Render AutocompleteSuggestion rows with section headers, favicons,
titles, and secondary text in the child-window popup instead of just
plain strings.
Move the AppKit popup and inline completion onto the shared suggestion
model, and share the base64 PNG decoding helper with the application
menu icon loading path.
Replace the AppKit popover because it becomes key and steals focus from
the location field while suggestions are visible. Use a borderless
child window anchored to the toolbar item instead, and dismiss it on
resize or focus changes so the browser chrome stays predictable.
Teach LibWebView autocomplete to query HistoryStore before falling back
to remote engines and move the wiring out of the AppKit frontend.
Refine matching so scheme and www. boilerplate do not dominate results,
short title and substring queries stay quiet, and history tracing can
explain what the ranking code is doing.
The bytecode interpreter only needed the running execution context,
but still threaded a separate Interpreter object through both the C++
and asm entry points. Move that state and the bytecode execution
helpers onto VM instead, and teach the asm generator and slow paths to
use VM directly.
Qt does not use IOCP's in their underlying Windows event loop
implementation; however, QWinEventNotifier allows us to register a wait
on a process handle that has SYNCHRONIZE access rights. This means an
event will be signalled when that process terminates which emits the
QWinEventNotifier::activated signal.
Co-authored-by: Aliaksandr Kalenik <kalenik.aliaksandr@gmail.com>
This page renders the bookmarks as a tree and hook context menu events
up to the UI's bookmarks bar context menus to allow editing bookmarks.
Users can also drag-and-drop bookmark items around.
Generalize the backing store sharing abstraction into SharedImage, which
represents shared GPU memory independently of Skia and can be used to
share memory between different processes or different GPU contexts.
AppKit doesn't need the bookmarks bar reorganization that Qt needed to
handle more mouse events. But let's rename these member fields to match
what they are now called in Qt.
In addition to Ctrl+PageDown and Ctrl+PageUp, also support Ctrl+Tab and
Ctrl+Shift+Tab to switch between browser tabs. This is the common
behavior in other browsers.
Unfortunately, the NSMenu-based bookmarks implementation cannot support
context menus within arbitrary folders. The NSMenu consumes the right-
click events during its own menu tracking, and we are not able to see
those events from our interface.
This patch replaces the NSMenu for folders with a custom NSPopup. This
implementation makes it much easier to handle context menus.
This allows us to get rid of the annoying Build/.../Lagom directory
and helps to deduplicate a bunch of logic while also drastically
simplifying everything.
This is largely based of off the work done by Andrew Kaster in #5918.
Having this toggle makes much more sense, especially if there will be
more UIs in the future.
Co-authored-by: Andrew Kaster <andrew@ladybird.org>
The parent owner of the TabBar changes as it is added to layout objects
within other widgets. So the parent we set at the beginning is no longer
the parent when we right-click on a tab. We now just store the TabWidget
directly.
QTabWidget's built-in tab bar management made it difficult to properly
position the new tab button, and caused visual artifacts with the
separator line across themes.
Replace QTabWidget with a custom TabWidget that places a QTabBar
alongside an explicit new tab button. This us gives full control over
the tab bar row layout.
We initially create the search field with no frame or size constraints.
The causes AppKit to log verbose warnings at start up. Here, we set an
intrinsic size initially, which becomes overridden once the window has
an actual size.
The HamburgerMenu subclass overrode showEvent() to reposition the menu
using move() after QMenu::popup() had already shown it. On the first
open, rect().width() could return a stale value, causing the menu to
appear detached from the button and truncated to a single item. The
move() call also bypassed Qt's screen boundary adjustments.
Instead, we can replace the HamburgerMenu with a normal QMenu, and use a
custom HamburgerButton class for the menu's tool button. In this class,
we override mousePressEvent() to call popup() with the correct right-
aligned position before the menu is shown. This lets Qt manage sizing
and screen boundaries correctly.
Changing Ctrl+click to a secondary click is incorrect. It prevents
sites from using Ctrl+click themselves. Instead, just maybe open the
context menu in mousedown for primary clicks with Ctrl pressed.
