. This is likely not going to improve performance as Atom's
already have a unique u32 that is used as the Hash but it safes a few
bytes for the RandomState that is normally in the HashMap.
Signed-off-by: Narfinger <Narfinger@users.noreply.github.com>
Testing: Hash function changes should not change functionality, we
slightly decrease the size and unit tests still work.
Signed-off-by: Narfinger <Narfinger@users.noreply.github.com>
Adds a new memory report that aggregates the size of all scroll trees
from within all pipelines for each web view.
Testing: Acessing `about:memory`
Fixes: #38726
---------
Signed-off-by: criskell <96352451+criskell@users.noreply.github.com>
This data structure is now unused after changes to the way that hit
testing works.
Testing: This is just removing dead code so no tests are necessary.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Properly send `mouseleave` events when the cursor moves between
`<iframe>`s. This allows a better handling of cursor changes and status
text updates. Specifically, we do not need to continuously update the
cursor and the value can be cached in the `Document`. In addition,
status updates can now be sent properly when moving focus between
`<iframe>`s.
Note that style updates for `:hover` values are still broken, but less
so than before. Now the hover state on the `Node` is updated, but for
some
reason the restyle isn't taking place properly. This maintains the
status quo as far as behavior goes when hover moves between `<iframe>`s.
This change also adds a helper data structure to `Document` which will
eventually be responsible for event handling.
Testing: Cursor and status change are currently very hard to test as
the API test harness makes this difficult at the moment.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Instead of calculating this value in the compositor, calculate it in
`ScriptThread` now that it is straightforward to get this value from the
layout spatial tree. This allows removing some tricky callback code in
the Compositor.
Testing: This shouldn't change any observable behavior so is covered by
existing tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Instead of using WebRender hit testing to update the cursor, base it on
layout hit tests. This allows removing the majority of WebRender hit
test items and finally opens up the possibility of adding support for
custom cursors. In addition, this change fixes an issue where cursors
were not set properly on areas of the viewport that extended past the
page content.
Testing: This is difficult to test as verifying that the cursor changed
properly is beyond the capabilities of Servo's test harnesses.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Before, the compositor was responsible for doing the hit testing during
input events within a page. This change moves that hit testing to
layout. With this change, epoch mismatches are no longer a bit deal and
we can simply ignore them, as the Constellation and Script will take
care of ignoring hit tests against scroll nodes and browsing contexts
that no longer exist. This means that hit testing retry support can be
removed.
Add the concept of a Script `HitTest` that transforms the coarse-grained
renderer hit test into one that hit tests against the actual layout
items.
Testing: Currently we do not have good tests for verifying the behavior
of
input events, but WebDriver tests should cover this.
Fixes: This is part of #37932.
Fixes: #26608.
Fixes: #25282.
Fixes: #38090.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Co-authored-by: kongbai1996 <1782765876@qq.com>
When calculating the node to world transform for use in bounding box
queries, cache the values of the transform. In addition, when scroll
offsets change, ensure that the cached values are invalided properly.
This change necessitated the storage of children for each node in the
tree, so that we can walk both up and down the tree. The purpose of this
part of the change is to increase performance when doing multiple
queries and prepare the tree for hit testing.
In addition, this change also tries to take into account sticky offsets,
using the algorithm from WebRender to calculate sticky offsets. This is
also going to be important for hit testing.
Testing: Newly passing tests:
- /css/css-position/position-sticky-dynamic-ancestor-001.html
- /css/css-tables/tentative/position-sticky-container.html
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Request a reflow when doing page zoom and only modify the scaling of the
WebView scene after the first root pipeline display list with the new
zoom is ready. In addition:
- store zoom limits in `Scale` types
- send `ViewportDetails` along with the display list so that we can
detect when the root pipeline scale is ready.
Testing: This is quite hard to test as it requires verification that
contents are zoomed appropriately at the right time.
Fixes: #38091.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This change is in accordance with highlighting the usage of various zoom
options.
`pinch_zoom`: `Mobile-style` zoom using pinch gesture
`page_zoom`: `Desktop-style` zoom (`Ctrl` + `+`/`-`)
It just renames the variable `viewport_zoom` to `pinch_zoom` for better
clarity
Testing: Testing not required, as it's just renaming of variable .
---------
Signed-off-by: Shubham Gupta <shubham13297@gmail.com>
Signed-off-by: Xiaocheng Hu <xiaochengh.work@gmail.com>
Co-authored-by: Xiaocheng Hu <xiaochengh.work@gmail.com>
This patch contains 2 components:
1. Instead of passing `self.pinch_zoom_level().get()` while checking
`zoom_result`, initialize it in `combined_magnification`. Ideally, this
part shouldn't have any effect on behavior.
2. Separates the logic for PinchZoom and ViewportZoom. So, when a new
page is opened, it will start with its own viewport zoom scale (rather
than the previous scale multiples).
i.e `self.pinch_zoom_level().get() * 1.0 * magnification`
```rust
let mut combined_magnification = 1.0;
...
