And let the old shadow_root(), which was only supposed to be used by
bindings, be called shadow_root_for_bindings() instead.
This makes it much easier to read DOM code, and we don't have to worry
about when to use shadow_root_internal() or why.
(cherry picked from commit f4bdf562127c12a7af18029777a88d4260af82d3)
We now honor the shadowrootmode attribute on template elements while
parsing, and instantiate a shadow tree as required by the spec.
(cherry picked from commit 9eb4b91168145def5677cd41057d34ecf369372b)
The bulk of this function is moved to a new "attach a shadow root"
helper, designed to be used both from here and the HTML parser.
(cherry picked from commit 043ad0eb7644a529c1ffbd4de9a4b89771808c84,
and amended to make `is_valid_shadow_host_name()` static.
The expectation with these messages is that no other code is allowed to
run while we're waiting for the response to arrive.
Not meeting this requirement did not degrade any (observable)
functionality, however.
These were made invalid with P1787, and Clang (19) trunk started warning
on them with https://github.com/llvm/llvm-project/pull/80801.
(cherry picked from commit 397774d42272fff8dbc6d8d53616d79667d6608a)
The default implementation of die() causes the client process to simply
exit cleanly. This prevents any tests from recognizing that something
went wrong, as the process exits with a code of 0. With this patch, we
still just exit when the connection dies, but with a fatal signal. In
the future, we will want to launch a new RequestServer process and
re-establish client connections.
(cherry picked from commit d1ec32e28feee10c2c36f4cd0d496b8a95d375f8)
8263e0a619 added `-Wno-coroutine-missing-unhandled-exception` to the
cmake build, and the coroutine code does not build without it.
Unbreaks building e.g. `js` with GN.
These can use an EnvironmentCoordinate for caching, just like normal
binding lookups. Saves a bunch of time for repeated typeof checks.
(cherry picked from commit 60a05ef414f804fe0c44d20fe13c795800826368)
This is actually safe everywhere but in the topmost program scope.
For web compatibility reasons, we have to flush all top-level bindings
to the environment, in case a subsequent separate <script> program
comes looking for them.
(cherry picked from commit 9d21d88374cf440ddf381e02a597a59ad329ed62)
Instead of displaying locals as "locN", we now show them as "name~N".
This makes it a lot easier to follow bytecode dumps, especially in
longer functions.
Note that we keep displaying the local index, to avoid confusion in case
there are multiple separate locals with the same name in one executable.
(cherry picked from commit 0aa8cb7dac60c88eac3bb7674e3fe575cf1da60b)
When resuming execution of a suspended function, we must not overwrite
any cached `this` value with something from the ExecutionContext.
This was causing an empty JS::Value to leak into the VM when resuming
an async arrow function, since the "this mode" for such functions is
lexical and thus ExecutionContext will have an empty `this`.
It became a problem due to the bytecode optimization where we allow
ourselves to assume that `this` remains cached after we've executed a
ResolveThisBinding in this (or the first) basic block of the executable.
Fixes https://github.com/LadybirdBrowser/ladybird/issues/138
(cherry picked from commit a91bb72dabe54849657720e36e19a3d3b737055f)
Instead, let's create a new class called AvailablePortDatabase that will
handle updating (from the Internet, by using our repository) of the list
and instantiating an object that could be used for querying.
By unsetting `CXXFLAGS` in order to build Lagom code generators without
instrumentation, we were also removing `-stdlib=libc++`, so Clang ended
up using the runner image's libstdc++ 9 headers. These are too old, and
don't contain `<coroutine>`, leading to a build failure.
Fixes#24578
This allows RS to start connections in parallel without actively waiting
for any possible handshakes.
Doing so gives us a nearly-3x speedup on the average connection latency.
Previously this just yielded the time slice, which blocked the event
loop until the other side read some data from the socket.
This "fixed" impl is still equally bad, but at least it doesn't block
the event loop.
These were showing up on profiles as quite hot (>5%), we can avoid all
the unnecessary assertions by doing them once in advance and using
pointers for the rest of the function.
This makes AES-GCM about 6% faster.
These options supplements the original threshold option, because it's
now possible to specify a threshold and add either minimum/maximum size
for additional filtering.
It's also possible to not use the old threshold option (that Tim
Schumacher told me it was inherited from coreutils) and only specify the
2 new options to create a filtering.
This reverts commit 59cb7994c6.
This change caused a bug where getImageData returned the image in
BGRA8888 format instead of RGBA8888.
(cherry picked from commit d2f9ba7db1d8dd1dd611e6a3f40b9125b5aa054d)
