The spec is unclear about when exactly we should parse the style sheet.
Previously we'd do so before calling this algorithm, which was
error-prone, as seen by the bug fixed by the previous commit. The spec
for step 1 of "create a CSS style sheet" says:
1. Create a new CSS style sheet object and set its properties as
specified.
The definitions linked are UA-defined enough that it seems reasonable to
put the parsing here. That simplifies the user code a little and makes
it harder to mess up. It does raise the question of what to do if
parsing fails. I've matched our previous behaviour by just logging and
returning in that case.
While I'm modifying this method, I've also converted the bool params to
enums so they're a little clearer to read.
Before this change, we assigned the style sheet's location *after* its
content rules were parsed and added to it. This meant any `@import`s
would try to fetch their style sheet before they knew the URL they
should fetch it relative to.
This change adds “default step” and “step scale factor” handling for all
remaining HTMLInputElement input types for which the spec defines such
and that we didn’t yet have handling for.
With this change we save a copy of of scroll state at the time of
recording a display list, instead of actual ScrollState pointer that
could be modifed by the main thread while display list is beings
rasterized on the rendering thread, which leads to a frame painted with
inconsistent scroll state.
Fixes https://github.com/LadybirdBrowser/ladybird/issues/4288
This will be needed by Streams. To support this, we now store callbacks
in a hash table, keyed by an ID. Callers may use that ID to remove the
callback at a later point.
While debugging a spec-compliant implementation of ReadableStreamPipeTo,
I spent a lot of time inspecting promise internals. This is much less
noisy if we halve the number of temporary promises.
To prepare for introducing a CSS::URL type, we need to qualify any use
of LibURL as `::URL::foo` instead of `URL::foo` so the compiler doesn't
get confused.
Many of these uses will be replaced, but I don't want to mix this in
with what will likely already be a large change.
Instead of wrapping all non-movable members of TransportSocket in OwnPtr
to keep it movable, make TransportSocket itself non-movable and wrap it
in OwnPtr.
We were handling removing the style sheet from the shadow root, but not
appending to it. Fixing this also revealed a bug that a removed link
element would always try to remove from the document's list, as the
root is no longer the shadow root it's in. The fix is to use the passed
in old root to remove the style sheet from.
Fixes the cookie banner on https://nos.nl/
There's a bit of a UTF-8 assumption with this change. But nearly every
caller of these methods were immediately creating a String from the
resulting ByteString anyways.
With this change, the responsibility for prepending messages with their
size and ensuring the entire message is received before returning it to
the caller is moved to TransportSocket. This removes the need to
duplicate this logic in both LibIPC and MessagePort.
Another advantage of reducing message granularity at IPC::Transport
layer is that it will make it easier to support alternative transport
implementations (like Mach ports, which unlike Unix domain sockets are
not stream oriented).
The special empty value (that we use for array holes, Optional<Value>
when empty and a few other other placeholder/sentinel tasks) still
exists, but you now create one via JS::js_special_empty_value() and
check for it with Value::is_special_empty_value().
The main idea here is to make it very unlikely to accidentally create an
unexpected special empty value.
Start work on a speculative HTML Parser in Swift. This component will
walk ahead of the normal HTML parser looking for fetch() requests to
make while the normal parser is blocked. This work exposed many holes in
the Swift C++ interop component, which have been reported upstream.
Skia has a check in debug mode to verify that surface is only used
within one thread. Before this change we were violating this by
allocating surfaces on the main thread while using and destructing them
on the rendering thread.
It turned out that some web applications want to send fairly large
messages to WebWorker through IPC (for example, MapLibre GL sends
~1200KiB), which led to failures (at least on macOS) because buffer size
of TransportSocket is limited to 128KiB. This change solves the problem
by wrapping messages that exceed socket buffer size into another message
that holds wrapped message content in shared memory.
Co-Authored-By: Luke Wilde <luke@ladybird.org>
This is very clearly a very dangerous API to have, and was causing
a crash on Linux as a result of a stack use-after-free when visiting
https://www.index.hr/.
Fixes#3901
Deleteing set_surface() makes DisplayListPlayer API a bit more intuitive
because now caller doesn't have to think whether it's necessary to
restore previous surface after execution, instead DisplayListPlayer
takes care of it by maintaining a stack of surfaces.