ScrollZoomEvent::ViewportZoom(magnification) => {
combined_magnification *= magnification
},
...
let pinch_zoom_result = match self.set_pinch_zoom_level(self.pinch_zoom_level().get() * combined_magnification)
```
Testing: This change adds a new WPT test.
Fixes: #37314
---------
Signed-off-by: Shubham Gupta <shubham13297@gmail.com>
WPT tests are expected to create screenshots as soon as everything is
loaded. If an animation is happening adding the "reftest-wait" class to
the root element is appropriate way to delay the screenshot. Previously,
the
test harness was waiting for all animations to finish, but that is just
leading to many timeouts. Removing that code fixes the timeouts.
Two Servo-specific tests are also updated as they were written with
Servo's previous behavior in mind.
Testing: There are test result updates for this change. Many TIMEOUTS
now either correctly
PASS OR FAIL.
Fixes: #36931.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Similar to #37960, previously, `AvailHeight`, `AvailWidth`, `Height`,
`Width` ask compositor for screen metrics. This PR moves the request to
embedder.
This simplifies code, and reduces workload of compositor, which is
busier most of time.
Testing: No behaviour change. Updated some tests. `Width/Height` matches
other browsers.
---------
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
Previously, `screenX`, `screenY`, `outerHeight`, `outerWidth`, `moveBy`,
`resizeBy` ask compositor for window rectangle, which then return
"inner" rectangle after consulting Embedder.
This PR
1. removes `GetClientWindowRect` from compositor, and directly let
script ask embedder.
2. add `window_size` to `ScreenGeometry`
3. add a lot of docs to `ScreenGeometry`
Testing: `tests\wpt\mozilla\tests\mozilla\window_resizeTo.html` can now
pass for Headed Window.
Fixes: #37824
---------
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
[RootCause]:Tochdown executes normally, but touchup fails during hittest
and does not send a touchup event to the script, causing the script to
save an incorrect number of active_touch_points. The application
receives an incorrect event.touches.length, causing a logic error and
preventing the carousel from sliding.
[Solution]:
When hit test on compositor fails, we also need to send the
EmbedderToConstellationMessage::ForwardInputEvent message to
constellation. In the script's handle_touch_event, if it exits early, we
also need to update active_touch_points.
Testing:
Fixes: #37763
---------
Signed-off-by: kongbai1996 <1782765876@qq.com>
Previously, our Servo-specific spatial tree scroll offsets were opposite
to
that of WebRender and also the web platform. This is due to the fact,
likely, that `winit` wheel directionality is also flipped. This change
has both the Servo spatial tree and the API take offsets that are
consistent with the web.
Any possible changes to the meaning of wheel directionality will be
handled in a followup change.
This is a breaking change to the Servo API.
Testing: This change updates unit tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
The compositor always does the same thing with these events regardless
of the phase, so I think it is completely unecessary.
Testing: This shouldn't change behavior at all, so is covered by
existing tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Currently, the hover state will stay when the mouse moves out of the
webview, this PR fixes it
Testing: Hover on the `About` on servo.org, and then move the mouse up
to the browser UI, see the hover state resets
Signed-off-by: Tony <legendmastertony@gmail.com>
Instead of taking into account whether the entire WebView starts
animating, always send an animation tick when a pipeline moves from the
"not animating" to "animating" state. It could be that the WebView was
animating, but not painting if the animation was not producing display
lists. In that case, the required tick would never come, because it is
sent after a repaint.
Testing: Added a new WPT test.
Fixes: #37458.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Previously, the Constellation would immediately ask the Compositor to
shut down a pipeline, even before the ScriptThread finished shutting it
down. This meant that the Compositor might remove a Pipeline and then
re-add it if the ScriptThread sent a Pipeline-related message (such as a
new display list) in the meantime.
This change makes it so that the Compositor waits for both the
Constellation and the ScriptThread to finish shutting down a Pipeline
before removing its data. In addition, the Constellation no longer
synchronously waits on the Compositor when shutting down Pipelines. This
was important when the Compositor would talk to the ScriptThread
directly, but isn't necessary any longer.
Testing: This is very hard to test, because it depends on the creation
and destruction of many iframes and the particular timing of of all
the messaging between Servo bits. That said, this was tested manually
by observing the completion of Speedometer 2.1.
Fixes: #37458.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Revert "Remove duplication in dispatch_pending_point_input_events"
(https://github.com/servo/servo/pull/37426) which causes a serious bug.
Signed-off-by: batu_hoang <longvatrong111@gmail.com>
There are currently five places that scroll offsets are stored:
- DOM: A set of scroll offsets used for script.
- Layout: An array of scroll offsets that is used for tracking
layout-side scroll offsets.
- Layout: The scroll offsets stored in the `ScrollTree`. These are
currently unset and unused.
- Compositor: The scroll offsets stored in the `ScrollTree` mirrored
from layout.
- WebRender: The scrolled offsets stored in the WebRender spatial tree.
This change is the first step in combining the first three into the
layout `ScrollTree`. It eliminates the extra array of scroll offsets
stored in layout in favor of the storing them in the `ScrollTree`. A
followup change will eliminate the ones stored in the DOM.
