Before all timers were managed by the Constellation process, meaning
that they had to trigger IPC calls to be scheduled and fired. Currently,
timers are only used in the `ScriptThread`, so it makes sense that they
are per-process.
This change restores the timer thread functionality that existed before
avoided entirely. Completion is done using a callback that is sent to
the timer thread similarly to how fetch is done. This allows reusing the
existing task queue without making any new channels.
Fixes#15219.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Block layout uses some heuristics to guess whether margins are separated
by clearance and then don't collapse. These heuristics now take the
min-content, max-content, fit-content and stretch sizing keywords into
account.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
In order to compute the inline min-content and max-content contributions
of an anonymous block, we were finding its min-content and max-content
inline size with a SizeConstraint coming from the block size of the box.
However, anonymous blocks do not establish a containing block for their
contents, so this patch uses a SizeConstraint from the block size of the
containing block.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
The nightly build jobs run on Ubuntu 20.04 to ensure the glibc version
is compatible with WPT.fyi runners. But this version of Ubuntu has an
older clang version (12) that is not compatible with mozjs.
This patch makes the nightly job for Linux explicity install Clang 14
using the KyleMayes/install-llvm-action@v2 action. The regular Linux job
still uses system Clang installation as there are failures when using
the github action.
Fixes#34713.
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
- The default value for SwapInterval is 1,
so setting it to 1 changes nothing
- We don't clear the screen immediately anymore, which was the only
useage of the egl bindings
this saves us from duplicated EGL bindings.
Signed-off-by: Jonathan Schwender <schwenderjonathan@gmail.com>
If a table element had e.g. `width: 0px`, we were assuming that this was
its intrinsic min-content and max-content contributions.
However, tables are always at least as big as its min-content size, so
this patch floors the intrinsic contributions by that amount.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
This change creates a new struct `IFrameCollection` that is used to
cache the list of `<iframe>`s in a `Document` as long as the
`Document`'s DOM has not changed. This prevent constantly iterating the
entire DOM during *update the rendering*, which runs up to 60 times per
second as well as for other operations.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
The new images published in servo/ci-runners#12 should have `uv`
installed already and the initial build of servo triggered during the
base image construction will force the installation of the Python
version mentioned at the time of the image construction (3.12). When
.python-version changes, we can no longer use the .venv baked into the
image and must recreate the it to avoid activating the environment.
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
When an element is a shadow root, lay out the shadow root elements
instead of the non-shadow children.
This fixes some tests and introduces some failures, due to bugs in the
Shadow DOM implementation. In general, this is very low impact as the
Shadow DOM is still disabled by default. At least this gets elements
rendering when the preference is turned on though.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This reverts commit bc0c8366f8,
relanding commit 11424f90b3.
Reason: it wasn't the cause behind #34688.
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
in each layout logic, in order to correctly resolve sizing keywords.
This patch adds a new `Sizes` struct which holds the preferred, min and
max sizing values for one axis, and unifies the logic to resolve the
final size into there.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
* Include unimplemented console methods in idl file
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Fix console.assert signature
The condition is optional and there can be multiple messages.
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Implement console.trace
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* ./mach fmt
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Log stack trace when calling console.trace
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Update wpt expectations
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Include line/column info in console.trace logs
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Move option out of constant
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Update mozjs
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
---------
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Don't unnecessarily clone responses twice when sending them to devtools
Signed-off-by: Simon Wülker <simon.wuelker@arcor.de>
* Store http response reason instead of inferring it later
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>
Consider:
```html
<div style="position: relative; width: 50px; height: 50px; border: solid; margin: 5px">
<div style="position: absolute; top: 0; bottom: 0; height: max-content">
<canvas width="25" height="25" style="background: cyan; height: 100%"></canvas>
</div>
</div>
```
In order to determine the inline min/max-content sizes, we need a
tentative block size as the input, which only takes extrinsic values
into account.
In this case `height: max-content` is intrinsic, so we were treating it
as `height: initial`, which would behave as a definite `height: stretch`.
Therefore, the canvas was able to resolve its percentage.
However, it seems weird to treat an explicitly intrinsic keyword in an
extrinsic way, and Blink doesn't do it. So now we treat the tentative
block size as indefinite, therefore the percentage behaves as auto.
This adds a new test, we were previously failing 6 subtests, now only 3.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
We have been having intermittent issues with sccache since
around the recent v0.9 release of sccache which only
appear when building mozjs from source.
Due to recent changes with readablestreams, we are not using
the prebuilt mozjs artifacts anymore and always building from
source, causing these problems to regularly block the merge
queue. Bumping SCCACHE_GHA_VERSION helps but has been observed
to only be a temporary solution until the next breakage.
Closes#34571
Signed-off-by: Jonathan Schwender <schwenderjonathan@gmail.com>