The identifier "Protocol" is claimed by Objective-C and Swift for use
by the language's built-in protocol conformance feature, which is
similar to Rust traits or Java interfaces.
Rename LibProtocol -> LibRequests, and its namespace from Protocol to
Requests to accomodate this.
We use instances of `Gfx::Bitmap` to move pixel data all the way from
raw image bytes up to the Skia renderer. A vital piece of information
for correct blending of bitmaps is the alpha type, i.e. are we dealing
with premultiplied or unpremultiplied color values?
Premultiplied means that the RGB colors have been multiplied with the
associated alpha value, i.e. RGB(255, 255, 255) with an alpha of 2% is
stored as RGBA(5, 5, 5, 2%).
Unpremultiplied means that the original RGB colors are stored,
regardless of the alpha value. I.e. RGB(255, 255, 255) with an alpha of
2% is stored as RGBA(255, 255, 255, 2%).
It is important to know how the color data is stored in a
`Gfx::Bitmap`, because correct blending depends on knowing the alpha
type: premultiplied blending uses `S + (1 - A) * D`, while
unpremultiplied blending uses `A * S + (1 - A) * D`.
This adds the alpha type information to `Gfx::Bitmap` across the board.
It isn't used anywhere yet.
This removes the --enable-callgrind-profiling flag, and replaces it with
a --profile-process=<process-name> flag. For example:
ladybird --profile-process=WebContent
ladybird --profile-process=RequestServer
This allows profiling any helper process with callgrind.
This removes the --debug-web-content flag, and replaces it with a
--debug-process=<process-name> flag. For example:
ladybird --debug-process=WebContent
ladybird --debug-process=RequestServer
This allows attaching gdb to any helper process.
Currently, if we want to add a new e.g. WebContent command line option,
we have to add it to all of Qt, AppKit, and headless-browser. (Or worse,
we only add it to one of these, and we have feature disparity).
To prevent this, this moves command line flags to WebView::Application.
The flags are assigned to ChromeOptions and WebContentOptions structs.
Each chrome can still add its platform-specific options; for example,
the Qt chrome has a flag to enable Qt networking.
There should be no behavior change here, other than that AppKit will now
support command line flags that were previously only supported by Qt.
The TabBar itself does not stretch the entire width of the TabWidget,
because it leaves space for the width of the new-tab button. So, we
manually tell Qt to paint the TabBar's background into the gap.
Fixes#768.
Makes use of the fact that QTabWidget automatically reserves space
for "corner widgets", and gives us the ability to override their
location with a custom style.
Also, determine the height of the new-tab button using the height of the
tab bar, instead of a hard-coded 32px which was too tall on MacOS.
The default limit (at least on Linux) causes us to run out of file
descriptors at around 15 tabs. Increase this limit to 8k. This is a
rather arbitrary number, but matches the limit set by Chrome.
Skia painter is visibly faster than LibGfx painter and has more complete
CSS transforms support. With this change:
- On Linux, it will try to use Vulkan-backend with fallback to
CPU-backend
- On macOS it will try to use Metal-backend with fallback to
CPU-backend
- headless-browser always runs with CPU-backend in layout mode
If no header includes the prototype of a function, then it cannot be
used from outside the translation unit it was defined in. In that case,
it should be marked as `static`, in order to avoid possible ODR
problems, unnecessary exported symbols, and allow the compiler to better
optimize those.
If this warning triggers in a function defined in a header, `inline`
needs to be added, otherwise if the header is included in more than one
TU, it will fail to link with a duplicate definition error.
The reason this diff got so big is that Lagom-only code wasn't built
with this flag even in Serenity times.
Enforce the use of the CPU backend in test mode to ensure that ref-tests
produce consistent results across different computers, as this
consistency cannot be achieved with the GPU backend.
This resolves a bug where if you opened a link in a new tab and quickly
went back to the original, the navigation buttons would update for the
new page shortly after.
Previously, the browser would crash when opening a task manager window
with the `--enable-qt-networking` flag set because we were passing the
default WebContentOptions to the underlying WebContentView.
This change adds a `--certificate` option to both WebContent and
WebWorker, which allows one or more custom root certificate paths to be
specified. Certificates are then loaded from these paths when Qt
networking is used.
This allows WPT tests that require a https connection to be run locally
with Qt networking.
We were not doing anything with the SwipeRefreshLayout, and it
interfered with touch events. We may want to bring this back at some
point? But probably only after we handle Android scrolling correctly.
This change adds a `--use-lagom-networking` flag to the WebWorker
process. Qt networking is used if this flag isn't passed. The flag is
passed the UI launches the WebWorker process unless the Qt chrome is
being run with the `--enable-qt-networking` flag.
Using Qt networking when running web platform tests improves
performance significantly. The time to run the subset of tests we run
on CI drops from 21.9 seconds to 8.2 seconds on my machine.
