DirectoryIterator, and so `Desktop::AppFile::for_each()`, doesn't
guarantee anything about the order, so we manually sort afterwards.
Which means using a manual version of NonnullRefPtrVector, since that
doesn't support `quick_sort()`.
Having files in Base's user `.config` folder means that every time the
Serenity image is built, all user settings in that file are thrown away.
So, let's not do that! :^)
Modified the default value for the homepage url to match what was in
Browser.ini, so there is no visible change.
Browser has a bunch of settings, but most are non-trivial to add here.
So far, these are implemented:
- Homepage URL
- Whether to close download windows when they complete
The others will be added in subsequent commits.
LibDSP can greatly benefit from this nice FFT implementation, so let's
move it into the fitting library :^)
Note that this now requires linking SoundPlayer against LibDSP. That's
not an issue (LibDSP is rather small currently anyways), as we can
probably make great use of it in the future anyways.
This patch makes use of all the new fallible APIs in LibGUI together
with TRY() to catch and propagate errors. The main error getting caught
is allocation failures while trying to construct the Help UI.
It's quite interesting to see how the code changes as more and more
fallible calls get branded as such by wrapping them in TRY().
There's a lot of repetitive "TRY(try_foo())" going on right now. Once
this becomes the dominant programming pattern, we can drop the "try_"
prefix everywhere. :^)
With this change, System::foo() becomes Core::System::foo().
Since LibCore builds on other systems than SerenityOS, we now have to
make sure that wrappers work with just a standard C library underneath.
This fixes#10940.
Previously, all presses of the Delete key without a modifier in Browser
were uselessly consumed by the "Delete" action in the bookmark context
menu.
Almost all synthesizer code in Piano is removed in favor of the LibDSP
reimplementation.
This causes some issues that mainly have to do with the way Piano
currently handles talking to LibDSP. Additionally, the sampler is gone
for now and will be reintroduced with future work.
This makes the default image fit perfectly into the default viewport,
which means the first fit_image_to_view call will end up with a scale of
exactly 1. This scale level has no visual artifacts, which is the more
intuitive 'default' behavior.
Fixes#10975.
This simplifies a bunch of error handling and makes the main function
quite a bit shorter.
It will become shorter yet, as we get better at propagating errors. :^)
Previously, when the --unlink flag was passed to CrashReporter, it
unlinked the coredump file immediately after reading it.
This change makes it so the coredump file is deleted when CrashReporter
exits.
Right now, this is a bit of a hack. We can't set a keymap to only apply
to the test area, so we set the system keymap instead, while also
keeping track of the "real" current keymap. Whenever the settings are
applied, we update what that "real" keymap is, and when we exit, we
revert to that keymap.
Basically, it behaves as you would expect, apart from it also affecting
other applications you are typing in while the KeyboardSettings window
is open with a different keymap selected.
This is functionally the same as before, as selecting a tab fires Show
and HideEvents. But this way, we don't need to directly access the
TabWidget, which will make using SettingsWindow simpler.
For an empty glyph (all pixels initially clear), this makes no
difference. However, if the glyph for the selected code point already
contains some set pixels, pasting used to "add" the set pixels, instead
of overwriting.
Also add slightly richer parse errors now that we can include a string
literal with returned errors.
This will allow us to use TRY() when working with JSON data.
The processor parameter values are displayed with two decimal places by
default. However, when these values become very large and exceed about 7
text symbols, the text is too long to fit the label and it'll simply not
show up. This commit fixes that by disabling the decimal place for such
large values, which allows us to show values up to 9,999,999, be it
only at integer precision.
This fixes all current code smells, bugs and issues reported by
SonarCloud static analysis. Other issues are almost exclusively false
positives. This makes much code clearer, and some minor benefits in
performance or bug evasion may be gained.
Having the delete key handling be done via an action limits our ability
to support key modifiers (e.g. ctrl+delete deleting the word in front of
the cursor).
The fact that it was an action _did_ allow us to have a delete button in
the TextEditor UI. However, this is an odd choice in the first place
that isn't common in other text editors, so I just removed it.
This change calculates the face-normal of the triangle by adding
the three vertex-normals and then normalizing. This results in an
average of the three vertex-normals.
Same as Vector, ByteBuffer now also signals allocation failure by
returning an ENOMEM Error instead of a bool, allowing us to use the
TRY() and MUST() patterns.
The path returned by GUI:FilePicker is stored on the stack when the
callback is executed. The player only stored a StringView to the path
however it should take ownership of the path instead since the path is
accessed even after the file menu open action has returned.
"Frame" is an MPEG term, which is not only unintuitive but also
overloaded with different meaning by other codecs (e.g. FLAC).
Therefore, use the standard term Sample for the central audio structure.
The class is also extracted to its own file, because it's becoming quite
large. Bundling these two changes means not distributing similar
modifications (changing names and paths) across commits.
Co-authored-by: kleines Filmröllchen <malu.bertsch@gmail.com>
Previously, the active window being centered, and then the inactive one
being higher, meant having an awkward gap at the bottom. This is a
simple fix, and not perfect, but it helps. :^)
The GUI for this is a bit odd, especially since we only have one flag,
but otherwise adding new flags would require modifying ThemeEditor. At
least it is consistent with the other theme properties.