The display list is an immutable data structure, so once it's created,
rasterization can be moved to a separate thread. This allows more room
for performing other tasks between processing HTML rendering tasks.
This change makes PaintingSurface, ImmutableBitmap, and GlyphRun atomic
ref-counted, as they are shared between the main and rendering threads
by being included in the display list.
Previously, all `GC::Cell` derived classes were Weakable. Marking only
those classes that require this functionality as Weakable allows us to
reduce the memory footprint of some frequently used classes.
PrimitiveString is now internally either UTF-8, UTF-16, or both.
We no longer convert them to/from ByteString anywhere, nor does VM have
a ByteString cache.
This fixes the frame-ancestors WPT tests from crashing when an iframe
is blocked from loading. This is because it would get an undefined
location.href from the cross-origin iframe, which causes a crash as it
expects it to be there.
Instead of reparsing the style attributes every time we instantiate
the internal shadow tree for a text input element, we now parse them
once (in the internal CSS realm) and reuse them for all elements.
Roughly a ~10% speedup on Speedometer 2.1 :^)
The upcoming generated types will match those for pseudo-classes: A
PseudoElementSelector type, that then holds a PseudoElement enum
defining what it is. That enum will be at the top level in the Web::CSS
namespace.
In order to keep the diffs clearer, this commit renames and moves the
types, and then a following one will replace the handwritten enum with
a generated one.
Instead of bothering the GC heap with a bunch of DOMRect allocations,
we can just pass around CSSPixelRect internally in many cases.
Before this change, we were generating so much DOMRect garbage that
we had to do a garbage collection *every frame* on the Immich demo.
This was due to the large number of intersection observers checked.
We still need to relax way more when idle, but for comparison, before
this change, when doing nothing for 10 seconds on Immich, we'd spend
2.5 seconds updating intersection observers. After this change, we now
spend 600 ms.
When decoding data into bitmaps, we end up with different alpha types
(premultiplied vs. unpremultiplied color data). Unfortunately, Skia only
seems to handle premultiplied color data well when scaling bitmaps with
an alpha channel. This might be due to Skia historically only supporting
premultiplied color blending, with unpremultiplied support having been
added more recently.
When using Skia to blend bitmaps, we need the color data to be
premultiplied. ImmutableBitmap gains a new method to enforce the alpha
type to be used, which is now used by SharedResourceRequest and
CanvasRenderingContext2D to enforce the right alpha type.
Our LibWeb tests actually had a couple of screenshot tests that exposed
the graphical glitches caused by Skia; see the big smiley faces in the
CSS backgrounds tests for example. The failing tests are now updated to
accommodate the new behavior.
Chromium and Firefox both seem to apply the same behavior; e.g. they
actively decode PNGs (which are unpremultiplied in nature) to a
premultiplied bitmap.
Fixes#3691.
This adds a basic settings page to manage persistent Ladybird settings.
As a first pass, this exposes settings for the new tab page URL and the
default search engine.
The way the search engine option works is that once search is enabled,
the user must choose their default search engine; we do not apply any
default automatically. Search remains disabled until this is done.
There are a couple of improvements that we should make here:
* Settings changes are not broadcasted to all open about:settings pages.
So if two instances are open, and the user changes the search engine
in one instance, the other instance will have a stale UI.
* Adding an IPC per setting is going to get annoying. It would be nice
if we can come up with a smaller set of IPCs to send only the relevant
changed settings.
When the return key is pressed, we try to handle it as a commit action
for input elements. However, we would then go on to actually insert the
return key's code point (U+000D) into the input element. This would be
sanitized out, but would leave the input element in a state where it
thinks it has text to commit. This would result in a change event being
fired when the return key is pressed multiple times in a row.
We were also firing the beforeinput/input events twice for all return
key presses.
To fix this, this patch changes the input event target to signify if it
actually handled the return key. If not (i.e. for textarea elements),
only then do we insert the code point. We also must not fall through to
the generic key handler, to avoid the repeated input events.
Previously, we would only invalidate style when setting the `media` IDL
attribute; changing the attribute via `setAttribute()` and
`removeAttribute()` had no immediate effect.
There are two changes happening here: a correctness fix, and an
optimization. In theory they are unrelated, but the optimization
actually paves the way for the correctness fix.
