This gets rid of the last use of the offset_margin_width() magic number.
I initially tried using `character_width() * 10` and found it was not
accurate enough with between-character spacing, so instead this measures
a specific string. All offset strings should be the same width in a
fixed-width font.
Compare the x position with the start of the hex characters, which is
m_padding pixels to the right of hex_start_x. (And the same for the
text characters.) If the position is within the padding area, clamp it
so it's on the nearest character.
This was previously covered-up somewhat by using the buggy
m_address_bar_width value to calculate the start and end x positions of
these areas. Fixing that in the previous commit made this more obvious.
We repeat the same calculations a lot, and it's easy to make mistakes.
m_address_bar_width and offset_margin_width() have basically the same
purpose, but both are a little wrong. This removes the former, but
we'll get to the latter soon.
Subtracting 1 from both axes causes a kink in the line, because the
start point isn't also translated. The clip rect prevents us from
painting too far down, so we can just remove the translated() call.
Ensure that the Utf-8 encoded "mapping character" in String (from
InputBox) gets converted to Utf-32 code point. Before, only the first
byte of Utf-8 char sequence was used as a Utf-32 code point.
That used to result in incorrect characters getting written in the
keymap file for all the "non-ascii" characters used as the mapping
character.
The change ensures that functions that parse string values for terminal
settings fall back to a default value, if there is no value present or
the value is invalid.
Automarks are similar to bookmarks placed by the terminal, allowing the
user to selectively remove a single command and its output from the
terminal scrollback.
This commit implements a single way to add marks: automatically placing
them when the shell becomes interactive.
To make sure the shell behaves correctly after its expected prompt
position changes, the terminal layer forces a resize event to be passed
to the shell on such (possibly) partial clears; this also has the nice
side effect of fixing the disappearing prompt on the preexisting "clear
including history" action: Fixes#4192.
The set_main_widget<T>() function only calls the construct and not
`try_create()` that the GMLCompiler generates so all calls to
`find_descendant_of_type_named()` would result in null
pointers that would resolve in a crash.
This is a fairly simple JSON format: A single array, containing objects,
with the Annotation fields as key:value pairs.
When reading a file, we let invalid or missing keys fall back to the
default values. This is mostly intended to set a pattern so that if we
add new fields in the future, we won't fail to load old annotations
files. If loading the file fails though, we keep the previously loaded
set of annotations.
The comment appears as a tooltip when hovering over the annotation.
A couple of properties of the TextEditor would ideally be set in GML,
but either don't have a setter exposed, or the GML compiler doesn't
recognise the enum. I'll fix those up after the current big GML
compiler PR gets merged.
Where it was straightforward to do so, I've updated the users to also
use ByteStrings for their file paths, but most of them have a temporary
String::from_byte_string() call instead.
We don't need to forward Cookie-related IPCs from the WebView through
the Tab to the BrowserWindow. We can skip the Tab like we do in other
chromes.
This is primarily to reduce overhead when changing Cookie IPCs in future
patches.
FontDatabase.h with its includes add up to quite a lot of code. In the
next commit, compiled GML files are going to need to access the
FontWeight enum, so let's allow them to do that without pulling in lots
of other things.
Also, change users to include FontWeight.h instead of FontDatabase.h
where appropriate.
This reverts commit 7c88ab2836.
Post-merge review revealed that the checks added in this PR prevent
any and all non-executable files from being displayed in Assistant,
which is a regression from intended functionality.
By default, a SpinBox's value should be unlimited, (or as close as we
can get to that,) and then the GML or code can impose a limit if
needed. This saves the developer from entering an arbitrary "big" max
value when they want the value to have no maximum.
I've audited the use of SpinBox and added `min: 0`, and removed a `max`,
where appropriate. All existing SpinBoxes constructed in code have a
range set explicitly as far as I can tell.
Allow the user to highlight sections of the edited document, giving them
arbitrary background colors. These annotations can be created from a
selection, or by manually specifying the start and end offsets.
Annotations can be edited or deleted by right-clicking them.
Any color can be used for the background. Dark colors automatically make
the text white for easier readability. When creating a new annotation,
we use whatever color the user last picked as this is slightly more
likely to be the one they want.
Icons contributed by Cubic Love.
Co-authored-by: Cubic Love <7754483+cubiclove@users.noreply.github.com>
Previously, constructing a `UnsignedBigInteger::from_base()` could
produce an incorrect result if the input string contained a valid
Base36 digit that was out of range of the given base. The same method
would also crash if the input string contained an invalid Base36 digit.
An error is now returned in both these cases.
Constructing a BigFraction from string is now also fallible, so that we
can handle the case where we are given an input string with invalid
digits.
`JsonValue::to_byte_string` has peculiar type-erasure semantics which is
not usually intended. Unfortunately, it also has a very stereotypical
name which does not warn about unexpected behavior. So let's prefix it
with `deprecated_` to make new code use `as_string` if it just wants to
get string value or `serialized<StringBuilder>` if it needs to do proper
serialization.
A bunch of users used consume_specific with a constant ByteString
literal, which can be replaced by an allocation-free StringView literal.
The generic consume_while overload gains a requires clause so that
consume_specific("abc") causes a more understandable and actionable
error.
Because of this the unary operations got applied to the result of the
operation-in-progress instead of the current argument as shown here:
`16 + 9 <sqrt> =`
Previous output: `sqrt(16 + 9)` = `5`
Expected output: `16 + sqrt(9)` = `19`
Instead of spawning these processes from the WebContent process, we now
create them in the Browser chrome.
Part 1/N of "all processes are owned by the chrome".
This is not necessary, as pressing Enter will activate the button
whether it is in focus or not. This makes the equals button behave the
same as all other buttons.