This fix allows us to move the knob wherever we click inside the slider.
The 'jump_to_cursor()' mechanism wasn't working properly because the
player was overwriting the value we had just clicked.
Even though they are called content_margins,
they are actually only ever used to determine where
a Widget is supposed to be grabbable.
So all methods and members are renamed accordingly.
Derivatives of Core::Object should be constructed through
ClassName::construct(), to avoid handling ref-counted objects with
refcount zero. Fixing the visibility means that misuses like this are
more difficult.
This commit is separate from the other Applications/Libraries changes
because it required additional adaption of the code. Note that the old
code did precisely what these changes try to prevent: Create and handle
a ref-counted object with a refcount of zero.
Derivatives of Core::Object should be constructed through
ClassName::construct(), to avoid handling ref-counted objects with
refcount zero. Fixing the visibility means that misuses like this are
more difficult.
For better visibility of wether the editing focus is on the hex or the
ascii view, render a blinking caret instead of a solid cell background.
For that to work, it's also necessary to change the way selection works.
The selection shouldn't extend to the current position but up to the
byte before it.
The old behavior of restarting the timer after every second click could
result in double-click-chains (or triple+ clicks), which does not feel
like the right behavior.
By resetting the double-clicking timer, you need to perform a new full
double-click to make the arrows change color again.
The shuffling algorithm uses a naïve bloom filter to provide random
uniformity, avoiding items that were recently played. With 32 bits,
double hashing, and an error rate of ~10%, this bloom filter should
be able to hold around ~16 keys, which should be sufficient to give the
illusion of fairness to the shuffling algorithm.
This avoids having to shuffle the playlist itself (user might have
spent quite a bit of time to sort them, so it's not a good idea to mess
with it), or having to create a proxy model that shuffles (that could
potentially use quite a bit of memory).
This is a first pass at refactoring SoundPlayer so that the View widget
is decoupled from the player itself.
In doing so, this fixed a couple of issues, including possibly
inconsistent states (e.g. player could be paused and stopped at the
same time).
With the change, Player actually controls the show, and calls methods
overriden by its subclasses to perform actions, such as update the Seek
bar; the hard work of massaging the raw data is done by the Player
class, so subclasses don't need to reimplement any of these things.
This also removes some copies of playlist management code that happened
to be copied+pasted inside callbacks of buttons -- it now lives inside
a neatly packaged Playlist class, and the Player only asks for the next
song to play.
In addition, the menu bar has been slightly rearranged.
Add option to reverse primary and secondary buttons in Mouse Settings.
- WindowServer.ini: add default entry
- switch-mouse-buttons.png: new icon for settings entry
- Mouse.gml/MouseWidget.*: new settings dialog
- ClientConnection/WindowManager/Server: window message for settings
- EventLoop.cpp: swap buttons 1 and 2 if settings are on
From what I think, the array should consist of point indexes that have
been matched instead of just the last one.
For example, these are the array contents after searching 'file' for
'File Manager':
- Before: [ 3 ]
- Now: [ 0, 1, 2, 3 ]
Besides that, this greatly improves the scoring logic, as it can now
calculate bonuses.
Closes: #8310
They were previously taking up 9% of samples in a profile of PixelPaint
while selecting a mask, and as a result of moving them to the header
they were inlined, which effectively eliminated them from the profile.
The old versions were renamed to JS_DECLARE_OLD_NATIVE_FUNCTION and
JS_DEFINE_OLD_NATIVE_FUNCTION, and will be eventually removed once all
native functions were converted to the new format.
When the image is flipped or rotated, the window is resized to ensure
that the image still fits in the frame. However, currently the original
bitmap rect is used, which doesn't take into account the scaling
factor. Fix this by using the scaled rect instead.
Previously, when the last layer got deleted, the active layer was
set to nullptr, causing a crash.
Now, we create a new transparent background layer with the image
dimensions instead.
Since there's no global API for being able to just assign a callback
function to config changes, I've made an inline struct in desktop
mode with the sole purpose of checking to see if the Wallpaper
entry has changed, and then updates GUI::Desktop.
It's pretty neat seeing the wallpaper change as soon as you edit the
config file :^)
Prior this patch, you couldn't remove any files from the context menu
if you didn't have write access to them.
It was incorrect, as the write permission for files means that you can
modify the contents of the file, where for directories it means that
you can create, rename, and remove the files there.
The memory and CPU graphs fail to display anything when the memory size
is larger than 2**31 bytes, because of the small range of int. This
commit makes replaces the type with size_t. Hopefully nobody will have
18 quintillion bytes of memory before this gets replaced. :^)
There's a subtle difference here. A "block box" in the spec is a
block-level box, while a "block container" is a box whose children are
either all inline-level boxes in an IFC, or all block-level boxes
participating in a BFC.
Notably, an "inline-block" box is a "block container" but not a "block
box" since it is itself inline-level.
I used "git grep -FIn http://" to find all occurrences, and looked at
each one. If an occurrence was really just a link, and if a https
version exists, and if our Browser can access it at least as well as the
http version, then I changed the occurrence to https.
I'm happy to report that I didn't run into a single site where Browser
can't deal with the https version.