Fixes the autofire shortcut not working in the Humble Mozilla Bundle's
asm.js FTL.
This moves normal/double/triple click checking into WebContent, the
client only has to send a click count in order to activate a double
or triple click in the content. This means that the AppKit UI will no
longer fire multiple double clicks when clicking in place more than 3
times. This matches the behavior of other browsers on macOS.
We will now also fire the click event regardless of whether a dblclick
event will follow, as the spec requires.
When an element is fullscreened, we now hide the application menu and
location bar. The user can hover over the top of the window to make
them show. When fullscreen is requested for the application directly
from the UI, we do not hide the location bar. This matches Safari's
behavior.
We were incorrectly calling presentMetalFrame while force-cpu-painting
was enabled, so updateLayer never had any chance to be called.
However, in addition to that, it turned out that there's a subclass of
CALayer that is used by NSView by default which cannot be created in a
makeBackingLayer override, so updateLayer was never going to be called
anyway. In order to update the layer at the requisite time, a new
LadybirdWebViewContentLayer had to be added to simply call the delegate
on our view.
Setting the contentsRect on the layer was also incorrect, and caused
the contents to shift and stretch until WebContent's backing stores
shrank to fit the new view size.
WebView::FontPlugin was the only implementation of the abstract
FontPlugin base class. Its dependencies (LibGfx, LibCore) are
already visible to LibWeb.
Remove the virtual dispatch by making FontPlugin concrete and
absorbing the WebView::FontPlugin implementation directly.
EventLoopPluginSerenity was the only implementation of the abstract
EventLoopPlugin base class. Its methods simply wrapped Core::EventLoop
calls with GC function unwrapping.
Remove the virtual dispatch by making EventLoopPlugin concrete and
absorbing the EventLoopPluginSerenity implementation directly.
This adds the following features to the Qt UI:
- An "exit fullscreen" button that displays when entering fullscreen
This will disappear after a certain amount of time. Moving the
cursor to the top of the screen will make it reappear.
- Switching tabs exits fullscreen
The look and feel of this is preliminary and should be expanded upon
in the future.
These IPC methods should be expanded in the future to allow WebContent
to specify what UI elements should be kept/removed, for example, the
navigation UI.
The set_viewport_size and set_device_pixel_ratio IPC messages were sent
separately, potentially causing a race condition when the DPR changes
(e.g. moving a window between screens): the DPR message would arrive
and use a stale viewport size, computing a temporarily wrong CSS
viewport. Combine both into a single set_viewport IPC that updates the
device viewport size and DPR together.
This aligns our behaviour closer to other browsers, which
_mostly_ consider file scheme URLs as opaque. For test
purposes, allow overriding this behaviour with a commandline
flag.
Instead of copying the Bitmap that wraps the IOSurface, we can just
present the IOSurface directly. This significantly reduces CPU usage in
the UI process, particularly at high refresh rates such as 120Hz where
it would saturate a full CPU core.
This is done by using CAMetalLayer and blitting the IOSurface to the
next drawable buffer, which handles triple buffering, locking the
IOSurface and vsync automatically. This also allows the Metal HUD to
work, but the only accurate stat is the frame intervals/FPS because
it's in the UI layer, not WebContent. However, that's still useful
to detect frame drops.
When triple clicking on text, we should select the entire paragraph, or
entire line in <input>s and <textarea>s. If the mouse button is held
down and the user starts dragging, the selection expands with additional
paragraphs or lines.
This expands on the work of Kai Wildberger (PR #7681) but was adjusted
for the work that happened previously to support double click + drag
moves and includes triple click support for our Qt UI.
Co-authored-by: Kai Wildberger <kiawildberger@gmail.com>
This introduces a simple FileDownloader to download files in the UI
process from RequestServer. We use this to download the context menu
image - this download is likely to hit the disk cache.
We currently always save screenshots to the Downloads folder. We will
now always ask for a save location.
This will just let an upcoming feature to save images from web pages
behave the same way. We will want the user to be able to choose a file
name, since the file name from the URL might be nonsense or already
exist.