Previously, some non-default process statistics were displayed in a
long human readable format in the processes tab, which could negatively
affect readability. A shorter format is now used in the processes tab,
while the process properties window retains the longer format.
Before the key events for shortcuts such as ALT-F4 did not work as the
widget was swallowing up the shortcut. This changes ignores the event
if nothing else matched
- Register the widget
- Register property getters and setters
- Rename getters and setters to match the property names, as required by
the GML compiler. The names min/max/value are chosen to match SpinBox.
- Prevent a crash when the minimum is less than the maximum (which can
happen while editing the GML).
The calendar now verifies that the user can't input invalid date ranges,
i.e. date ranges where the end is before the start. The UI elements do
this implicitly, by adjusting the values when changing the dates in an
illegal way (e.g. when picking an end date that is before the start
date, the end date will change to the start date).
Previously, the EventManager stored the calendar events as a raw
JsonArray of objects. Now, we parse the JSON into a Vector<Event>
structure and store that in the EventManager. This makes it easier to
access the events from the outside, as you now don't have to know the
JSON structure anymore.
Since the meridiem selection was removed in the previous commit and
there currently doesn't seem to be a way to change the system time
format, I'm changing the SpinBoxes here to accept 24-hour format values
for now.
Previously, we had two versions of MonthListModel for the AddEventDialog
and the DatePickerDialog. Now, a unified version is in the Calendar.h
file, which can be used easily by anyone. Since that model and the
MeridiemListModel weren't used anymore in the AddEventDialog, I have
also removed them from there.
Instead of assigning, then sometimes reassigning these colors, set them
once. This makes it easier to see how we prioritize the different
factors that affect the styling.
`highlight_flag` is now `selected` since it represents if the byte is
within the current selection.
This change fixes the initial tool selection when pixelpaint is started
with a path. Previously an already existing editor was expected when
the default tool was initially propagated - which was not the case if
pixelpaint was launched to directly load an existing image.
These IPCs are different than other IPCs in that we can't just set up a
callback function to be invoked when WebContent sends us the screenshot
data. There are multiple places that would set that callback, and they
would step on each other's toes.
Instead, the screenshot APIs on ViewImplementation now return a Promise
which callers can interact with to receive the screenshot (or an error).
Instead of returning HeapBlock memory to the kernel (or a non-type
specific shared cache), we now keep a BlockAllocator per CellAllocator
and implement "deallocation" by basically informing the kernel that we
don't need the physical memory right now.
This is done with MADV_FREE or MADV_DONTNEED if available, but for other
platforms (including SerenityOS) we munmap and then re-mmap the memory
to achieve the same effect. It's definitely clunky, so I've added a
FIXME about implementing the madvise options on SerenityOS too.
The important outcome of this change is that GC types that use a
type-specific allocator become immune to use-after-free type confusion
attacks, since their virtual addresses will only ever be re-used for
the same exact type again and again.
Fixes#22274
Instead of using a scan code, which for scan code set 2 will not
represent the expected character mapping index, we could just use
another variable in the KeyEvent structure that correctly points to the
character index.
This change is mostly relevant to the KeyboardMapper application, and
also to the WindowServer code, as both handle KeyEvents and need to
use the character mapping index in various situations.
This is a cleanup patch, moves a chunk of repeated code to one place
instead of assigning variables with the same values twice in two
different places of code.
Currently if users select last bytes in HexEditor with mouse in either
Hex or Text mode, they will be able to move cursor on the byte outside
bounds. If then they try to write something in either of those modes,
app will crash.
This patch moves the recently added "replace" cursor to always be on the
last byte of the selection instead of being on the byte after the last
selected byte.
This patch changes cursor type from caret to black box for both Hex and
Text modes, because right now the way how blinking caret looks like is
more closer to "insert" mode in similar editors, whereas the real
behavior of this cursor is more of a "replace" mode seen in similar
editors like GHex.
In a bunch of cases, this actually ends up simplifying the code as
to_number will handle something such as:
```
Optional<I> opt;
if constexpr (IsSigned<I>)
opt = view.to_int<I>();
else
opt = view.to_uint<I>();
```
For us.
The main goal here however is to have a single generic number conversion
API between all of the String classes.
This commit un-deprecates DeprecatedString, and repurposes it as a byte
string.
As the null state has already been removed, there are no other
particularly hairy blockers in repurposing this type as a byte string
(what it _really_ is).
This commit is auto-generated:
$ xs=$(ack -l \bDeprecatedString\b\|deprecated_string AK Userland \
Meta Ports Ladybird Tests Kernel)
$ perl -pie 's/\bDeprecatedString\b/ByteString/g;
s/deprecated_string/byte_string/g' $xs
$ clang-format --style=file -i \
$(git diff --name-only | grep \.cpp\|\.h)
$ gn format $(git ls-files '*.gn' '*.gni')
This patch adds the new DynamicWidgetContainer for the right panel
elements. This allows the user to collapse, expand or even detach
widgets to a separate window. The collapsed or expanded state is
persisted so that they are restored after application startup.
With this change it is possible to shrink the size of widgets to a
minimum in order to give more space to other currently important
widgets.
This pr fixes a problem where the application would crash if the mouse
was moved while the editor was still loading the image. With this
change tool related mouse move events are discarded as long as there is
no active tool.
This changes how the initial tool is beeing set on application startup.
With this change the initial tool is set the same way as when the tool
is changed via the frontend. Thus it also updates the tool properties
window which wasn't done before leaving the properties empty as long
the tool was changed for the first time.
If the Inspector widget already exists, be sure to inspect the page when
it is re-opened. However, this should be a no-op if the page was already
inspected (as any existing Inspector will be reset if a new page load
began).
Note this is not an issue in the AppKit chrome.