Before this commit, the HTML tokenizer would attempt to look for named
character references by checking from after the `&` until the end of
m_decoded_input, which meant that it was unable to recognize things like
named character references that are inserted via `document.write` one
byte at a time. For example, if `∉` was written one-byte-at-a-time
with `document.write`, then the tokenizer would only check against `n`
since that's all that would exist at the time of the check and therefore
erroneously conclude that it was an invalid named character reference.
This commit modifies the approach taken for named character reference
matching by using a trie-like structure (specifically, a deterministic
acyclic finite state automaton or DAFSA), which allows for efficiently
matching one-character-at-a-time and therefore it is able to pick up
matching where it left off after each code point is consumed.
Note: Because it's possible for a partial match to not actually develop
into a full match (e.g. `¬indo` which could lead to `⋵̸`),
some backtracking is performed after-the-fact in order to only consume
the code points within the longest match found (e.g. `¬indo` would
backtrack back to `¬`).
With this new approach, `document.write` being called one-byte-at-a-time
is handled correctly, which allows for passing more WPT tests, with the
most directly relevant tests being
`/html/syntax/parsing/html5lib_entities01.html`
and
`/html/syntax/parsing/html5lib_entities02.html`
when run with `?run_type=write_single`. Additionally, the implementation
now better conforms to the language of the spec (and resolves a FIXME)
because exactly the matched characters are consumed and nothing more, so
SWITCH_TO is able to be used as the spec says instead of RECONSUME_IN.
The new approach is also an optimization:
- Instead of a linear search using `starts_with`, the usage of a DAFSA
means that it is always aware of which characters can lead to a match
at any given point, and will bail out whenever a match is no longer
possible.
- The DAFSA is able to take advantage of the note in the section
`13.5 Named character references` that says "This list is static and
will not be expanded or changed in the future." and tailor its Node
struct accordingly to tightly pack each node's data into 32-bits.
Together with the inherent DAFSA property of redundant node
deduplication, the amount of data stored for named character reference
matching is minimized.
In my testing:
- A benchmark tokenizing an arbitrary set of HTML test files was about
1.23x faster (2070ms to 1682ms).
- A benchmark tokenizing a file with tens of thousands of named
character references mixed in with truncated named character
references and arbitrary ASCII characters/ampersands runs about 8x
faster (758ms to 93ms).
- The size of `liblagom-web.so` was reduced by 94.96KiB.
Some technical details:
A DAFSA (deterministic acyclic finite state automaton) is essentially a
trie flattened into an array, but it also uses techniques to minimize
redundant nodes. This provides fast lookups while minimizing the
required data size, but normally does not allow for associating data
related to each word. However, by adding a count of the number of
possible words from each node, it becomes possible to also use it to
achieve minimal perfect hashing for the set of words (which allows going
from word -> unique index as well as unique index -> word). This allows
us to store a second array of data so that the DAFSA can be used as a
lookup for e.g. the associated code points.
For the Swift implementation, the new NamedCharacterReferenceMatcher
was used to satisfy the previous API and the tokenizer was left alone
otherwise. In the future, the Swift implementation should be updated to
use the same implementation for its NamedCharacterReference state as
the updated C++ implementation.
The intent is that this will replace the separate Task Manager window.
This will allow us to more easily add features such as actual process
management, better rendering of the process table, etc. Included in this
page is the ability to sort table rows.
This also lays the ground work for more internal `about` pages, such as
about:config.
We previously had PropertyOwningCSSStyleDeclaration and
ResolvedCSSStyleDeclaration, representing the current style properties
and resolved style respectively. Both of these were the
CSSStyleDeclaration type in the CSSOM. (We also had
ElementInlineCSSStyleDeclaration but I removed that in a previous
commit.)
In the meantime, the spec has changed so that these should now be a new
CSSStyleProperties type in the CSSOM. Also, we need to subclass
CSSStyleDeclaration for things like CSSFontFaceRule's list of
descriptors, which means it wouldn't hold style properties.
So, this commit does the fairly messy work of combining these two types
into a new CSSStyleProperties class. A lot of what previously was done
as separate methods in the two classes, now follows the spec steps of
"if the readonly flag is set, do X" instead, which is hopefully easier
to follow too.