We used to have several of our own window types to close. We now only
have Tab windows. Our delegate handler is invoked even for windows such
file picker modals. The result was that we would close the modal window,
but the application would be left in a modal-active state.
We now behave like Safari. If cmd+w is pressed while a file picker modal
is open, no window is closed, and an error beep is played by AppKit.
Add a thread-safe deferred_invoke() API on WeakEventLoopReference that
queues work onto the owning thread's event queue and wakes that thread
via EventLoopManager hooks. This avoids calling wake() from foreign
threads during teardown.
Implement current_thread_handle()/wake_thread() in each backend and
track per-thread data so handles are validated before waking:
- Unix: wake via per-thread wake pipe
- Windows: wake via thread wake event
- macOS: wake via stored CFRunLoopRef
- Qt: wake via event target or QEventLoop::wakeUp()
- Android: wake via stored ALooper
The base class deferred_invoke() adds work to the thread event queue
and calls wake(), but that's not sufficient for CFRunLoop. We also
need to signal the deferred_source via CFRunLoopSourceSignal() so the
run loop knows to call its perform callback which processes the queue.
Without this signal, deferred invocations posted from other threads
(such as IPC I/O threads) would not be processed, causing deadlocks
when waiting for sync IPC responses.
wake() was using CFRunLoopGetCurrent() which returns the calling
thread's run loop. When called from an I/O thread to wake the main
thread's event loop, this would wake the wrong run loop.
Fixed by using the stored m_impl->run_loop which is captured at
construction time and refers to the event loop's owning thread.
When the user focuses the address bar and presses Escape, the original
URL is now restored. Previously, the address bar would show whatever
text the user had typed, or be empty if they had cleared it.
This matches the AppKit implementation and Chrome/Firefox behavior.
Fixes https://github.com/LadybirdBrowser/ladybird/issues/7264
When the user focuses the address bar and presses Escape, the original
URL is now restored. Previously, the address bar would show whatever
text the user had typed, or be empty if they had cleared it.
Clicking outside the address bar keeps the typed text (matching
Chrome/Firefox behavior) — only Escape restores the original URL.
Fixes https://github.com/LadybirdBrowser/ladybird/issues/7264
The Android build is likely bitrotted, and breaking changes need to be
made to the PlaybackStream interface. This should be reverted and the
implementation updated if the Android build is maintained again.
Once upon a time, we needed the UI-specific event loops outside of the
UI process. This is no longer the case. Let's move the event loops back
to the UI folder to remove the awkward interface library we were left
with.
This patch adds an API for posting a Core::Event::Type to the thread
event queue without requiring a full Core::Event object.
We use this API to avoid heap allocations for every timer and notifier
activation event.
The `Bitmap` type was referring to to its internal pixel format by a
name that represents the order of the color components as they are layed
out in memory. Contrary, the `Color` type was using a naming that where
the name represents the order of the components from most to least
significant byte when viewed as a unsigned 32bit integer. This is
confusing as you have to keep remembering which mental model to use
depending on which code you work with.
To unify the two, the naming of RGBA-like colors in the `Color` type has
been adjusted to match the one from the Bitmap type. This seems to be
generally in line with how web APIs think about these types:
* `ImageData.pixelFormat` can be `rgba-8unorm` backed by a
`Uint8ClamedArray`, but there is no pixel format backed by a 32bit
unsigned type.
* WebGL can use format `RGBA` with type `UNSIGNED_BYTE`, but there is no
such format with type `UNSIGNED_INT`.
Additionally, it appears that other browsers and browser-adjacent
libraries also think similarly about these types:
* Firefox:
https://github.com/mozilla-firefox/firefox/blob/main/gfx/2d/Types.h
* WebKit:
https://github.com/WebKit/WebKit/blob/main/Source/WebCore/platform/graphics/PixelFormat.h
* Skia:
https://chromium.googlesource.com/skia/+/refs/heads/main/include/core/SkColorType.h
This has the not so nice side effect that APIs that interact with these
types through 32bit unsigned integers now have the component order
inverted due to little-endian byte order. E.g. specifying a color as hex
constant needs to be done as `0xAABBGGRR` if it is to be treated as
RGBA8888.