- In addition the `ScrollState` data structure is eliminated as these
are
now stored in a `HashMap` everywhere when passing them via IPC.
- The offsests stored in layout can now never scroll past the boundaries
of the scrolled content.
Testing: This should not change behavior and is thus covered by existing
WPT tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: stevennovaryo <steven.novaryo@gmail.com>
The compositor was accepting scroll offsets from the ScriptThread
without checking their boundaries. In some cases this could cause a
temporary discrepancy with the rendered scroll offset. This change makes
it so that all offset updates for scroll ayers in the compositor do not
scroll past the scroll boundaries of the node.
Testing: Two new tests pass with this change:
- `/css/css-position/sticky/position-sticky-left-003.html`
- `/css/css-position/sticky/position-sticky-top-003.html`
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Use own `ViewportDescription` to `clamp()` initial_scale.
If use `self` it will use current module `ViewportDescription`, which is
old one because new `ViewportDescription` is not set yet.
Testing: Tested Locally
Fixes: #37338
Signed-off-by: Shubham Gupta <shubham13297@gmail.com>
Follow up to [hit_test failed occasionally when the touch event is
sent](https://github.com/servo/servo/issues/35788#top) and
https://github.com/servo/servo/issues/36676#issuecomment-2882917136,
this PR adds a retry for hit tests with expired epoch in result.
Hit tests with outdated epoch mean it is too early to perform the hit
test.
Solution: retry hit test for the event on the next webrender frame.
The retry should guarantee that:
- Keep the correct order of events
- Retry time is not too long
Test cases: `./mach test-wpt --product servodriver -r
tests\wpt\tests\webdriver\tests\classic\element_click\events.py`
cc: @xiaochengh , @yezhizhen
---------
Signed-off-by: batu_hoang <longvatrong111@gmail.com>
Implement missing synchronization in `dispatch_actions` of `WebDriver`.
https://w3c.github.io/webdriver/#dispatching-actions
> The user agent event loop has spun enough times to process the DOM
events generated by the last invocation of the >[dispatch tick
actions](https://w3c.github.io/webdriver/#dfn-dispatch-tick-actions)
steps.
- Add a way for `ScriptThread` to notify `WebDriver` about the
completion of input commands.
- Add a `webdriver_id` field for `InputEvent`. `ScriptThread` uses it to
distinguish WebDriver events and sends notification.
Tests:
`./mach test-wpt --product servodriver -r
tests\wpt\tests\webdriver\tests\classic\element_click\events.py` pass if
`hit_testing` pass. Check
[issue](https://github.com/servo/servo/issues/36676#issuecomment-2882917136)
cc: @xiaochengh
---------
Signed-off-by: batu_hoang <longvatrong111@gmail.com>
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
When multiple WebViews are updating scroll events, instead of taking the
list of `WebView`s to avoid a double-borrow, batch scroll events into a
single transaction. This should make processing slightly more efficient
and avoids having to take the vector of WebViews.
Testing: No behavior change here and this aspect of WebView interaction
is untestable currently.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
After `InputEvent::Touch` processed by Script, it sends
`TouchEventProcessed` back to Constellation, which goes through hit-test
etc. and loops back to Script again. However,
`WebViewRenderer::simulate_mouse_click` should not send
`MouseButtonAction::Click` due to #36413
Testing: Manually tested by converting mouse to touch
cc @jdm @xiaochengh
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
Previously, when processing animations, the compositor would sent a tick
message to each pipeline. This is an issue because now the
`ScriptThread` always processes rendering updates for all `Document`s in
order to ensure properly ordering. This change makes it so that tick
messages are sent for an entire WebView. This means that each
`ScriptThread` will always receive a single tick for every time that
animations are processed, no matter how many frames are animating. This
is the first step toward a refresh driver.
In addition, we discard the idea of ticking animation only for
animations and or only for request animation frame callbacks. The
`ScriptThread` can no longer make this distinction due to the
specification and the compositor shouldn't either.
This should not really change observable behavior, but should make Servo
more efficient when more than a single frame in a `ScriptThread` is
animting at once.
Testing: This is covered by existing WPT tests as it mainly just improve
animation efficiency in a particular case.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
There is a `WebView` in libservo (new) and a `WebView` in compositing
(old). Nowadays, the "real" `WebView` is the one in the libservo. The
`WebView` in `compositing` is really about rendering the contents of a
`WebView` from libservo. In addition there is also a trait exposed by
the compositor called `RendererWebView` which is a way for the
compositor to talk to libservo without a circular dependency.
This changes does some renames to make things clearer and so that there
is One Less WebView™:
- `compositing::WebView` -> `compositing::WebViewRenderer` (this is the
same kind of naming as `ServoRenderer`).
- `compositing::RendererWebView` -> `compositing::WebViewTrait`
Testing: This is just a couple renames so should be covered by existing
tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
2025-04-21 08:47:39 +00:00
Renamed from components/compositing/webview.rs (Browse further)
