This change creates a `constellation_traits` crate. Previously messages
to the `Constellation` were in the `compositing_traits` crate, which
came about organically. This change moves these to a new crate which
also contains data types that are used in both compositing/libservo and
script (ie types that cross the process boundary). The idea is similar
to `embedding_traits`, but this is meant for types not exposed to the
API.
This change allows deduplicating `UntrustedNodeAddress`, which
previously had two versions to avoid circular dependencies.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Simply how `ProgressiveWebMetrics` works:
1. Keep only a single struct instead of one in layout and one script
that both implement the `ProgressiveWebMetrics` trait. Since layout
and script are the same thread these can now just be a single
`ProgressiveWebMetrics` struct stored in script.
2. Have the compositor be responsible for informing the Constellation
(which informs the ScripThread) about paint metrics. This makes
communication flow one way and removes one dependency between the
compositor and script (of two).
3. All units tests are moved into the `metrics` crate itself since there
is only one struct there now.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This will allow removing the dependency of the compositor on
`script_traits`, which should make our internal dependency chain a lot
easier to deal with.
Part of #35984.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Now that Stylo considers `servo` as the default feature, Servo doesn't
need to specify `features = ["servo"]`.
Also use the same crate names as Stylo, rather than renaming them with
`package`.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
This is one of the first big steps toward making the compositor work
per-WebView. It moves the collection of pipelines into the per-WebView
data structure in the compositor as well as the pending paint metrics.
This means that more messages need to carry information about the
WebView they apply to. Note that there are still a few places that we
need to map from `PipelineId` to `WebViewId`, so this also includes a
shared mapping which tracks this. The mapping can be removed once event
handling is fully per-WebView.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Delan Azabani <dazabani@igalia.com>
* Use 2024 style edition
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Reformat all code
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
---------
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
The `OffscreenRenderingContext` does not need to be double-buffered.
Instead, when resizing the framebuffer, create a new one and blit the
old contents onto the new surface. This allows immediately displaying
the contents without having to render paint the WebRender scene one more
time. In addition to speeding up the rendering pipeline, the goal here
is to reduce flickering during resizes (though there is more work to
do).
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This is the first step toward removing `WindowMethods`, which will
gradually be integrated into the `WebView` and `WebViewDelegate`. Sizing
of the `WebView` is now handled by the a size associated with a
`RenderingContext`. `WebView`s will eventually just paint the entire
size of their `RenderingContext`. Notes:
- This is transitionary step so now there is a `WebView::resize` and a
`WebView::move_resize`. The first is the future which will resize the
`WebView` and its associated `RenderingContext`. The second is a
function that the virtual `WebView`s that will soon be replaced by a
the one-`WebView` per `WebView` model.
- We do not need to call `WebView::move_resize` at as much any longer
because the default size of the `WebView` is to take up the whole
`RenderingContext`.
- `SurfmanRenderingContext` is no longer exposed in the API, as a
surfman context doesn't naturally have a size unless a surface is
bound to it.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This moves the GL accelerated media setup out of `RenderingContext`
which prevents making libservo dependo on the Wayland and X11 versions
of surfman explicitly. This support is experimental and (honestly) a bit
broken. I've confirmed that this works as well as it did before the
change.
The main thing here is that the configuration, which currently needs
surfman types, moves to servoshell. In addition:
1. Instead of passing the information to the Constellation, the setup is
stored statically. This is necessary to avoid introducing a
dependency on `media` in `webrender_traits`. It's quite likely that
`media` types should move to the internal embedding API to avoid
this. This is preserved for a followup change.
2. The whole system of wrapping the media channels in an abstract type
is removed. They could be either mpsc channels or IPC channels. This
was never going to work because mpsc channels cannot be serialized
and deserialized with serde. Instead this just uses IPC channels. We
also have other ways of doing this kind of abstraction in Servo so we
do not need another. The `mpsc` version was hard-coded to be
disabled.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Expose two easy-to-use wrappers around `SurfmanRenderingContext` that
make the API simpler to use:
- `WindowRenderingContext`: This `RenderingContext` is a newtype around
`SurfmanRenderingContext` takes a `raw-window-handle` display and window
and creates a full window rendering context.
- `SoftwareRenderingContext`: is wraps `SurfmanRenderingContext` and
adds a swap chain in order to expose a software GL rendering context.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Create a new `RenderingContext` which is used to render to a
`SurfmanRenderingContext`-related offscreen buffer. This allows having a
temporary place to render Servo and then blitting the results to a
subsection of the parent `RenderingContext`.