We could alleviate this by providing endian-independent APIs to callers.
But I suspect long-term we might want to think differently about bitmap
data anyway, e.g. to better support HDR in the future. However, such
changes would be more involved than just unifying the naming as done
here. So I considered that out of scope for now.
When a shortcut is an alternative key, macOS first fires a key event for
the primary key. Then if it is not handled, it fires an event for the
alternative key. For example, we see these two events in a row when we
press cmd and =/+:
NSEvent: type=KeyDown flags=0x100108 chars="=" keyCode=25
NSEvent: type=KeyDown flags=0x100108 chars="+" keyCode=24
For dead key processing, we don't handle these events right away. By the
time we get to doCommandBySelector, when we call [NSApp currentEvent],
we see the following events:
NSEvent: type=KeyDown flags=0x100108 chars="=" keyCode=24
NSEvent: type=AppDefined flags=0
The AppDefined event is internally posted by our event loop. So it seems
we cannot rely on currentEvent being the key-down event we are looking
for.
Instead, we can store the key-down event that we last saw.
When a new tab is created, the location bar should automatically
receive keyboard focus so users can immediately start typing a URL.
However, the onURLChange callback was stealing focus back to the
web view when loading the new tab page.
This fix:
1. Calls focusLocationToolbarItem when activating a new tab
2. Prevents onURLChange from stealing focus when loading the new
tab page URL
Fixes#1512
The autocomplete popover on macOS was stealing focus from the location
bar when suggestions were displayed. This change saves the current
first responder before showing the popover and restores it immediately
after, ensuring the user can continue typing without interruption.
In 11b8bbe one thing that was claimed was that we now properly set the
Notifier's actual fd on the NotifierActivationEvent. It turns out that
claim was false because a crucial step was forgotten: actually set the
m_notifier_fd when registering. Despite that mistake, it ultimately was
irrelevant as the methods on NotifierActivationEvent are currently
unused code. We were posting the event to the correct Notifier receiver
so the on_activation was still getting invoked.
Given they are unused, NotifierActivationEvent can be defined the same
way as TimerEvent is, where we just pass the event type enum to the
Event base class. Additionally, NotificationType can be moved to
the Notifier header as this enum is now always used in the context of
creating or using a Notifier instance.
When building Ladybird with the Qt UI framework on Windows, we were
getting the following warning in the CMake target post build step:
"Could not find any translations in <binary_dir>\vcpkg_installed\
x64-windows\translations\Qt6 (developer build?)". We now tell the
deploy script that we have no translations, which removes the warning.
We set bInheritHandles to TRUE for all child processes we spawn. Some
of the types of objects that support handle inheritence is all of the
STD handles (STD_INPUT_HANDLE, STD_OUTPUT_HANDLE, STD_ERROR_HANDLE) and
the console screen buffer. This means if Ladybird and all the child
service processes it launches/communicates with our console apps, only
a single console needs to be allocated and all child process output
their logs to that single console.
The function currently has 2 purposes: (1) To copy dependent dlls for
executables to output binary directory. This ensures that these helper
processes can be ran after a build given not all DLLs from vcpkg libs
get implicitly copied to the bin folder. (2) Allow fully background
and/or GUI processes to use the Windows Subsystem. This prevents
unnecessarily launching a console for the process, as we either require
no user interaction or the user interaction is all handled in the GUI.
When trying to build Ladybird on macOS 14.3, it fails with the error:
```
No visible @interface for 'NSToolbar' declares the selector
'setAllowsDisplayModeCustomization:' (clang arc_may_not_respond)
```
This is caused by macOS < 15 not having the @interface definition for
in NSToolbar for setAllowsUserCustomization:(BOOL).
By dynamically calling the method, we can avoid the error altogether.
We only supported headless clipboard management in test-web. So when WPT
tests the clipboard APIs, we would blindly try to access the Qt app,
which does not exist.
