This change makes out-of-flow blocks to be considered for joining
only to anonymous blocks that have inline children. It finally
solved the problem that out-of-flow break anonymous blocks into
chunks causing wrong layout without regressing Acid2.
This patch does two things:
- We now use u32 instead of size_t for the hops and index fields
in EnvironmentCoordinate. This means we're limited to an environment
nesting level and variable count of 4Gs respectively.
- Instead of wrapping it in an Optional, EnvironmentCoordinate now has
a custom valid/invalid state using a magic marker value.
These two changes reduce the size of Identifier by 16 bytes. :^)
Before this change, each AST node had a 64-byte SourceRange member.
This SourceRange had the following layout:
filename: StringView (16 bytes)
start: Position (24 bytes)
end: Position (24 bytes)
The Position structs have { line, column, offset }, all members size_t.
To reduce memory consumption, AST nodes now only store the following:
source_code: NonnullRefPtr<SourceCode> (8 bytes)
start_offset: u32 (4 bytes)
end_offset: u32 (4 bytes)
SourceCode is a new ref-counted data structure that keeps the filename
and original parsed source code in a single location, and all AST nodes
have a pointer to it.
The start_offset and end_offset can be turned into (line, column) when
necessary by calling SourceCode::range_from_offsets(). This will walk
the source code string and compute line/column numbers on the fly, so
it's not necessarily fast, but it should be rare since this information
is primarily used for diagnostics and exception stack traces.
With this, ASTNode shrinks from 80 bytes to 32 bytes. This gives us a
~23% reduction in memory usage when loading twitter.com/awesomekling
(330 MiB before, 253 MiB after!) :^)
Making floats to join anonymous block caused regressions in
Acid2 Test so let's leave it to be only absolute blocks who
might be joined into anonymous block when possible.
We can't be nuking the ESO while its owned execution context is still on
the VM's execution context stack, as that may lead to a use-after-free.
This patch solves this by adding a `context_owner` field to each context
and treating it as a GC root.
When timeouts are implemented, the start node used to find elements may
not remain valid for the entire duration of the timeout. For example,
the active document element may change, or the start node may be removed
from the DOM.
To handle this, we will need to re-evaluate the start node on each
iteration of the find() operation. This patch wraps the steps to do so
in a lambda to be executed on each iteration.
This getter and setter were previously labelled as a "hack" and used to
disable style invalidation on attribute changes during the HTML parsing
phase (as it caused big sites's loading to be slow). These functions
are currently not used, so they can be removed:^)
This commit adds inline spec comments to the part of the parser that
ends up calling HTMLScriptElement::prepare().
The code is tweaked to match the spec more closely.
This change makes calculate_static_position to return content box
for both x and y (at least for the case when children are not inline).
It makes it possible to be consistent about x and y when calculating
box offset inside layout_absolutely_positioned_element.
We were mistakenly trying to append UTF-16 code units to a StringBuilder
via the append(char) API. This patch fixes that by accumulating the
result in a Vector<u16> instead.
This'll be a bit worse for performance, since we're now doing additional
UTF-16 string conversions, but we're going for correctness at this stage
and can worry about performance later.
Because of how we output lines in the loop above, if we leave that loop
when the last line was not the same in both files, then either
`file1_line` or `file2_line` has not been output yet.
`process_remaining()` does not print that line either, since it
immediately reads a new line. So, output the previously-missing line
before we call that. :^)
These are based on the ones in Firefox:
`$(selector, element = document)`:
Equivalent to `element.querySelector(selector)`.
`$$(selector, element = document)`:
Equivalent to `element.querySelectorAll(selector)`.
This holds the return value of the expression that was last entered into
the browser console. If that last expression returned an error of some
kind, `$_` will be `undefined`. This matches the behaviour in Firefox.
This allows us to expose extra functions and properties to the console,
such as `$0`, without them being available to website scripts.
`ConsoleEnvironmentSettingsObject` is basically a stub, since we require
an `EnvironmentSettingsObject` but it has abstract methods.
If the page's URL is changed from within the WebContent process (e.g.
window.location.href is set from JavaScript, or a navigation is started
from WebDriver), we would not store the URL at all within OOPWV. This
prevented actions like reloading from working properly, as the browser
would reload whatever URL was previously entered in the URL box.
Two Blocking modals in the same modal chain which aren't descended
one from the other will block each other's input rendering the chain
noninteractive. This has caused confusion in the past for builders so
this warning makes the behavior explicitly forbidden.
This was causing a slight delay when opening ComboBox ListViews.
As an easy first optimization, don't bother computing this at all
for frameless windows and those not type Normal.
Makes the Audio applet, Taskbar clock, CommandPalette, EmojiPicker,
and Assistant work as Popup windows. Popups are frameless, unmovable,
and unresizable by default, in addition to their preemptive function.
Also sets Assistant not to obey widget min size so its search result
area resizes correctly
and the CaptureInput mode. They are a source of unneeded complexity
in WindowServer and have proven prone to regressions, so this patch
replaces them with a simple input preemption scheme using Popups.
Popup windows now have ergonomics similar to menus: When open,
a popup preempts all mouse and key events for the entire window
stack; however, they are fragile and will close after WindowServer
swallows the first event outside them. This is similar to how combo
box windows and popups work in the classic Windows DE and has the
added benefit of letting the user click anywhere to dismiss a popup
without having to worry about unwanted interactions with other
widgets.
with WindowInput{Preempted,Restored} Events and allow Widgets to save
the state of their focus preemption. As of now, only Popups will
preempt input and trigger these events.
For flate and lzw filters, the data can be transformed by this
predictor function to make it compress better. For us this means that
we have to undo this step in order to get the right result.