The goal with this change is to remove the details of how servoshell
renders from the `Compositor` and prepare for the compositor-per-WebView
world.
Co-authred-by: Ngo Iok Ui (Wu Yu Wei) <yuweiwu@pm.me>
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
* layout: Add AxesScrollSensitivity to enable control of scroll in axis
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout_2013: Be compatible with AxesScrollSensitivity
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: update struct AxesScrollSensitivity to euclid::Vector2D
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* display_list: implement From<Overflow> for ScrollSensitivity
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: simplify and reuse scroll related logic
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout_2013: simplify and reuse scroll related logic
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout, layout_2013: revert AxesScrollSensitivity to pair struct
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: Reimport ComputedOverflow as #35103 depends on it
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: Add AxesOverflow to replace PhysicalVec
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: implement scroll of viewport for different axes
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* layout: explicitly handle overflow match
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
* Update components/shared/webrender/Cargo.toml
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
---------
Signed-off-by: Euclid Ye <yezhizhenjiakang@gmail.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
There are a few methods are still difficult to implement without
the help of surfman. To simplify the trait methods, all methods that
return surfman types are removed. They are either handled by
embedders themselves or abstract to simpler types that servo
components need. The most noticeable changes are:
- Methods related to native surface are moved to servo_glue. The
embedder should decide when to remove/replace the surface and it's
outside of servo's scope.
- Methods required by servo media now return exact media types for it.
The other major change is sevevral difficult trait methods that are
reuiqred by WebGL and Servo media have default implementation. So they
can be optional for users to implement.
Signed-off-by: Wu Wayne <yuweiwu@pm.me>
* Remove unused deps
This doesn't seem to remove any deps from the workspace.
Signed-off-by: Jonathan Schwender <schwenderjonathan@gmail.com>
* ohos: Remove gaol dependency
Signed-off-by: Jonathan Schwender <schwenderjonathan@gmail.com>
---------
Signed-off-by: Jonathan Schwender <schwenderjonathan@gmail.com>
When building scroll frames, add a special
`StackingContextContent::Fragment` type for a hit test that covers all
scroll frame contents. This makes it so that you don't have to be
hovering over actual content to scroll the scroll frame.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
There is a command-line argument to override the default window size,
but not one for overriding the default screen resolution. This is
important for testing pages that use screen size to have different
behavior.
In addition to adding the new option this change:
- Renames the `--resolution` command-line argument to `--window-size`
to remove ambiguity with the `--screen-size` argument.
- Passes the screen size as device independent (device pixels scaled by
HiDPI factor) to Servo internals. Not only it make it simpler to pass
the `--window-size` override, it makes more sense. Different screens
can have different HiDPI factors and these can be different from the
scale of the window. This makes the screen HiDPI factor totally
independent of the one that Servo uses for the window.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Instead of exposing many different kinds of messages to the compositor
that are routed through the constellation, expose a single message type
which can be sent across IPC channels. In addition, this IPC channel and
the route to the crossbeam channel with the compositor is created along
with the `CompositorProxy`, simplifying what needs to be passed around
during pipeline initialization.
Previously, some image updates (from video) were sent over IPC with a
special serialization routine and some were sent via crossbeam channels
(canvas). Now all updates go over the IPC channel `IpcSharedMemory` is
used to avoid serialization penalties. This should improve performance
and reduce copies for video, but add a memory copy overhead for canvas.
This will improve in the future when canvas renders directly into a
texture.
All-in-all this is a simplification which opens the path toward having a
standard compositor API and reduces the number of duplicate messages and
proxying that had to happen in libservo.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
* clippy: Fix large size difference between variants in components/shared/webrender/lib.rs
Signed-off-by: chickenleaf <vyaan2000@gmail.com>
* clippy: Fix large size difference between variants of ScriptToCompositorMsg
Signed-off-by: chickenleaf <vyaan2000@gmail.com>
* Remove trailing white space
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
---------
Signed-off-by: chickenleaf <vyaan2000@gmail.com>
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
Instead of exposing many different kinds of messages to the compositor
that are routed through the constellation, expose a single message type
which can be sent across IPC channels. In addition, this IPC channel and
the route to the crossbeam channel with the compositor is created along
with the `CompositorProxy`, simplifying what needs to be passed around
during pipeline initialization.
Previously, some image updates (from video) were sent over IPC with a
special serialization routine and some were sent via crossbeam channels
(canvas). Now all updates go over the IPC channel `IpcSharedMemory` is
used to avoid serialization penalties. This should improve performance
and reduce copies for video, but add a memory copy overhead for canvas.
This will improve in the future when canvas renders directly into a
texture.