Note that the AppKit UI has no such restriction, as the NSPasteboard is
accessible even without a GUI.
This prevents the AppKit UI from sending unknown MIME types to LibWeb,
where we would previously blindly dereference the result of parsing the
MIME string. We now ignore unknown MIME types as well.
To detect system time zone changes on Windows, the event we need to look
for is WM_TIMECHANGE. The problem is how the callback with said message
actually gets invoked is very particular. (1) We must have an active
message pump (event loop) for the message to ever be processed. (2) We
must be a GUI application as WM_TIMECHANGE messages are seemingly sent
to top level windows only. It doesn't say that in the docs for the
event, but attempts of creating a LibTest-based application with a
message pump and a message only window and never receiving the event
point to that probably being true.
This workaround is built off the fact that Qt's message pump defined
internally in QEventDispatcherWin32::processEvents does in fact receive
WM_TIMECHANGE events, even though it is not exposed as a QEvent::Type.
Given the requirements stated above it makes sense that it works here as
the message pump is executing in a QGuiApplication context. So we use a
native event filter to hook into the unexposed WM_TIMECHANGE event and
forward it along to the on_time_zone_changed() callback.
Note that if a Windows GUI framework is done in the future, we'll have
to re-add support to ensure the TimeZoneWatcher still gets invoked.
Clipboard handling largely has nothing to do with the individual web
views. Rather, we interact with the system clipboard at the application
level. So let's move these implementations to the Application.
This lets us avoid each UI needing to handle link clicks directly, and
lets actions stored in LibWebView avoid awkwardly going through the link
click callbacks to open URLs.
On macOS Tahoe, it is now recommended to show menu item icons. We use
system symbols for this now. Symbols do not have constant variable names
and must be found via the SF Symbols app.
The symbols chosen here were to match Safari as close as possible.
On macOS Tahoe, the zoom button's border is visible when the button
itself is hidden. This feels like a macOS bug, but for now let's hide
the borders as well.
By migrating the debug menu to LibWebView, the AppKit and Qt UIs are now
in sync - the AppKit UI was previously missing some actions.
Further, this inadvertently fixes bugs around applying debug settings to
new web views, especially across site-isolated processes. We were
previously not applying settings appropriately; this now "just works" in
the LibWebView infra.
This migrates all duplicated context menus from the UIs to LibWebView.
The context menu actions are now largely handled directly in LibWebView,
with some UI-specific callbacks added to display e.g. confirmation
dialogs.
Actions that only ever apply to a specific web view are stored on the
ViewImplementation itself. Actions that need to be dynamically applied
to the active web view are stored on the Application.
We currently duplicate a lot of code to handle application/context menus
and actions. The goal here is to hold the data for the menus and actions
in LibWebView. Each UI will then be able to generate menus from the data
on-the-fly.
The structures added here are meant to support generic and checkable
actions, action groups, submenus, etc.
The BUILD_RPATH/INSTALL_RPATH CMake infrastructure is not supported
on Windows, but we want to ensure Ladybird executables are runnable
after the build phase so there can be an efficient dev loop.
lagom_copy_runtime_dlls() can be used by executable targets so all
their dependent dlls are copied to their output directory in their
post build step.
Initializing m_current_screen may not be necessary, but the initial
value of m_preferred_color_scheme is sent to WebContent, so failing to
give it an initializer would cause bugs.
When we manually enter a new URL and hit enter, the web view gets
focused. But when the URL changes for other reasons, such as starting
Ladybird with a URL argument, the location field was still focused.
The faux position we created here is adjusted by the device pixel ratio
later on, which would invoke integer overflow on screens with a DPR
greater than 1.
Instead of creating special data for a mouse move event, let's just add
an explicit leave event handler.
Previously, when the mouse left the WebView, the currently hovered node
would remain hovered (including the scroll bar). This felt a bit awkward
and is not how other browsers behave.
Due to removal of local ca-certificates we need to use system's
certificate. However, on Android it is stored in multiple files.
Ladybird doesn't support multiple certificates yet, so we just
concatenate all of them into one big file.
