| Age | Commit message (Collapse) | Author |
|---|
|
This hack allows for Boxes that have a background to be painted and a
border to accurately paint their border-radii if needed.
For that the box in with the background is drawn is extended to the
bordered_rect. The border is later drawn over this regardless.
Previously when drawing a Box that had all three, background, border
and a border-radius, there could be some white between the filling and
the border.
|
|
This applies to thicker borders. When drawing them we now don't
chamfer the corner if there is a border-radius present.
|
|
The struct BorderRadiusData contains the four radii of the box.
In case the specified borders are too large for the dimensions of the
box, they get scaled down.
|
|
Even if the Box doesn't have a border-radius, we'll pass the rendering
of the Box and its corners to the Painter.
|
|
It supports all four corners.
|
|
This adds a function to draw a circle specified by a center point (
relative to the given Rect) and a radius. The circle arc is only drawn
inside the specified Rect as to allow for circle arc segments.
Technically this was already possible using draw_elliptical_arc(), but
the algorithm is quite involved and lead to wonky arcs when trying to
draw circle arc segments.
|
|
This paints a rectangle with rounded corners each specified by a
radius.
|
|
This takes care of the 1, 2, 3 and 4 parameter shorthand of the border-
radius identifier.
There are more as well as the ominous '/' character but that is for
another time. The 2 and 3 parameter versions are weird enough already.
I don't think anybody uses anything other than the 1 or 4 parameter
version or even the elliptical stuff.
|
|
|
|
And resolve the shorthands.
|
|
|
|
So other Boxes can override this function.
|
|
|
|
|
|
This can currently highlight tag names and attribute names/values.
|
|
|
|
|
|
This was breaking ports linting, which runs the script with the
'showproperty' option. This check is not needed for some other options
as well, so let's do it conditionally.
|
|
If you type in a filename that doesn't exist, show an error message
instead of closing the FilePicker "successfully."
|
|
Some people apparently like to type in full absolute paths into the
filename box of GUI::FilePicker. So let's handle that as you'd expect
by using the full path as the selected path.
|
|
Return a String instead of a LexicalPath. Also call it a path instead
of a file since that's what we're really returning.
|
|
Ask for a bold_variant() of the default_fixed_width_font() instead.
|
|
Instead use default_font().bold_variant() in cases where we want a bold
variant of the default font. :^)
|
|
Simply by checking whether a built libc.so exists, we should be able to
avoid strange build errors where that's not the case and just tell the
user upfront.
Fixes #7309.
|
|
Problem:
- `constexpr inline` is redundant because `constexpr` implies `inline`.
Solution:
- Remove redundancy.
|
|
|
|
The PS2KeyboardDevice can be free'd in try_to_initialize if the
initialization failed, resulting in an assertion.
|
|
Don't pollute the source directory with generated code.
|
|
Painting the wallpaper at 100% scale in the top left corner is a weird
mode and I don't think we need it. :^)
|
|
Previously, this code would create a bitmap with the size of the screen
resolution, draw the new wallpaper into it, and then scale it down for
display inside the MonitorWidget.
This was done on every paint event, which made the code quite slow and
allocation-happy.
Instead of this, we now scale down the new wallpaper to a miniature
of the same scale as the little MonitorWidget screen. The miniature is
then used for tiling, etc. The miniature is cached and reused across
paint events if nothing else changes.
|
|
An application that allows opening arbitrary files from the filesystem
needs to allow itself to access the filesystem, otherwise there's no
point in supporting the feature. :^)
And the "thread" pledge is needed for background thumbnail generation.
|
|
|
|
And use GENERATED_SOURCES (or add_dependencies) to make LibVT depend on
that target.
Fixes a FIXME.
|
|
Also map the author name I used before to my current proper name.
|
|
|
|
I introduced this bug in e95eb7a51, where it's possible that the
ProcessGroup is created, but we never add it to the list. Make sure we
check that we are in the list before removal. This only broke booting in
self-test mode oddly enough.
Reported-By: Andrew Kaster <andrewdkaster@gmail.com>
|
|
.mailmap files are documented here: https://git-scm.com/docs/gitmailmap
|
|
This avoids allocations for smp_broadcast() and smp_unicast() by
using the Function class.
|
|
This avoids allocations for deferred_call_queue().
|
|
When blocking a thread with a timeout we would previously allocate
a Timer object. This removes the allocation for that Timer object.
|
|
Removing the element from the intrusive linked list might not be safe
if doing so requires a lock. Instead this is something the caller
should have done so let's verify instead that we're not on any lists.
|
|
In cases with ambiguous captures involving pawns (where multiple pieces
could have made the capture), we were exporting invalid syntax for
the move:
`1. e4 e5 2. Bb5 c6 3. Bxc6 ddxc6`
Move 3 should be `Bxc6 dxc6`, but we were duplicating the d on the pawn
move.
|
|
|
|
Avoid allocating while holding the g_process_groups_lock spinlock, it's
a pattern that has a negative effect on performance and scalability,
especially given that it is a global lock, reachable by all processes.
|
|
Make ProcessGroup::find_or_create & ProcessGroup::create OOM safe, by
moving to adopt_ref_if_nonnull.
|
|
Currently in SMP mode we hard code support for up to only 8 processors.
There is no reason for this to be a dynamic allocation that needs to be
guarded by a spinlock. Instead use a Array<T* with inline storage of 8,
allowing each processor to initialize it self in place, avoiding all
the need for locks.
|
|
Spinlocks guard short regions, with hopefully no other locks being taken
in the process. Violating constraints usually had detrimental effects on
platform stability as well as performance and scalability. Allocating
memory takes it own locks, and can in some cases even allocate new
regions, and thus violates these tenants.
Move the AnonymousVMObject creation outside of the spinlock as
creation does not modify any shared state.
|
|
We had rules for .png and .jpg files, but we have not maintained the
list as support for other file formats has been added. To test these
changes test files have been committed for each of these formats.
This change updates the list with all of the binary image file types I
was able to find in the tree at the time of writing.
|
|
The wild card rules at the top of the .gitignore came from a time when
the build wrote back to the git repository and placed files right next
to the source. (Original commit that introduced them 37c27e2e, they were
later consolidated into the root .gitignore in 802d4dc) We have since
moved to cmake, and these rules have become obsolete, and they just
cause issues where we need to go and add negations for these rules in
order for things to work.
A previous change attempted to remove the top wild card rules (PR #4565)
but it was later reverted, as they forgot to remove the top ignore
everything rule '*', so all files were ignored. This change just removes
all of these rules that no longer make sense, restoring a bit of sanity.
*.o,*.d,*.a rules were also from when the build wrote to the repository,
they are now defunct. The same goes for the *Endpoint.h and CMakeFiles
rules.
The lowercase build directory can be removed as we've standardized on
the uppercase 'Build' directory as the root of the build output dir.
|
|
Problem:
- Bitmasks are duplicated.
- Bitmasks are C-style arrays.
Solution:
- Move bitmasks to BitmapView.h.
- Change C-style arrays to be AK::Array for added safety.
|
