The issue noted in the removed comment no longer seems to apply. The URL
loaded here is triggered before WebDriver even issues the POST /session
command.
When we launch Ladybird, we currently:
1. Create a BrowserWindow, whose constructor navigates to the configured
new tab page URL.
2. Then navigate to either "about:blank" or whatever URL is provided via
the command line.
This patch removes step 2 and forwards the URL from the command line (if
any) to BrowserWindow. BrowserWindow's constructor then either navigates
to that URL or the new tab page URL.
Currently, if the JS console is open and tied to the last opened tab in
the browser window, it will prevent the main process from exiting when
the last tab is closed. This change explicitly closes that tab before
closing the window (if it's the last tab), allowing Qt to delete the
Tab object.
This will help a lot with developing chromes for different UI frameworks
where we can see which helper classes and processes are really using Qt
vs just using it to get at helper data.
As a bonus, remove Qt dependency from WebDriver.
With the CMake build, the ladybird binary ends up in
$build/bin/ladybird.app/Contents/MacOS. Which is a bit unfortunate for
a few reasons, but the main one is that --enable-lagom-networking
doesn't work with ./Meta/serenity.sh run lagom ladybird.
Let's copy the file to the expected location. There's very likely better
solutions we can use in the future.
This object is available as `window.internals` (or just `internals`) and
is only accessible while running in "test mode".
This first version only has one API: gc(), which triggers a garbage
collection immediately.
In the future, we can add more APIs here to help us test parts of the
engine that are hard or impossible to reach via public web APIs.
The Qt docs are not that clear, but to get the number of steps the
wheel was scrolled you divide by 120 (which when multiplied by
wheelScrollLines() gives the scroll offset).
Documentation says that pixelDelta() is not reliable across platforms
so always using angleDelta() should produce more predictable scrolling
behaviour.
Rather than directing mouse wheel events to the QAbstractScrollArea,
send them to the WebContent process, allowing it to determine if they
should be consumed by the element or used for page scrolling.
The implementation of this plugin is meant to eventually replace all
current audio plugins in Ladybird. The benefits over the current Qt-
based audio playback plugin in Ladybird are:
- Low latency: With direct access to PulseAudio, we can ask for a
specific latency to output to allow minimal delay when pausing or
seeking a stream.
- Accurate timestamps: The Qt audio playback API does not expose audio
time properly. When we have access directly to PulseAudio APIs, we can
enable their timing interpolation to get an accurate monotonically-
increasing timestamp of the playing audio.
- Resiliency: With more control over how the underlying audio API is
called, we have the power to fix most bugs we might encounter. The
PulseAudio wrappers already avoid some bugs that occur with QAudioSink
when running through WSLg.
Previously, a QTimer was used to start processing of our event queue in
the main Qt event loop. Unfortunately, QTimers are not thread-safe, and
disallow starting of a timer from a different thread than it was
created in.
Instead, use a dummy QObject to post a custom QEvent to the main loop
from whatever thread we like, and process our event queue when it is
received by our dummy object.
The logic for `EventLoopImplementationQt::pump()` caused calls to
`EventLoop::pump(PumpMode::DontWaitForEvents)` to not consume events
posted to the Qt event queue. This caused the window to be unresponsive
even when polling the event loop, if waiting was not desired.
If a URL is not valid we try navigating to https:// + the url. It's
better to ask AK::Url if it thinks the url is valid than put a big list
of prefixes here, with this obscure protocols like Gemini are now
recognised and with `--enable-lagom-networking` can be viewed in
Ladybird (thanks to #2218).
Now that all the classes for Ladybird are in the Ladybird namespace, we
don't need them named as Ladybird::FooBarLadybird. For the Qt-specific
classes, we can tack on a Qt at the end for clarity, but FontPlugin and
ImageCodecPlugin no longer have anything to do with Qt.
We were super inconsistent about this, with most "new" classes living in
the Ladybird namespace, while "old" ones were in the global namespace,
or even sitting in the Browser namespace.
Ladybird's RequestServer needs to depend on its generated IPC header
files to ensure they are generated before RequestServer is compiled,
which we currently bundle into LibWebView.
LibTLS still can't access many parts of the web, so let's hide this
behind a flag (with all the plumbing that entails).
Hopefully this can encourage folks to improve LibTLS's algorithm support
:^).
Re-organize our helper files here a bit, to make a clearer distinction
between Qt-specific helpers and generic non-serenity helpers.
A future commit will move Lagom specific code from LibSQL to ladybird
as well, so that we can see about future generic apis for spawning
helper procesess.
