This confirmed works on Xcode 16, and Xcode 16.1 Beta 2, with CMake 3.28
or higher.
On linux, the 6.0.0 release from swiftly is still missing my libstdc++
workaround, so it needs a snapshot to work.
When the detected SDK for CMAKE_OSX_SYSROOT and friends has the same
version as your current macOS system version, CMake helpfully doesn't
set CMAKE_OSX_DEPLOYMENT_TARGET. Unfortunately, in this case, swiftc
will default to macOS 10.4, which is absolutely ancient. Grab the target
triple from the -print-target-info JSON when CMAKE_OSX_DEPLOYMENT_TARGET
is not provided at configure time.
The IPCs to request a page's text, layout tree, etc. are currently all
synchronous. This can result in a deadlock when WebContent also makes
a synchronous IPC call, as both ends will be waiting on each other.
This replaces the page info IPCs with a single, asynchronous IPC. This
new IPC is promise-based, much like our screenshot IPC.
This change allows the user to specify the format of the log file to be
generated by the `WPT.sh` script. Multiple logging arguments may now be
specified.
The supported logging arguments are: `--log-raw`, `--log-unittest`,
`--log-xunit`, `--log-html`, `--log-mach`, `--log-tbpl`,
`--log-grouped`, `--log-chromium`, `--log-wptreport` and
`--log-wptscreenshot`. These arguments act the same as the equivalent
arguments supported by `wpt run`.
The short `--log` argument may also be used as an alias for `--log-raw`.
UI event handlers currently return a boolean where false means the event
was cancelled by a script on the page, or otherwise dropped. It has been
a point of confusion for some time now, as it's not particularly clear
what should be returned in some special cases, or how the UI process
should handle the response.
This adds an enumeration with a few states that indicate exactly how the
WebContent process handled the event. This should remove all ambiguity,
and let us properly handle these states going forward.
There should be no behavior change with this patch. It's meant to only
introduce the enum, not change any of our decisions based on the result.
This adds a new script for linting WebIDL files, and adds it to the set
of scripts Meta/lint-ci.sh runs. Initially, this script does just one
thing: normalizes IDL definition lines so they start with four spaces.
This script only checks Tests/AK, and verifies that all source files
that match Tests/AK/*.cpp are listed in the CMakeLists.txt.
This is a bit excessive. We don't have this check for any other test
files. This sort of error will definitely ™️ be caught in review.
CMake reads CMakePresets.json, which is before it reads CMakeLists.txt.
This causes CMake Error: Unrecognized "version" field if the version of
CMake is older than support for presets, or the version field of
presets.
The fix is to check CMake version in ladybird.sh before trying to create
the build directory.
Co-Authored-By: Andrew Kaster <andrew@ladybird.org>
You can now build with STYLE_INVALIDATION_DEBUG and get a debug stream
of reasons why style invalidations are happening and where.
I've rewritten this code many times, so instead of throwing it away once
again, I figured we should at least have it behind a flag.
Instead of switching on the PropertyID and doing a boatload of
comparisons, we reorder the PropertyID enum so that all inherited
properties are in two contiguous ranges (one for shorthands,
one for longhands).
This replaces the switch statement with two simple range checks.
Note that the property order change is observable via
window.getComputedStyle(), but the order of those properties is
implementation defined anyway.
Removes a 1.5% item from the profile when loading https://hemnet.se/
This header held a bunch of utility functions shared across several code
generators. The only user of any of these utilities now is the public
suffix generator. Move the one used function to that generator, and
remove the common header.
Depending on usage, `@layer` has two forms, with two different CSSOM
types. One simply lists layer names and the other defines a layer with
its contained rules.
This change should move us forward toward emoji support, as we are no
longer limited by our own OpenType implementation, which was failing
to parse the TrueType Collection format used to store emoji fonts
(at least on macOS).
This is only used for CSS style sheets. One case wants it as a String,
and the others don't care, but will in future also want to have the
source as a String.
When trying to use pkgconfig for finding libjxl, the build fails
trying to link the cross-compiler's libc++.
Using this way libjxl also requires hwy library.
Findlibjxl.cmake was taken from SDL_image and altered to include its license.
This project is a part of the Xcode-shipped toolchain on macOS, but
needs built from source on other platforms. However, using the Xcode
version of the framework leads to a bunch of rpath confusion when
trying to link it the expected way. I suspect that there will be a
more intuitive way to link this library from the toolchain when it
stabilizes. So we'll build it everywhere :)
Instead of using a global setting, let's set this per-target. This
prevents conflicts when importing third-party dependencies that do
not tolerate the mode being "default".