While some of the changes are due to syntax and have no effect on
the functionality, others are more important.
The major changes are
- Use Painter instead of now removed DeprecatedPainter
- Remove the need of LibArchive by using Java's native zip support
- Have a proper C++23 compiler by updating NDK to r29 beta 2
- Store user data and config in an user accessible way
- Update AGP to 8.11 and update compile target to SDK level 35
These files are lifted from the ladybird-gtk4 repository and adapted
to work with the Qt UI port. They are installed by default on Linux, but
can be installed via a CMake option on other platforms.
Co-Authored-By: Sergey Bugaev <bugaevc@serenityos.org>
Co-Authored-By: Nicolas Ramz <nicolas.ramz@adevinta.com>
Co-Authored-By: Beckett Normington <beckett@b0ba.dev>
Co-Authored-By: Xexxa <93391300+Xexxa@users.noreply.github.com>
We haven't required a local copy of the ca-certificates since switching
to OpenSSL as the backend for TLS. Remove the script to download the
PEM file, and update the tests to use the system's CA certificates.
This replaces the --devtools-port flag with a --devtools flag, which
optionally accepts a port. If the --devtools flag is set, we will now
automatically launch the DevTools server.
In particular, we need to defer creating the process manager until after
we have decided whether or not to create a UI-specific event loop. If we
create the process manager sooner, its event loop signal registration
does not work, and we don't handle child processes exiting.
You would have to just know that you need to define the constructor with
this declaration. Let's allow subclasses to define constructors as they
see fit.
This is causing errors on the WPT runner, which does not have a display
output. To do this requires shuffling around the Main::Arguments struct,
as we now need access to it from overridden WebView::Application methods
after construction.
Now that headless mode is built into the main Ladybird executable, the
headless-browser's only purpose is to run tests. So let's move it to the
testing directory and rename it to test-web (a la test-js / test-wasm).
We currently create a separate headless-browser application to serve two
purposes:
1. Allow headless browsing to take a screenshot of a page or print its
layout tree / internal text.
2. Run the LibWeb test framework.
This patch migrates (1) to the main Ladybird executable. The --headless
flag enables this mode. This matches the behavior of other browsers, and
means we have one less executable to ship at distribution time.
We want to avoid creating too many AppKit / Qt facilities in headless
mode. So this involves some shuffling of application init to ensure we
don't create them until after we've parsed the command line arguments.
Namely, we avoid creating the NSApp in AppKit and QCoreApplication in
Qt. Doing so also requires that we don't create the application event
loop until we've parsed the command line as well, because the loop we
create depends on whether we're creating those UI facilities.
We currently create web views through the headless Application, so that
the Application can store the views for easy iteration/management. We
would like to move HeadlessWebView into LibWebView to make headlessness
a feature of the primary Ladybird binaries. In order to do so, we need
to remove this indirection, as we won't have this test-only Application
class in Ladybird.
In a clean environment, building only headless-browser neglected to
install the resource files needed to run the browser. These were only
installed by the main Ladybird target.
Especially on new setups, it is handy for `headless-browser --run-tests`
to "just work". Anyone using ladybird.py will have LADYBIRD_SOURCE_DIR
set in their environment already.
Currently the window resize events are used to resize the webview.
But when extra window items appear, for example: search or DevTools
Bar then the webview is clipped out of the window. This also happens
when the tab bar is opened and closed. Listening to layout events
resolves these clipping issues completely.
We currently rely on UI-specific APIs to interact with the system
clipboard from AppKit and Qt. We do not have access to these from
headless-browser.
We should ultimately implement these APIs without relying on the UI as
a middle-man. For now, store a clipboard item so that we may exercise
the clipboard WPT tests.
We currently have a single IPC to set clipboard data. We will also need
an IPC to retrieve that data from the UI. This defines system clipboard
data in LibWeb to handle this transfer, and adds the IPC to provide it.
The about:settings page has gotten beefy enough that it is a bit of a
paint to manage in a single HTML file. This patch breaks the settings
into several modules to make this page easier to mantain.