There is still some functionality in CSSStyleDeclaration that belongs in
CSSStyleProperties, but I'll do that next. To avoid a huge diff for
"CSSStyleDeclaration-all-supported-properties-and-default-values.txt"
both here and in the following commit, we don't apply the (currently
empty) CSSStyleProperties prototype yet.
This also implements the `:high-value` and `:low-value` that are in the
spec.
Same note as before about this being based on the very-drafty CSS Forms
spec. In fact, some of this isn't even in that spec yet. Specifically,
the `:suboptimal-value` and `:even-less-good-value` names are undecided
and subject to change. However, it's clear that this is a pseudo-class
situation, not a pseudo-element one, so I think this is still an
improvement, as it allows styling of the `::fill` pseudo-element
regardless of what state it is in.
Relevant spec issue: https://github.com/openui/open-ui/issues/1130
This spec is very early on, and likely to change. However, it still
feels preferable to use these rather than the prefixed -webkit ones.
Plus, as we have a `::fill` on range inputs, we can use that for styling
the bar instead of inserting CSS from C++.
A Storage object may be created with an existing storage bottle. For
example, if you navigate from site.com/page1 to site.com/page2, they
will have different localStorage objects, but will use the same bottle
for actual storage.
Previously, if page1 set some key/value item, we would initialize the
byte count to 0 on page2 despite having a non-empty bottle. Thus, if
page2 set a smaller value with the same key, we would overflow the
computed byte count, and all subsequent writes would be rejected.
This was seen navigating from the chess.com home page to the daily
puzzle page.
Exactly one place seems to define these at the moment: service worker
clients. Since we don't have a type for these and just use
EnvironmentSettingsObject, I've put it there.
This allows us to parse the Content-Security-Policy header and
Referrer-Policy header from navigation responses and actually allow
them to start having an effect.
Site isolation is a common technique to reduce the chance that malicious
sites can access data from other sites. When the user navigates, we now
check if the target site is the same as the current site. If not, we
instruct the UI to perform the navigation in a new WebContent process.
The phrase "site" here is defined as the public suffix of the URL plus
one level. This means that navigating from "www.example.com" to
"sub.example.com" remains in the same process.
There's plenty of room for optimization around this. For example, we can
create a spare WebContent process ahead of time to hot-swap the target
site. We can also create a policy to keep the navigated-from process
around, in case the user quickly navigates back.
When setting scroll position during page load we need to consider
whether we actually have a fragment to scroll to. A script may already
have run at that point and may already have set a scroll position.
If there is an actual fragment to scroll to, it is fine to scroll to
that fragment, since it should take precedence. If we don't have a
fragment however, we should not unnecessarily overwrite the scroll
position set by the script back to (0, 0).
Since this problem is caused by a spec bug, I have tested the behavior
in the three major browsers engines. Unfortunately they do not agree
fully with each other. If there is no fragment at all (e.g. `foo.html`),
all browsers will respect the scroll position set by the script. If
there is a fragment (e.g. `foo.html#bar`), all browsers will set the
scroll position to the fragment element and ignore the one set by
script. However, when the fragment is empty (e.g. `foo.html#`), then
Blink and WebKit will set scroll position to the fragment, while Gecko
will set scroll position from script. Since all of this is ad-hoc
behavior anyway, I simply implemented the Blink/WebKit behavior because
of the majority vote for now.
This fixes a regression introduced in 51102254b5.
This removes a couple of places where we were constructing strings or
vectors just to transfer data over IPC. And passes some values by const&
to remove clangd noise.
12c6ac78e2 with fixed mistake when cache
slot is copied instead of being referenced:
```cpp
auto cache =
box.cached_intrinsic_sizes().min_content_height.ensure(width);
```
while it should've been:
```cpp
auto& cache =
box.cached_intrinsic_sizes().min_content_height.ensure(width);
```
This change moves intrinsic sizes cache from
LayoutState, which is local to current layout run,
to layout nodes, so it could be reused between
layout runs. This optimization is possible because
we can guarantee that these measurements will
remain unchanged unless the style of the element
or any of its descendants changes.
For now, invalidation is implemented simply by
resetting cache on whole ancestors chain once we
figured that element needs layout update.
The case when layout is invalidated by DOM's
structural changes is covered by layout tree
invalidation that drops intrinsic sizes cache
along with layout nodes.