At the same time, simplify CMakeLists magic for libraries that want to
include Swift code in the library. The Lib-less name of the library is
now always the module name for the library with any Swift additions,
extensions, etc. All vfs overlays now live in a common location to make
finding them easier from CMake functions. A new pattern is needed for
the Lib-less modules to re-export their Cxx counterparts.
We don't actually generate any such events ourselves. But Google Lens
will create one with the DataTransfer attribute set to that of any drop
event we send it.
Instead of CSSColorValue holding a Gfx::Color, make it an abstract class
with subclasses for each different color function, to match the Typed-OM
spec. This means moving the color calculations from the parsing code to
the `to_color()` method on the style value.
This lets us have calc() inside a color function, instead of having to
fully resolve the color at parse time. The canvas fillStyle tests have
been updated to reflect this.
The other test change is Screenshot/css-color-functions.html: previously
we produced slightly different colors for an alpha of 0.5 and one of
50%, and this incorrect behavior was baked into the test. So now it's
more correct. :^)
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.
USVString attributes Now replace any surrogates with the replacement
character U+FFFD and resolve any relative URLs to an absolute URL. This
brings our implementation in line with the specification.
The Web Platform Tests runner requires that some hostnames point to
localhost when running the tests locally. We now append these hostnames
to `/etc/hosts` if they aren't already present.
For a long time, we've used two terms, inconsistently:
- "Identifier" is a spec term, but refers to a sequence of alphanumeric
characters, which may or may not be a keyword. (Keywords are a
subset of all identifiers.)
- "ValueID" is entirely non-spec, and is directly called a "keyword" in
the CSS specs.
So to avoid confusion as much as possible, let's align with the spec
terminology. I've attempted to change variable names as well, but
obviously we use Keywords in a lot of places in LibWeb and so I may
have missed some.
One exception is that I've not renamed "valid-identifiers" in
Properties.json... I'd like to combine that and the "valid-types" array
together eventually, so there's no benefit to doing an extra rename
now.
Commit 35392d4d28 moved the
target_*_directories() calls (or rather their include()) before the
target declaration, so they fail because of the undefined target.
We can fix the problem by using global *_directories() instead.
Implements the corresponding encoders, selects the appropriate one when
encoding URL search params. If an encoder for the given encoding could
not be found, fallback to utf-8.
USVString is defined in the IDL spec as:
> The USVString type corresponds to scalar value strings. Depending on
> the context, these can be treated as sequences of either 16-bit
> unsigned integer code units or scalar values.
This means we need to account for surrogate code points by using the
replacement character.
This fixes the last test in https://wpt.live/url/url-constructor.any.html
In theory the clang module map should not have absolute paths for the
headers. Other Swift projects seem to use the -ivfsoverlay feature of
clang to work around this, but it seems difficult to get to work.
And modernize the cmake_parse_arguments() call at the top.
Ideally, we would pull these flags from the target we're generating
for, but the current CMake setup makes that prohibitively infeasible.
This ensures that we can get all the proper warnings on/off to get the
same diagnostics and other options when loading C++ headers into the
Swift frontend.
Add a simple shell script to update the local clangd configuration
according to the type of build selected by the user. Include
documentation on where the script might be useful when building
under different configurations.
For the SVG <use> element, we want to support loading HTML documents
that have a SVG element inside of it pointed to by the URL fragment.
In this situation we would need to fetch and parse the entire document
in SharedImageRequest (so that we can still cache the SVGs). Rename
SharedImageRequest to SharedResourceRequest to make the class a little
more generic for future usecases.
Previously we were assuming that the attribute return value was never
nullable and going to be returned in an Optional<IntegralType>, causing
complile errors for something such as: `attribute unsigned long?`.
Mirroring the pre-existing `generate_from_integral` function. This will
allow us to fix a bug that all of these if statements have in common -
no handling of nullable types.
This also adjusts the type casted for each integral to fully match that
stated by the spec.
The main incentive is much better performance. We could have gone a bit
further in optimizing the Skia painter to blit glyphs produced by LibGfx
more efficiently from the glyph atlas, but eventually, we also want Skia
to improve correctness.
This change does not completely replace LibGfx in text handling. It's
still used at all stages, including layout, up until display list
replaying.
It turns out we were already generating all the necessary include
statements, and we can simply remove all this goofy code soup that
uses the C preprocessor to speculatively look for include files.
This is `counter(name, style?)` or `counters(name, link, style?)`. The
difference being, `counter()` matches only the nearest level (eg, "1"),
and `counters()` combines all the levels in the tree (eg, "3.4.1").