Modules have some nice benefits over classic scripts, such as enabling
strict mode by default, and allowing each module to function without
polluting the global object.
Creating a URL should almost always go through the URLParser to
handle all of the small edge cases involved. This reduces the
need for URL valid state.
This is to prepare for custom search engines. If we use AK::format, it
would be trivial for a user (or bad actor) to come up with a template
search engine URL that ultimately crashes the browser due to internal
assertions in AK::format. For example:
https://example.com/crash={1}
Rather than coming up with a complicated pre-format validator, let's
just not use AK::format. Custom URLs will signify their template query
parameters with "%s". So we can do the same with our built-in engines.
When it comes time to format the URL, we will do a simple string
replacement.
This implements a setting to change the languages provided to websites
from `navigator.language(s)` and the `Accept-Language` header. Whereas
the existing Qt settings dialog allows users to type their language of
choice, this setting allows users to select from a predefined list of
languages. They may choose any number of languages and their preferred
order.
This patch only implements the persisted settings and their UI. It does
not integrate the choses languages into the WebContent process.
This seems to have broken in some recent-ish AppKit update. When we add
the status label to the view hierarchy / change its visibility state,
the NSApp is resetting the cursor to the standard cursor. By overriding
the cursorUpdate method to do nothing, we prevent this from happening.
This removes the old autoplay allowlist file in favor of the new site
setting. We still support the command-line flag to enable autoplay
globally, as this is needed for WPT.
This replaces the existing menu item that would open the Qt settings
dialog. That menu item still exists, but is no longer the default
settings action.
This adds a basic settings page to manage persistent Ladybird settings.
As a first pass, this exposes settings for the new tab page URL and the
default search engine.
The way the search engine option works is that once search is enabled,
the user must choose their default search engine; we do not apply any
default automatically. Search remains disabled until this is done.
There are a couple of improvements that we should make here:
* Settings changes are not broadcasted to all open about:settings pages.
So if two instances are open, and the user changes the search engine
in one instance, the other instance will have a stale UI.
* Adding an IPC per setting is going to get annoying. It would be nice
if we can come up with a smaller set of IPCs to send only the relevant
changed settings.
This adds a WebView::Settings class to own persistent browser settings.
In this first pass, it now owns the new tab page URL and search engine
settings.
For simplicitly, we currently use a JSON format for these settings. They
are stored alongside the cookie database. As of this commit, the saved
JSON will have the form:
{
"newTabPageURL": "about:blank",
"searchEngine": {
"name": "Google"
}
}
(The search engine is an object to allow room for a future patch to
implement custom search engine URLs.)
For Qt, this replaces the management of these particular settings in the
Qt settings UI. We will have an internal browser page to control these
settings instead. In the future, we will want to port all settings to
this new class. We will also want to allow UI-specific settings (such as
whether the hamburger menu is displayed in Qt).
In order to maintain a consistent look and feel between internal about:
pages going forward, let's use a central CSS file to define Ladybird
colors and some common form control styles.
The select dropdown was doing its own ad-hoc method of handling DPR. We
now handle it just like other context menus. Previously, the drop down
in the AppKit chrome was twice as large as it should be.
The "disable DevTools" button looked like a "close this notification"
button to me, and although a tooltip was set, it only showed up
immediately on the AppKit UI and not the Qt version.
This makes the behavior of clicking the disable button a lot clearer by
showing a button with "Disable" as its title.
JsonParser only holds a view into the provided string, the caller must
keep it alive. Though we can actually just use JsonValue::from_string
here instead.
The intent is that this will replace the separate Task Manager window.
This will allow us to more easily add features such as actual process
management, better rendering of the process table, etc. Included in this
page is the ability to sort table rows.
This also lays the ground work for more internal `about` pages, such as
about:config.
This commit:
- Prevents path traversal via the about: scheme
- Prevents loading about:inspector
- Requires about: URIs to be opaque paths
- Prevents crashes with invalid percent encoded paths
For example, consider the following IPC message:
do_something(u64 page_id, String string, Vector<Data> data) =|
We would previously generate the following C++ method to encode/transfer
this message:
void do_something(u64 page_id, String string, Vector<Data> data);
This required the caller to either have to copy the non-trivial types or
`move` them in. In some places, this meant we had to construct temporary
vectors just to send an IPC.