The spawn_helper_process method was introduced together with
get_paths_for_helper_process but was only ever used briefly to spawn
WebContent. Other helper processes (SqlServer, headless_browser etc)
are either execed or spawned with their own helpers & custom arguments.
The Qt relationship was removed in de31a8a4, so let's acknowledge that
and make it clearer to potential non-Qt ports that this file is usable
by them :^).
Rather than set a hacky scale factor on our QIconEngine, just set the
icon size for the toolbar icons. This avoids the icons being
unnecessarily scaled in other places (i.e. the context menu).
In order to follow spec text to achieve this, we need to change the
underlying representation of a host in AK::URL to deserialized format.
Before this, we were parsing the host and then immediately serializing
it again.
Making that change resulted in a whole bunch of fallout.
After this change, callers can access the serialized data through
this concept-host-serializer. The functional end result of this
change is that IPv6 hosts are now correctly serialized to be
surrounded with '[' and ']'.
Now that we no longer use QFont from LibWeb, we can also stop using
QGuiApplication in the WebContent process entirely.
This removes a whole bunch of unnecessary work from the event loop,
and also allows nice things like running headless-browser while
*actually* headless. :^)
Rather than render the icons to a 16x16 bitmap, keep them as vector
graphics and render them on request. This keeps the icons crisp on high
DPI displays.
The Qt StyleHint system didn't work on X11 anyway, and we ended up with
the default UI font being used for both `serif` and `sans-serif`.
Instead of asking Qt for something it can't do properly, let's just grab
the first available font from our list of fallbacks. This should give us
better results everywhere.
WebView::ViewImplementation now remembers which JS interpreter it
started with, and uses the same setting if the WebContent process
crashes and we have to spawn a new one.
In some places in the code QAction instances are added to a parent
directly on instantiation. In the case they are not, Valgrind
reports them as leaking.
We don't need the extra gradle files in our sources, the Qt CMake
integration will generate suitable ones for us.
Make sure that assets is always a folder, so that we can get the proper
layout for the ladybird-assets.tar.gz and CMake doesn't create a gzip
file with the name "assets".
Fix up the AndroidPlatform file and make sure it's linked into all the
applications that need it. Also make sure to copy all the application
shared libraries into the ladybird APK so that when we make them into
proper Services, the libs are already there.
This implementation detail of audio support in ladybird is a QObject
that needs moc'd by the moc tools. Putting it in its own file follows
the pattern we have for all the other QObjects in Ladybird.
Doing this removes the qt6-svg dependency and allows our rasterizer to
be used for these little icons (and happens to be a fair bit smaller
than the old SVGs).
We currently drop events which do not have text associated with them.
This prevents e.g. arrow keys from being able to be handled by web
elements. We now match Browser's behavior on Serenity, where these key
events are already propagated.
This change was a long time in the making ever since we obtained sample
rate awareness in the system. Now, each client has its own sample rate,
accessible via new IPC APIs, and the device sample rate is only
accessible via the management interface. AudioServer takes care of
resampling client streams into the device sample rate. Therefore, the
main improvement introduced with this commit is full responsiveness to
sample rate changes; all open audio programs will continue to play at
correct speed with the audio resampled to the new device rate.
The immediate benefits are manifold:
- Gets rid of the legacy hardware sample rate IPC message in the
non-managing client
- Removes duplicate resampling and sample index rescaling code
everywhere
- Avoids potential sample index scaling bugs in SoundPlayer (which have
happened many times before) and fixes a sample index scaling bug in
aplay
- Removes several FIXMEs
- Reduces amount of sample copying in all applications (especially
Piano, where this is critical), improving performance
- Reduces number of resampling users, making future API changes (which
will need to happen for correct resampling to be implemented) easier
I also threw in a simple race condition fix for Piano's audio player
loop.
For example, on https://xboygeniusx.bandcamp.com/album/the-record, a
song with a duration of 03:52 would actually complete in 03:33 on my
machine. This issue only affects Ladybird on Lagom; on Serenity, we
already take the entire 03:52 to play the song.
It's currently possible to seek to the total sample count of an audio
loader. We must limit seeking to one less than that count.
This mistake was duplicated in both AudioCodecPluginSerenity/Ladybird,
so the computation was moved to a helper in the base AudioCodecPlugin.
When the default audio device changes on the host, it's convenient to
automatically switch to that device rather than needing to reload the
page to update.
We are currently forcing audio to play with a sample size of 16 bits. We
are also feeding the output audio device a hard-set amount of samples
without considering the actual size of its sample buffer. This would
cause a wide array of issues when playing audio elements. On my Linux
machine, we would hear some cracking; on my macOS machine, audio was
quite garbled.