Although this feature is meant for images, I found at least a few
documents that use it all over the place, making this step very
important.
We previously compared two unrelated values to determine if we parsed
the xref table to completion. We now check if we added every subsection
instead, and double check to make sure we never read past the end.
This filter basically tells us that we are dealing with a JPEG.
Note that by serializing the resulting image we assume that this filter
is the last one in the chain, everything else would be highly unlikely.
We currently don't support ICC color spaces and fall back to a "simple"
one instead.
If no alternative is specified however, we are allowed to pick the
closest match based on the number of color components.
This fixes an issue where we would show an error message on every
subsequent paint event, making it impossible to close or use the
application further.
Some more architectural work is needed if we want to invalidate the
sidebar entries as well.
And TRY early during initialization in FontEditor to leave the app
in a valid state on error. Fixes OOM crashes when cloning the original
font for highlight modifications.
Like other apps, FontEditor now organizes widget visibility toggles
under a View->Layout submenu. The main toolbar and the status bar
can now be turned on and off.
Each GenericInterruptHandler now tracks the number of calls that each
CPU has serviced.
This takes care of a FIXME in the /sys/kernel/interrupts generator.
Also, the lsirq command line tool now displays per-CPU call counts.
These lambdas were marked mutable as they captured a Ptr wrapper
class by value, which then only returned const-qualified references
to the value they point from the previous const pointer operators.
Nothing is actually mutating in the lambdas state here, and now
that the Ptr operators don't add extra const qualifiers these
can be removed.
Even if the pointer value is const, the value they point to is not
necessarily const, so these functions should not add the qualifier.
This also removes the redundant non-const implementations of these
operators.
These functions are now implemented in terms of getpwent_r() which
allows us to remove two FIXMEs about global variable shenanigans.
I'm also adding tests for both APIs. :^)
This gives much better visual results than painting the path directly.
It also has the nice side effect that Type 1 fonts will now look much
more similar to TrueType fonts, which use the same class :^)
In addition, we can now cache glyph bitmaps for repeated use.
The custom TTF path rasterizer is actually generic enough for it to be
used for other fonts. To make this more clear, it now lives on its own
in the "Font" directory.
Registers the PreviewWidget for addition directly into GML. Fixes
its previous double Frame borders. Also standardizes the Apply and
Reset buttons as DialogButtons and spaces them consistently with
other apps. Gives the TabWidget some tasteful container margins.
Right now only the accessibility menu is shared across apps but
it's a good bit of boilerplate that can be consolidated and will
make adding ColorBlindnessFilters to other widgets trivial.
Still some TODOs here:
* We don't handle all capabilities (e.g. proxy)
* We don't match the capabilities against the running browser
But this will parse the capabilities JSON object received from the
WebDriver client.
The spec's text is pretty awkward here, but the way we've currently
transcribed it to C++ means we reject valid script timeouts. This meant
the following would fail:
TimeoutsConfiguration config {}; // Default values.
auto json = timeouts_object(config);
config = TRY(json_deserialize_as_a_timeouts_configuration(json));
Since ff2f31b LibWeb has segfaulted when clicking on links, as the
browsing context (a GCPtr) in the lambda was captured by reference
and was out of scope by the time the callback fired.
Before this could give the `must be followed by in` error before the
undeclared private identifier error. Fixing the `in` error would not
have resolved the other error so this order makes the errors more
actionable.
If we don't check that a private identifier is valid this can break the
assumption that we have a private environment when evaluation the
private identifier. Also an unknown private identifier this should
be a SyntaxError.
This is only visible with something like `Object.getOwnPropertyNames` on
the global object. All other declaration instantiations put the
functions on an environment making the order invisible.
Note that spec order is not quite tree order as in non-strict mode
functions which get hoisted out of blocks appear before top level
functions.
Co-authored-by: Hendiadyoin1 <leon.a@serenityos.org>
Before putting `throw 1;` in a test file did not fail the file, and it
ignored the rest of the file. Now we check the resulting completion from
run and fail if that is an error.
The way in which dialogs should be handled is configurable by the driver
capabilities object, which we don't support yet. So this implements just
the default mode to dismiss the dialog and return an error if there is
one open.
In the OOPWV, this means we need to refer to the dialog after it has
been open, so we now hold a pointer to whatever dialog is open.
This will be needed for WebDriver, which will require constructing and
controlling dialogs manually. Currently, InputBox will only set its text
value when the OK button is pressed. This changes InputBox to update its
text when done(ExecResult::OK) is invoked in any way.
This also makes the text_value() method public, allows for setting the
text value, and allows for moving-in the initial text value.
Currently, the WebContent process is completely blocked while waiting
for a response to a dialog request. This patch allows the IPC event loop
to continue executing while only blocking the HTML event loop.
This will allow other processes like WebDriver to continue to operate on
the WebContent process while a dialog is open.
We now layout foreign objects as if they form a nested block formatting
context. This probably isn't the most correct way to do this, but it's
a start.
This fixes an error when using auto-fit with grid-gap, as previously
were not taking into account the fact that more columns had been added
to the grid yet the occupation grid had not grown.
Previously were incorrectly initializing the grid gap tracks as were
using the GridSize float constructor, which creates a flexible length.
What is actually wanted is a fixed-size track of a certain size,
indicated in pixels.
This patch optimizes the drawing of aa-lines by rotating the drawn
rectangle to the direction where the line points. That enables us to
draw non-straight lines with the proper width.
If a aa-line is drawn that is infact a straigt line we now dont plot
those lines with multipe rectangles across the line - insted we draw
straight lines in one go with just one rectangle of proper size.
Stroking of lines has been enhanced to take care of the edges between
two lines with drawing the stroke till the intersections of two
connected lines.
