There are a lot of native C++ functions that will be used by both the
bytecode interpreter and jitted code. Let's put them in their own file
instead of having them in Interpreter.cpp.
This necessitated making the JIT::Compiler aware of the current
Bytecode::Executable, since that's where all the string literals are
held, but that seems like a good thing.
These push a "valid" unwind context on the stack and check_exception()
now knows how to jump to the (catch) handler if present.
(finally) finalizers will require some more work, but with this change,
we now have basic support for try...catch. :^)
We now establish a stack of "unwind contexts" similar to what the
bytecode interpreter does, but here, it's a stack of structs with
addresses to the catch and finally blocks.
Unwind contexts also have a "valid" flag, and the root unwind context
(always present, pushed on JIT code entry) has valid=false, which we
interpret in check_exception() as "return and let our caller deal with
the exception".
Anything in Compiler that may generate an exception should now also
call check_exception() ASAP to emit the code for handling this.
Nothing in PDF 1.7 spec 8.2.1 Destinations mentions the page being
`null`, but it happens in 0000372.pdf (for the root outline element)
and in 0000776.pdf (for every outline element, which looks like a
bug in the generator maybe) of 0000.zip from the pdfa dataset.
This patch brings a service to handle image decompression. With it comes
security enhancement due to the process boundary. Indeed, consequences
of a potential attack is reduced as only the decoder will crash without
perturbing the WebContent process.
It also allows us to display pages containing images that we claim to
support but still make us crash, like for not-finished-yet decoders.
As an example, we can now load https://jpegxl.info/jxl-art.html without
crashing the WebContent process.
This library offers tools to communicate with an ImageDecoder server
through IPC. There is currently no such executable for Lagom but that
shouldn't take long :^)
Changing `try_compute_width()` to return width and margins, instead of
mutating them in the box's state makes it works in cases when box has
min-width or max-width and this function needs to be called multiple
times.
Fixes https://github.com/SerenityOS/serenity/issues/21598