We now write samples of the size requested by the output audio device.
We also limit the samples we provide to the audio device to however many
bytes are available in its buffer.
The main thread in the WebContent process is often busy with layout and
running JavaScript. This can cause audio to sound jittery and crack. To
avoid this behavior, we now drive audio on a secondary thread.
Note: Browser on Serenity uses AudioServer, the connection for which is
already handled on a secondary thread within LibAudio. So this only
applies to Lagom.
Rather than using LibThreading, our hands are tied to QThread for now.
Internally, the Qt media objects use a QTimer, which is forbidden from
running on a thread that is not a QThread (the debug console is spammed
with messages pointing this out). Ideally, in the future AudioServer
will be able to run for non-Serenity platforms, and most of this can be
aligned with the Serenity implementation.
The data we want to send out of the WebContent process is identical for
audio and video elements. Rather than just duplicating all of this for
audio, generalize the names used for this IPC for all media elements.
This also encapsulates that data into a struct. This makes adding new
fields to be sent much easier (such as an upcoming field for muting the
element).
This makes pages that use CSS rules like '@media (max-device-width:
600px)' render more correctly.
Without this change device-width and height queries would return 0.
This creates (and installs upon WebContent startup) a platform plugin to
play audio data.
On Serenity, we use AudioServer to play audio over IPC. Unfortunately,
AudioServer is currently coupled with Serenity's audio devices, and thus
cannot be used in Ladybird on Lagom. Instead, we use a Qt audio device
to play the audio, which requires the Qt multimedia package.
While we use Qt to play the audio, note that we can still use LibAudio
to decode the audio data and retrieve samples - we simply send Qt the
raw PCM signals.
On macOS, CMake incorrectly tries to add and/or remove rpaths from files
that it has already processed when it performs installation. Setting the
rpaths during the build process ensures that they are only set once, and
as a bonus, makes installation slightly more performant.
Fixes#10055.
Currently, we only look at the relative path `./{helper}/{helper}`,
which fails if the working directory is not the same as the directory
where the ladybird binary lives.
This will make it a lot easier to understand what went wrong, especially
when the failure occurs on CI but not at home.
And of course, use LibDiff to generate the diff! :^)
Instead of starting a new headless-browser for every layout & text test,
headless-browser now gets a mode where it runs all the tests in a single
process.
This is massively faster on my machine, taking a full LibWeb test run
from 14 seconds to less than 1 second. Hopefully it will be a similarly
awesome improvement on CI where it has been soaking up more and more
time lately. :^)
This looks a lot more "at home" than usual pixel art logo on
non-SerenityOS systems. :^)
Also, stop using site favicons as the app icon as that made it
annoyingly hard to find Ladybird in task switchers sometimes.
This allows us to create "text tests" in addition to "layout tests".
Text tests work the same as layout tests, but dump the document content
as text and exit upon receiving the window "load" event.
Previously, calling `.right()` on a `Gfx::Rect` would return the last
column's coordinate still inside the rectangle, or `left + width - 1`.
This is called 'endpoint inclusive' and does not make a lot of sense for
`Gfx::Rect<float>` where a rectangle of width 5 at position (0, 0) would
return 4 as its right side. This same problem exists for `.bottom()`.
This changes `Gfx::Rect` to be endpoint exclusive, which gives us the
nice property that `width = right - left` and `height = bottom - top`.
It enables us to treat `Gfx::Rect<int>` and `Gfx::Rect<float>` exactly
the same.
All users of `Gfx::Rect` have been updated accordingly.
The goal here is to reduce the amount of WebContent client APIs that are
duplicated across every ViewImplementation. Across our three browsers,
we currently:
Ladybird - Mix some AK::Function callbacks and Qt signals to notify
tabs of WebContent events.
Browser - Use only AK::Function callbacks.
headless-browser - Drop most events on the floor.
Instead, let's only use AK::Function callbacks across all three browsers
to propagate events to tabs. This allows us to invoke those callbacks
directly from LibWebView instead of all three browsers needing to define
a trivial `if (callback) callback();` override of a LibWebView virtual
function. For headless-browser, we can simply not set these callbacks.
As a first pass, this only converts WebContent events that are trivial
to this approach. That is, events that were simply passed onto the tab
or handled without much fuss.
This is to match Browser, where ownership of all "subwidgets" is placed
on the tab as well. This further lets us align the web view callbacks to
match Browser's OOPWV as well, which will later let us move them into
the base LibWebView class.