This commit replaces all TLS connection code with wolfssl.
The certificate parsing code has to remain for now, as wolfssl does not
seem to have any exposed API for that.
Previously, new tabs always had the `auto` color scheme, regardless of
what the user has selected before.
Replace the 3 individual slots with a `set_preferred_color_scheme`
method.
LibWebView uses the Application object in WebContentClient.cpp, so we
need one to exist. It doesn't really do anything, since we never add
processes to it, but it just has to exist.
This updates the Android WebViewImplementationNative to match changes
made in commit 5285e22f2a.
We remove the async_set_viewport_rect call because the parent
ViewImplementation::handle_resize call does that for us.
The current version of the Android NDK uses Clang 17, which uses the
name c++2b instead of c++23.
This is the same flag we use in Meta/gn/build/BUILD.gn for macOS.
It looks like some things have moved around since the last time the
Android build worked. So, update the incorrect paths to point to where
they should.
This ensures that vcpkg downloads and builds all dependencies for
Android. We add it as a CMAKE_TOOLCHAIN_FILE that then chainloads the
Android NDK's toolchain file, as per the vcpkg documentation.
GPU painter that uses AccelGfx is slower and way less complete compared
to both default Gfx::Painter and Skia painter. It does not make much
sense to keep it, considering Skia painter already uses Metal backend on
macOS by default and there is an option to enable GPU-accelerated
backend on linux.
This call is used to inform the chrome that it should display a tooltip
now and avoid any hovering timers. This is used by <video> tags to
display the volume percentage when it is changed.
Now instead of sending the position in which the user entered the
tooltip area, send just the text, and let the chrome figure out how to
display it.
In the case of Qt, wait for 600 milliseconds of no mouse movement, then
display it under the mouse cursor.
This allows developers on macOS to open Ladybird.app in Instruments.
Add some documentation for how to use the command as well. It is enabled
automatically when CMAKE_BUILD_TYPE is not Release or RelWithDebInfo.
By Setting setBordered propperty on header buttons to `Yes` this
path makes the whole button clickable. Previously the only the
icon was clickable, now it's easy to click.
This large commit also refactors LibWebView's process handling to use
a top-level Application class that uses a new WebView::Process class to
encapsulate the IPC-centric nature of each helper process.
In the upcoming changes, Skia painter will be switched to Metal backend,
so we can no longer assume `pitch = width * 4` while reading Gfx::Bitmap
that wraps IOSurface populated by writing into MTLTexture that has
padded scanlines.
Previously, autocomplete was cleared before the results for the current
query were retrieved. The new results would then be added when the
network request completed. This resulted in a noticable flicker. The
results are now updated when the request for the current query is
completed.
There is a small behavior change in that the query itself is no longer
included in the autocomplete dropdown unless the list would otherwise
be empty.
This was only put in place because on SerenityOS, we wanted to use the
pid of the WebContent process, rather than the peer pid for the socket.
When launching with SystemServer, this got annoying. However, now that
we only have the case that the UI process spawns processes, we can
get rid of this hack. This restores the ability to see WebContent
processes' statistics in the task manager widget.
This is the same behavior as RequestServer, with the added benefit that
we know how to gracefully reconnect ImageDecoder to all WebContent
processes on restart.
The placeholder text shows "Search or enter web address" which doesn't
tell the user about *how* the search will be performed. Other popular
open source browsers show the search engine that will be used. For
example:
Chromium: "Search **engine** or type a URL"
Firefox: "Search with **engine** or enter address"
This change changes the placeholder text of the location bar to show
"Search with **engine** or enter web address".
Using mmap-allocated memory for backing stores does not allow us to
benefit from using GPU-accelerated painting, because all the performance
increase we get is mostly negated by reading the GPU-allocated texture
back into RAM, so it can be shared with the browser process.
With IOSurface, we get a framebuffer that is both shareable between
processes and can be used as underlying memory for an OpenGL/Metal
texture.
This change does not yet benefit from using IOSurface and merely wraps
them into Gfx::Bitmap to be used by the CPU painter.
Allows WebContentClient to get pid of WebContent process right after
creation, so there is no window between forking and
notify_process_information() IPC response, when client doesn't know the
pid.
In the upcoming changes, we are going to switch macOS to using an
IOSurface for the backing store. This change will simplify the process
of sharing an IOSurface between processes because we already have the
MachPortServer running in the browser, and WebContent knows how to
locate the corresponding server.
LibLocale was split off from LibUnicode a couple years ago to reduce the
number of applications on SerenityOS that depend on CLDR data. Now that
we use ICU, both LibUnicode and LibLocale are actually linking in this
data. And since vcpkg gives us static libraries, both libraries are over
30MB in size.
This patch reverts the separation and merges LibLocale into LibUnicode
again. We now have just one library that includes the ICU data.