I measured improvement on couple websites:
- Mail list on GMail 28ms -> 6ms
- GitHub large code page 47ms -> 36ms
- Discord chat history 15ms -> 8ms
(Time does not include `commit()`)
This was completely unnecessary, and we can just let the internal
DOM tree changes trigger partial layout updates instead.
Noticed we were repeatedly dropping layout trees on ChatGPT and this
was one of the culprits.
This change implements the requirements for the “suffering from an
overflow” and “suffering from an underflow” algorithms for
HTMLInputElement constraint validation.
This patch removes those unused 2 algorithms:
1. `fetch_internal_module_script_graph`
2. `fetch_descendants_of_a_module_script`
Those 2 algorithms were removed in spec and are not used in our
codebase.
The User Timing, Resource Timing and Navigation Timing specifications
have not been updated to account for the queue method to also append to
the underlying performance buffer and it's method of checking it's
full.
This separates the functionality into a different function, with a flag
to indicate whether to use the custom full buffer logic or not.
This change fixes a bug in our implementation of the “step base”
algorithm at https://html.spec.whatwg.org/#concept-input-min-zero. We
were using the “value” IDL/DOM attribute in a particular step, where the
spec instead actually requires using the “value” content attribute.
This change implements HTMLInputElement type=email constraint validation
in conformance with the current spec requirements (which happens to also
produce behavior that’s interoperable with other existing engines).
This is a bit weird in the spec in it passing through a string here
instead of a URL record. However, the string being used in this
case should only ever be a valid URL string if it is not the empty
string.
Instead of relying on the implicit URL constuctor. Parsing should
never fail here as urlString before adding the suffix is already
parsed above, and the suffix should only be a valid query string.
This is the same type as what is spec'd. We cannot use a URL record
for this member as the spec in some scenarios will set and compare
the URL string to an invalid URL value, such as the empty string.
With implicit string constructors for the URL class removed
explicitly using URL::Parser::basic_parse makes the code look
quite silly in those places.
This change implements HTMLInputElement type=url constraint validation
in such a way as to match the behavior in other existing engines (which
is, however, very different from what the spec currently requires).
The DOMParsing spec is in the process of being merged into the HTML one,
gradually. The linked spec change moves XMLSerializer, but many of the
algorithms are still in the DOMParsing spec so I've left the links to
those alone.
I've done my best to update the GN build but since I'm not actually
using it, I might have done that wrong.
Corresponds to 2edb8cc7ee
These form the basis of Content Security Policy. A policy is a
collection of directives that are parsed from either the
Content-Security-Policy(-Report-Only) HTTP header, or the `<meta>`
element.
The directives are what restrict the operations can be performed in the
current global execution context. For example, "frame-ancestors: none"
tells us to prevent the page from being loaded in an embedded context,
such as `<iframe>`.
You can see it a bit like OpenBSD's pledge() functionality, but for the
web platform: https://man.openbsd.org/pledge.2
Note that "becomes browsing-context connected" is defined as:
> When the insertion steps are invoked with it as the argument and it is
> now browsing-context connected.
This fixes an issue where WPT editing tests would clone the entire DOM
thousands of times and re-fetch all the linked CSS files once per clone.
Otherwise finalization step of initial `about:blank` navigation might
cancel user-initiated navigations by changing ongoing navigation id.
This is implemented by marking navigable as ready to start processing
navigation in SHTQ task, because we know for sure this task cannot be
processed until finalization step of initial `about:blank` navigation
is done.
This commit implements the main "render blocking" behavior for link
elements, drastically reducing the amount of FOUC (flash of unstyled
content) we subject our users to.
The document will now block rendering until linked style sheets
referenced by parser-created link elements have loaded (or failed).
Note that we don't yet extend the blocking period until "critical
subresources" such as imported style sheets have been downloaded
as well.
This change fixes a bug that can be reproduced with the following steps:
```js
const iframe = document.createElement("iframe");
document.body.appendChild(iframe);
iframe.contentWindow.location.href = ("http://localhost:8080/demo.html");
```
These steps are executed in the following order:
1. Create iframe and schedule session history traversal task that adds
session history entry for the iframe.
2. Generate navigation id for scheduled navigation to
`http://localhost:8080/demo.html`.
3. Execute the scheduled session history traversal task, which adds
session history entry for the iframe.