This change removes wrappers inherited from Gfx::Typeface for WOFF and
WOFF2 fonts. The only purpose they served is owning of ttf ByteBuffer
produced by decoding a WOFF/WOFF2 font. Now new FontData class is
responsible for holding ByteBuffer when a font is constructed from
non-externally owned memory.
Override the vcpkg/scripts/detect_compiler behavior of always pulling
$CC and $CXX at the time that vcpkg install is determined to need called
by forcing $ENV{CXX} and $ENV{CC} to our CMake-determined compiler.
This prevents strange behavior such as running the following:
./Meta/ladybird.sh run
make changes...
ninja -C Build/ladybird
Where the second build step would be run without CC or CXX set in the
environment, causing a total cache miss from vcpkg and a full rebuild.
It also helps prevent full rebuilds when an IDE passes a slightly
different compiler to the build step than ladybird.sh.
AK will depend on some vcpkg dependencies, so the Lagom tools build will
need to know how to use vcpkg. We can do this by sym-linking vcpkg.json
to Meta/Lagom (as vcpkg.json has to be in the CMake source directory).
We also need a CMakePresets.json in the source directory, which can just
include the root file. The root CMakePresets then needs to define paths
relative to ${fileDir} rather than ${sourceDir}.
- `-fstack-protection-strong` enables stack canaries for functions where
addresses of local variables are taken or arrays/structures
containing arrays are allocated on the stack.
- `-fstrict-flex-arrays=2` causes the compiler to only treat arrays with
unknown bounds (`[]`) or zero-length-arrays (`[0]`) as *flexible array
members*, allowing the sanitizers to emit bounds checks for structs
with proper arrays as their last member.
More rigorous options (such as AArch64 pointer authentication, Control
Flow Integrity, _FORTIFY_SOURCE) should be investigated in the future,
however this is a good baseline.
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.
These used to be enabled in `serenity_compile_options.cmake` for
Serenity builds and were removed in 9b05fb98. This is a slightly more
conservative subset of those, with ones that are enabled by default
omitted.
This should prevent our code quality regressing in the long run.
This change updates the Meta/check-debug-flags.sh script to avoid an
apparent Bach 3.2 parser bug. Specifically, it takes a comment and some
code of a process substitution and moves it into a separate function.
Otherwise, without this change, trying to run the check-debug-flags.sh
script with Bash 3.2 fails with the following error:
line 39: bad substitution: no closing `)' in <(
...apparently because Bash 3.2 chokes on the comment (and doesn’t choke
if the comment is completely removed).
Relates to https://github.com/LadybirdBrowser/ladybird/issues/283
This change makes all the pre-commit CI scripts runnable under Bash 3.2,
by replacing “mapfile” invocations in them code that first explicitly
creates an array, and then uses a while loop to populate the array.
Otherwise, without this change, the scripts all fail to run under Bash
3.2 — due to lack of support for “mapfile”.
Fixes https://github.com/LadybirdBrowser/ladybird/issues/283
This also drops bash from the list of homebrew dependencies in the build
instructions — because with this change, homebrew bash (v4) is no longer
needed; things will now work with the Apple-provided bash (v3.2)
We no longer have multiple locations including AK (e.g. LibC). So let's
avoid awkwardly defining the AK library across multiple CMake files.
This is to allow more easily adding third-party dependencies to AK in
the future.
Since we support the multi-memory proposal, we should skip tests that
validate that we have only one memory. Once multi-memory gets included
in the main WebAssembly specification (and the testsuite is updated), we
can revert this commit.
Because `nan:arithmetic` and `nan:canonical` aren't bound to a single
bit pattern, we cannot check against a float-containing SIMD vector
against a single value in the tests. Now, we represent `v128`s as
`TypedArray`s in `testjs` (as opposed to using `BigInt`s), allowing us
to properly check `NaN` bit patterns.
Some spec-tests check the bit pattern of a returned `NaN` (i.e.
`nan:canonical`, `nan:arithmetic`, or something like `nan:0x200000`).
Previously, we just accepted any `NaN`.
`linking.wast` has an unusual pattern for invoke commands, which is now
accounted-for. Also, special unicode characters are now properly
serialized in JavaScript as string literals (this is only relevant for
`names.wast`).
Trying to build VulkanLoader from source is a giant headache of
unnecessary packages. Every modern distro has vulkan packages, let's
depend on those instead of trying to build something for both wayland
and X11.
We need to avoid using vcpkg's pkg-config on non-x86_64 platforms,
because they do very strange things to the default paths.
On Asahi Linux and other 16 KiB page distros, the user must also provide
a properly compiled version of GN.
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.