Further, this will let LibUnicode share the locale cache that previously
would only exist in LibLocale.
This patch adds a simple in-memory HTTP cache to each WebContent
process.
It's currently off by default (turn it on with --enable-http-cache)
since the validation logic is lacking and incomplete.
The location bar URL is no longer hidden when creating a new tab or
opening a new window that has an associated URL. Conversely, the
location bar is now always focused and the URL hidden when creating a
window or tab without an associated URL.
The location bar is focused when:
* Opening the browser from the command line with no URL arguments
* Opening a new tab (Ctrl+T)
* Opening a new window (Ctrl+N)
The location bar is not focused when:
* Opening the browser from the command line with one or more URLs
* Opening hyperlinks in a new tab
* Clicking a hyperlink with `target="_blank"`
This matches the behavior of other major browsers.
This adds a motion preference to the browser UI similar to the existing
ones for color scheme and contrast.
Both AppKit UI and Qt UI has this new preference.
The auto value is currently the same as NoPreference, follow-ups can
address wiring that up to the actual preference for the OS.
The URL is now not shown when a new tab is initially activated until
the location bar loses focus. This allows the user to see the location
bar placeholder text when opening a new tab. It also makes it easier to
paste URLs into the location bar after opening a new tab.
This commit adds the standard shortcuts for the Find Next and Find
Previous buttons on the find in page panel. These shortcuts are usually
F3 and Shift+F3 respectively, although Qt standard shortcuts may vary
across platforms.
On my system `QKeySequence::StandardKey::ZoomIn` includes both `Ctrl++`
and `Ctrl+=`, so explicitly adding the secondary `Ctrl+=` shortcut
ourselves results in an ambiguous shortcut error message being shown.
According to the Qt documentation the key bindings returned by
`QKeySequence::StandardKey::*` are platform specific, so we may still
need to add the secondary shortcut on some systems. Therefore, we now
check whether our secondary shortcut is already in the shortcut list
before adding it.
The placeholder text is there to prompt the user as to what could be
added in the address bar. The current text tells the user that they can
"Search or enter web address" even when the search setting is disabled.
When attempting to "Search" the user is instead sent to page ":", with
an error in the console:
WebContent(575249): (FIXME) Don't know how to navigate to :
This patch fixes this by checking whether the search feature is enabled
and setting the placeholder appropriately. This provides a slightly
better user experience.
Closes#132
Before we had HTTP::HeaderMap (which preserves multiple headers with the
same name), we collected multiple "Set-Cookie" headers and bundled them
together as a JSON array.
This was a huge hack, and now we can stop doing that, since LibWeb gets
access to the full set of headers now.
Instead of using a HashMap<ByteString, ByteString, CaseInsensitive...>
everywhere, we now encapsulate this in a class.
Even better, the new class also allows keeping track of multiple headers
with the same name! This will make it possible for HTTP responses to
actually retain all their headers on the perilous journey from
RequestServer to LibWeb.
If we get a suggestion from fontconfig, we try those fonts first, before
falling back on the hard coded list of known suitable fonts for each
generic family.
This makes WebView::Database wrap around sqlite3 instead of LibSQL. The
effect on outside callers is pretty minimal. The main consequences are:
1. We must ensure the Cookie table exists before preparing any SQL
statements involving that table.
2. We can use an INSERT OR REPLACE statement instead of separate INSERT
and UPDATE statements.
The "Enable Scripting", "Block Pop-ups" and "Enable Same-Origin Policy"
options are now set for every tab when toggled. They are also applied
to new tabs when they are created.
For some reason, moving the UI-specific CMake to its own files prevents
resource files from being copied to the Resource directory in the macOS
application. I'm not sure what the difference here is, but doing this
step during create_ladybird_bundle() works.
The main intention of this change is to have a consistent look and
behavior across all scrollbars, including elements with
`overflow: scroll` and `overflow: auto`, iframes, and a page.
Before:
- Page's scrollbar is painted by Browser (Qt/AppKit) using the
corresponding UI framework style,
- Both WebContent and Browser know the scroll position offset.
- WebContent uses did_request_scroll_to() IPC call to send updates.
- Browser uses set_viewport_rect() to send updates.
After:
- Page's scrollbar is painted on WebContent side using the same style as
currently used for elements with `overflow: scroll` and
`overflow: auto`. A nice side effects: scrollbars are now painted for
iframes, and page's scrollbar respects scrollbar-width CSS property.
- Only WebContent knows scroll position offset.
- did_request_scroll_to() is no longer used.
- set_viewport_rect() is changed to set_viewport_size().
This was no longer doing anything. We'll eventually want a way to pass
system default fonts to each WebContent process, but we don't need to
squeeze everything through this API that was really meant for Serenity's
very idiosyncratic font system.
This closes the window at WebContent process startup where we were
relying on Gfx::FontDatabase having some resolvable value in its default
font query.
