According to the Intel manual: "After reset, all bits (except bit 0) in
XCR0 are cleared to zero; XCR0[0] is set to 1."
Sadly we can't trust this, for example VirtualBox starts with
bits 0-4 set, so let's do it ourselves.
Fixes#5653
We can't always rely on the initial URL's path(), so use either the
user-specified argument or the URL path for determining the realpath,
depending on whether we got a file:// URL argument.
Fixes#4950.
The expression
(u8*)params.m_stack_location + stack_size
… causes UBSan to spit out the warning
KUBSAN: addition of unsigned offset to 0x00000002 overflowed to 0xb0000003
… even though there is no actual overflow happening here.
This can be reproduced by running:
$ syscall create_thread 0 [ 0 0 0 0 0xb0000001 2 ]
Technically, this is a true-positive: The C++-reference is incredibly strict
about pointer-arithmetic:
> A pointer to non-array object is treated as a pointer to the first element
> of an array with size 1. […] [A]ttempts to generate a pointer that isn't
> pointing at an element of the same array or one past the end invoke
> undefined behavior.
https://en.cppreference.com/w/cpp/language/operator_arithmetic
Frankly, this feels silly. So let's just use FlatPtr instead.
Found by fuzz-syscalls. Undocumented bug.
Note that FlatPtr is an unsigned type, so
user_esp.value() - 4
is defined even if we end up with a user_esp of 0 (this can happen for example
when params.m_stack_size = 0 and params.m_stack_location = 0). The result would
be a Kernelspace-pointer, which would then be immediately flagged by
'MM.validate_user_stack' as invalid, as intended.
..instead of a tool window. Tool windows are meant as accessories to
an application's main/primary windows, not to be primary windows
themselves.
Fixes#5667.
Previously the page background was always draw relative to the viewport
instead of following with the content. This should eventually become
an opt-in mode (via CSS "background-attachment") but for now let's have
the default behavior be that backgrounds scroll with content.
Also take this opportunity to move the background painting code from
the two web views to a shared location in InitialContainingBlockBox.
This adds 2 more flags, that help with the "select on start" invocation.
-s - makes us open the parent directory of the entry, and select it.
-r - makes FileManager to skip real path resolution for cases when we
want to select the symlink in parent directory.
Also, if the file path is passed as argument, not it will open parent
with the file selected.
If we set selection before the IconView is laid out, it has no size.
So it can't correctly calculate where to scroll. Forcing scroll after
the first resize fixes that.
Now that we use fragment for specifying starting selection in
FileManager we would benefit from providing it as argument instead of
setting it each time separately.
This commit adds a few basic variable substitution operations:
- length
Find the length of a string or a list
- length_across
Find the lengths of things inside a list
- remove_{suffix,prefix}
Remove a suffix or a prefix from all the passed values
- regex_replace
Replace all matches of a given regex with a given template
- split
Split the given string with the given delimiter (or to its
code points if the delimiter is empty)
- concat_lists
concatenates any given lists into one
Closes#4316 (the ancient version of this same feature)
Non-interactive shells (i.e. when running scripts) do not need this
functionality, so they are a boatload of wasted time.
This significantly reduces the script startup and shutdown times when
there are lots of executables in PATH or lots of entries in the history.
It is possible to set the run.sh mode via the SERENITY_RUN environment
variable, but the SERENITY_DISK_IMAGE="grub_disk_image" override for
qgrub mode was only checking $1. This makes qgrub mode work via 'ninja
run' without explicitly setting SERENITY_DISK_IMAGE:
SERENITY_RUN=qgrub ninja run
Instead of declaring a reserved area from byte 0x160 to 0x400, we
change the declaration of TimerStructure array to be 32 units.
Also, a static_assert was added, to ensure that the calculation is
right.
This reverts commit af22204488.
According to the HPET specification, each theoretical comparator takes
32 bytes in the MMIO space.
Although I hardly believe that any system will implement all 32
comparators, in practice if a machine happens to have more than 3
comparators, we need to address the comparators correctly if we want to
use them.
This fixes some precision issues in Painter::draw_triangle()
that caused the Cube demo to leave pixels empty between triangles
under certain angles.
Also adds some extra comments for clarity and early returns when
parts of the triangle do not need to be drawn.
It was previously using draw_tiled_bitmap, which always aligns the
tiles with the global origin and does not respect the alpha of the
source. Switch to a new Painter::blit_tiled helper which uses
Painter::blit under the hood, which has more correct behavior.