4. Ooops, navigation to `http://localhost:8080/demo.html` is aborted
because addings SHE for the iframe resets the navigation id.
This change fixes this by delaying all navigations until SHE for a
navigable is created.
We hold a raw pointer to the mouse selection target, which is a mixin-
style class inherited only by JS::Cell classes. By not visiting this
object, we sometime had a dangling reference to it after it had been
garbage collected.
Previously, they would stay open for the entire WebContent lifetime,
or until the server closed the connection. This was particularly
noticeable on collaborative websites/games such as
https://jigsawpuzzles.io/, where the user using Ladybird would stick
around even after they had navigated away.
This change — part of the HTML constraint-validation API (aka
“client-side form validation”) — implements the willValidate IDL/DOM
attribute/property for all form controls that support it.
Previously, the charset of name "UTF-16BE/LE" would be checked against
when following standards to convert the charset to UTF-8, but in
reality, the charsets "UTF-16BE" and "UTF-16LE" should be checked
separately.
Co-authored-by: Jelle Raaijmakers <jelle@ladybird.org>
This reduces the number of `.cpp` files that need to be recompiled when
one of the below header files changes as follows:
CSS/ComputedProperties.h: 1113 -> 49
CSS/ComputedValues.h: 1120 -> 209
This reduces the number of `.cpp` files that need to be recompiled when
one of the below header files changes as follows:
Painting/Command.h: 1030 -> 61
Painting/DisplayList.h: 1030 -> 60
Painting/DisplayListRecorder.h: 557 -> 59
One point to note is that I am not entirely sure what the result
of the pre-existing valueAsNumber test should be for this strange
case which does not lie exactly on a week/day boundary. Chrome
gives a negative timestamp, which seems more wrong than the result
we give, and neither gecko or WebKit appear to support the 'week'
type. So I'm considering this result acceptable for now, and this
may be something that will need more WPT tests added in the future.
Corresponds to part of https://github.com/whatwg/html/pull/9841 and then
https://github.com/whatwg/html/pull/11047
Adding `Auto` as a type state feels a little odd, as it's not an actual
type allowed in HTML. However, it's the default state when the value is
missing or invalid, which works out the same, as long as we never
serialize "auto", which we don't.
This is required to store Content Security Policies, as their
Directives are implemented as subclasses with overridden virtual
functions. Thus, they cannot be stored as generic Directive classes, as
it'll lose the ability to call overridden functions when they are
copied.
We were incorrectly treating cellpadding=0 as if the attribute was
missing. This commit fixes it so it behaves as `padding: 0` on cells.
When adding a test, I discovered that we were not invalidating style for
cells when their containing table's cellpadding attribute changed.
So this commit fixes that as well.
There's a quirk in HTML where the parser should ignore any line feed
character immediately following a `pre` or `textarea` start tag.
This was working fine when we could peek ahead in the input stream and
see the next token, but didn't work in character-at-a-time parsing with
document.write().
This commit adds the "can ignore next line feed character" as a parser
flag that is maintained across invocations, making it work in this
parsing mode as well.
20 new passes in WPT/html/syntax/parsing/ :^)
Instead of always inserting a new text node, we now continue appending
to an extisting text node if the parser's character insertion point is
a suitable text node.
This fixes an issue where multiple invocations of document.write() would
create unnecessary sequences of text nodes. Such sequences are now
merged automatically.
19 new passes in WPT/html/syntax/parsing/ :^)
We were neglecting to return after handling the `frameset` start tag,
which caused us to process it twice, once properly and once generically.
54 new passes in WPT/html/syntax/parsing/ :^)
Before this change, the explicit EOF inserted by document.close() would
instantly abort the parser. This meant that parsing algorithms that ran
as part of the parser unwinding on EOF would never actually run.
591 new passes in WPT/html/syntax/parsing/ :^)
This exposed a problem where the parser would try to insert a root
<html> element on EOF in a document where someone already inserted such
an element via direct DOM manipulation. The parser now gracefully
handles this scenario. It's covered by existing tests (which would
crash without this change.)
When constructing an entry list, XHR::FormDataEntry is created
manually and appended to the entry list instead of using the
spec-defined method of creating an entry.
We only need a Page for file:// urls. At some point we probably
needed it for other kinds of requests, but the current functionality
doesn't need to store the Page pointer on the ResourceLoader.