This isn't necessary because we weren't holding onto these parameters
anyways. We would construct an IPC::Message subclass with them (which
does require owning types), but then immediate encode the message to
an IPC::MessageBuffer and send it.
We now generate code such that we don't need to construct a Message. We
can simply encode the parameters directly without needing ownership.
This allows us to take view-types to IPC parameters.
So the above example now becomes:
void do_something(u64, StringView, ReadonlySpan<Data>);
The `cursor` property accepts a list of possible cursors, which behave
as a fallback: We use whichever cursor is the first available one. This
is a little complicated because initially, any remote images have not
loaded, so we need to use the fallback standard cursor, and then switch
to another when it loads.
So, ComputedValues stores a Vector of cursors, and then in EventHandler
we scan down that list until we find a cursor that's ready for use.
The spec defines cursors as being `<url>`, but allows for `<image>`
instead. That includes functions like `linear-gradient()`.
This commit implements image cursors in the Qt UI, but not AppKit.
This information is not particularly interesting, and it can be quite
verbose (such as style-propagation-for-long-continuation-chain.html,
which would log hundreds of lines of "a"s and "b"s).
These were previously skipped on macOS, due to hard to
fix inconsistencies in the exact produced pixels.
Turns out, this is not just a macOS thing. The same sort
of hard-to-spot slight pixel deviations are present on
arm64 Linux as well.
Explicitly link final targets with OpenSSL to ensure that the vcpkg
version is loaded instead of the system one.
Before this change we would inherit `libcrypto.so` and `libssl.so` from
other dependencies, like Qt, that do not have their RPATH rewritten.
This would cause the loader to prefer the system libraries over the
vcpkg ones causing all sorts of version mismatch issues.
The effectiveness of this change can be verified with
`readelf -d ./bin/Ladybird` showing `libcrypto.so` and `libssl.so` as
direct dependencies, before they would not appear. Additionally, `ldd`
will show `libcrypto.so` and `libssl.so` pointing to the vcpkg builds.
Tests have the glob run against the relative path of the test file.
Since this was never set for crash tests the '-f' argument to
headless browser would never match the global against any crash test.
For example, running `alert(1)` will pause the event loop, during which
time no JavaScript should execute. This patch extends this disruption to
microtasks. This avoids a crash inside the microtask executor, which
asserts the JS execution context stack is empty.
This makes us behave the same as Firefox in the following page:
<script>
queueMicrotask(() => {
console.log("inside microtask");
});
alert("hi");
</script>
Before the aforementioned assertion was added, we would execute that
microtask before showing the alert. Firefox does not do this, and now
we don't either.
This brings keyboard shortcuts for the inspector up with common
convention in FF and Chrome: Ctrl+Shift+C now also opens the inspector,
and F12, Ctrl+W, and Ctrl+Shift+I now close the inspector when the
inspector window is focused.
Resolves#972
We now ignore files imported from WPT, if they are in the root `common`
directory, or any directory named `resources`. This matches the
behavior of the WPT test harness.
Some tests take longer than others, and so may want to set a custom
timeout so that they pass, without increasing the timeout for all other
tests. For example, this is done in WPT.
Add an `internals.setTestTimeout(milliseconds)` method that overrides
the test runner's default timeout for the currently-run test.
Without this, a crashing ref test is able to take down the entire
process because of the `VERIFY(!m_pending_screenshot);` in
`take_screenshot()`. The dialog/prompt fields were not causing crashes
but clearing them feels more hygienic.
Add a formatter output to the flake (`nix fmt`), along with moving +
renaming the devshell so it will work by running `nix-shell` in the root
of the project.
This commit adds a "echo_server_port" property to `WebContentOptions`.
Additionally, it makes `Application::web_content_options()` return a
mutable reference instead of a const reference so that we can set the
port value from the fixture.