All-in-all this is a simplification which opens the path toward having a
standard compositor API and reduces the number of duplicate messages and
proxying that had to happen in libservo.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Instead of a blocking a layout thread on the generation of WebRender
`FontKey`s and `FontInstanceKey`s, generate the keys ahead of time and
send the font data to WebRender asynchronously. This has the benefit of
allowing use of the font much more quickly in layout, though blocking
display list sending itself on the font data upload.
In order to make this work for web fonts, `FontContext` now asks the
`SystemFontService` for a `FontKey`s and `FontInstanceKey`s for new web
fonts. This should happen much more quickly as the `SystemFontService`
is only blocking in order to load system fonts into memory now. In
practice this still drops layout thread blocking to fractions of a
millisecond instead of multiple milliseconds as before.
In addition, ensure that we don't send font data or generate keys for
fonts that are used in layout but never added to display lists. This
should help to reduce memory usage and increase performance.
Performance of this change was verified by putting a microbenchmark
around `FontContext::create_font` which is what triggered font key
generation.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This is done by no longer forwarding compositor-bound messages through
SystemFontService and making `FontContext` non-generic:
- Messages from the `FontContext` to the `Compositor` no longer need to be
forwarded through the `SystemFontService`. Instead send these messages
directly through the script IPC channel to the `Compositor`.
- Instead of adding a mock `SystemFontServiceProxy`, simply implement a
mock `SystemFontService` on the other side of an IPC channel in the
`font_context` unit test. This allows making `FontContext`
non-generic, greatly simplifying the code. The extra complexity moves
into the unit test.
These changes necessitate adding a new kind of `FontIdentifier`,
`FontIdentifier::Mock` due to the fact that local fonts have
platform-specific identifiers. This avoids having to pretend like the
system font service can have web fonts -- which was always a bit of a
hack.
These two changes are combined into one PR because they both require
extensive and similar chages in the font_context unit test which
dependended on the details of both of them.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This changes modifes the way that font data is sent over IPC channels.
Instead of serializing the data or sending it via IPC byte senders, font
data is copied into shared memory and a copy of the handle is sent over
the channel.
There is also the idea of sending the file handle of the on disk data of
system fonts. This could be implemented as a further followup once there
is an abstraction in `ipc-channel` over file handles.
To accomplish this, a `FontData` abstraction is added, which also allows
caching an in-memory shared `Arc<Vec<u8>>` version of the data (neeeded
by some APIs). This could also be a place for caching font tables in the
future.
Finally, the `FontCacheThread` is renamed to the `SystemFontService`
while the proxy for this is now named `SystemFontServiceProxy`.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
This is the first part of cleaning up unused WebRender resources.
Currently this only cleans up web font resources, but a more
full-featured implementation in the future could also clean up unused
system fonts.
Fixes#32345.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
The `gfx_traits` crate is gradually become just about text and fonts and
this is one of the few things there that is standing in the way of this.
Eventually `webrender_traits` will be the general cross-process
compositor API, so this sort of makes sense as well.
This moves mangement of web fonts to the per-Layout `FontContext`,
preventing web fonts from being available in different Documents.
Fixes#12920.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Mukilan Thiyagarajan <mukilan@igalia.com>
* Move WebRender related types to `webrender_traits`
This refactor moves several WebRender related types
from `compositing_traits`, `script_traits` and `net_traits`
crates to the `webrender_traits` crate.
This change also moves the `Image` type and associated
function out of `net_traits` and into the `pixels` crate.
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
* Move `script_traits::WebrenderIpcSender` to `webrender_traits::WebRenderScriptApi`
---------
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
For a long time, `gfx_traits` has held a lot of things unrelated to graphics
and also unrelated to the `gfx` crate (which is mostly about fonts).
This is a cleanup which does a few things:
1. Move non `gfx` crate things out of `gfx_traits`. This is important in
order to prevent dependency cycles with a different integration between
layout, script, and fonts.
2. Rename the `msg` crate to `base`. It didn't really contain anything
to do with messages and instead mostly holds ids, which are used
across many different crates in Servo. This new crate will hold the
*rare* data types that are widely used.
Details:
- All BackgroundHangMonitor-related things from base to a new
`background_hang_monitor_api` crate.
- Moved `TraversalDirection` to `script_traits`
- Moved `Epoch`-related things from `gfx_traits` to `base`.
- Moved `PrintTree` to base. This should be widely useful in Servo.
- Moved `WebrenderApi` from `base` to `webrender_traits` and renamed it
to `WebRenderFontApi`.
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>
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`.