Our existing coalescing mechanism for input events didn't prevent
multiple mousemove/mousewheel events from being processed between paint
cycles. Since handling these events can trigger style & layout updates
solely for hit-testing purposes, we might end up doing work that won't
be observable by a user and could be avoided by shceduling input events
processing to happen right before painting the next frame.
MediaQueryList will now remember if a state change occurred when
evaluating its match state. This memory can then be used by the document
later on when it's updating all queries, to ensure that we don't forget
to fire at least one change event.
This also required plumbing the system visibility state to initial
about:blank documents, since otherwise they would be stuck in "hidden"
state indefinitely and never evaluate their media queries.
We were previously assuming that dictionary members were always
required when being returned.
This is a bit of a weird case, because unlike _input_ dictionaries
which the spec marks as required, 'result' dictionaries do not seem to
be marked in spec IDL as required. This is still fine from the POV that
the spec is written as it states that we should only be putting the
values into the dictionary if the value exists.
We could do this through some metaprogramming constexpr type checks.
For example, if the type in our C++ representation was not an
Optional, we can skip the has_value check.
Instead of doing that, change the IDL of the result dictionaries to
annotate these members so that the IDL generator knows this
information up front. While all current cases have every single
member returned or not returned, it is conceivable that the spec
could have a situation that one member is always returned (and
should get marked as required), while the others are optionally
returned. Therefore, this new GenerateAsRequired attribute is
applied for each individual member.
This fixes a crash in the included test that regressed in 0adf261,
and is hit by the following HTML:
```html
<body></body>
<script>
const frame = document.body.appendChild(document.createElement("iframe"));
frame.contentDocument.open();
const child = frame.contentDocument.createElement("html")
const html = frame.contentDocument.appendChild(child);
frame.contentDocument.close();
</script>
```
I am not 100% sure this is fully the correct fix and there are other
cases which would not work properly. But it's definitely an improvement
to make the confuisingly named 'insert_an_eof' function of the tokenizer
actually do something.
Before this change, tasks associated with a destroyed document would get
stuck in the task queue forever, since document-associated tasks are not
allowed to run when their document isn't fully active (and destroyed
documents never become fully active again). This caused everything
captured by task callbacks to leak.
We now treat tasks for destroyed documents as runnable immediately,
which gets them out of the queue.
This fixes another massive GC leak on Speedometer.
Before this change, Agent held on to all of the live MutationObserver
objects via GC::Root. This prevented them from ever getting
garbage-collected.
Instead of roots, we now use a simple IntrusiveList and remove them
from it in the finalizer for MutationObserver.
This fixes a massive GC leak on Speedometer.
This is not really a context, but more of a set of parameters for
creating a Parser. So, treat it as such: Rename it to ParsingParams,
and store its values and methods directly in the Parser instead of
keeping the ParsingContext around.
This has a nice side-effect of not including DOM/Document.h everywhere
that needs a Parser.
We've historically asserted that no "saturated" size values end up as
final metrics for boxes in layout. This always had a chance of producing
false positives, since you can trivially create extremely large boxes
with CSS.
The reason we had those assertions was to catch bugs in our own engine
code where we'd incorrectly end up with non-finite values in layout
algorithms. At this point, we've found and fixed all known bugs of that
nature, and what remains are a bunch of false positives on pages that
create very large scrollable areas, iframes etc.
So, let's change it! We now clamp content width and height of boxes to
17895700 pixels, apparently the same cap as Firefox uses.
There's also the issue of calc() being able to produce non-finite
values. Note that we don't clamp the result of calc() directly, but
instead just clamp values when assigning them to content sizes.
Fixes#645.
Fixes#1236.
Fixes#1249.
Fixes#1908.
Fixes#3057.
While keyword_to_foo() does return Optional<Foo>, in practice the
invalid keywords get rejected at parse-time, so we don't have to worry
about them here. This simplifies the user code quite a bit.
Before this change, `m_needs_repaint` was reset in
`Document::record_display_list()` only when the cached display list was
absent. This meant that if the last triggered repaint used the cached
display list, we would keep repainting indefinitely until the display
list was invalidated (We schedule a task that checks if repainting is
required 60/s).
This change also moves `m_needs_repaint` from Document to
TraversableNavigable as we only ever need to repaint a document that
belongs to traversable.
We just need to create the backend context once and let Skia handle the
context's state.
On my machine, this reduces the load time for https://tweakers.net from
7.5s to 3.5s.
Using https://github.com/whatwg/html/pull/9457
(with some changes made to catch up with the current spec)
to fix a spec bug and a crash when removing a visible popover.
This extends the optimization introduced in the previous commit to
also apply to the inserted steps for an option element. This makes
the test:
https://wpt.live/html/select/options-length-too-large.html
Go from not ever completing due to how slow it was to running, to
finishing in 800ms on my PC :^)
We can definitely expand on this a bunch more, but using the metadata
provided in the children change notification we are able to skip
runnning the expensive selectedness algorithm on the <select> element.
This removes children changed from appearing in the profile of:
https://wpt.live/html/select/options-length-too-large.html
Currently, this metadata is only provided on the insertion steps,
though I believe it would be useful to extend to the other cases
as well. This metadata can aid in making optimizations for these
steps by providing extra context into the type of change which
was made on the child.
This is still called _way_ too often, and we need to be much smarter
about when this needs to run. But we get two wins from this very
naive implementation:
1. There is a inner text setter nested within the selectedness
steps uses this list of elements. This saves us building
up a list of elements again within these steps.
2. More importantly, it caches the number of selected elements.
This already allows us to skip a minor amount of work iterating
over the children again. But in future commits, this will serve
as one of the criteria for skipping running the selectedness
algorithm altogether for certain cases, which is a very big win.
A example future idea might be to append to this list directly when
something like appendChild is run instead of iterating over all of
the children again. But that's left as future work.
The spec only refers to this property as something that is on a
Window object, and as far as I can tell this API is only ever
exposed on a Window object anyhow.
This is fundametally broken. A microtask only finishes after all
javascript has finished running. The selectedness of a select element
is observable by javascript, so any changes which are made as a result
of children changing associated with a select element should be made
through the children update steps and friends.
To be properly compatible with calc(), the resolved() methods all need:
- A length resolution context
- To return an Optional, as the calculation might not be resolvable
A bonus of this is that we can get rid of the overloads of `resolved()`
as they now all behave the same way.
A downside is a scattering of `value_or()` wherever these are used. It
might be the case that all unresolvable calculations have been rejected
before this point, but I'm not confident, and so I'll leave it like
this for now.
The focus chain always consists of newly created GC::Root objects, so
the condition always produced `false`. The fix is to use GC::Root's
overloaded operator== method, which compares the pointers of the stored
type.
This fixes Figma dropdowns and context menus instantly disappearing
upon opening them. This is because they become focused when they insert
them. Because of this bug, it would fire blur events all the way up to
and including the window. Figma listens for the blur event on the
window, and when received, it will instantly hide dropdowns and context
menus. The intention behind this seems to be hiding them when the user
clicks off the browser window, or switches tab.
Corresponds to https://github.com/whatwg/html/pull/10904
However, we don't implement most of what is changed in that PR, so this
simply moves the FIXME'd getContextAttributes() method from one place
to another.
Previously, we were generating timestamps relative to the current time
of the monotonic clock. We now generate timestamps relative to the
event loop's last render opportunity time, per the spec.
It is not cheap to do this, so only do it once within this function.
There is definitely some caching that we can do here, but this will
require some smart invalidation to detect _relevant_ changes in
the children.
Using the raw value meant that 1em would be incorrectly treated as 1px,
for example.
I've updated our canvas-filters test to demonstrate this - without the
code change this would instead have an x-offset of 2px.
We achieve this by keeping track of the number of HTMLSlotElements
inside each ShadowRoot (do via ad-hoc insertion and removal steps.)
This allows slottables assignment to skip over entire shadow roots when
we know they have no slots anyway.
Massive speedup on https://wpt.fyi/ which no longer takes minutes/hours
to load, but instead a "mere" 19 seconds. :^)
Our layout tree requires that all containers either have inline or
non-inline children. In order to support the layout of non-inline
elements inside inline elements, we need to do a bit of tree
restructuring. It effectively simulates temporarily closing all inline
nodes, appending the block element, and resumes appending to the last
open inline node.
The acid1.txt expectation needed to be updated to reflect the fact that
we now hoist its <p> elements out of the inline <form> they were in.
Visually, the before and after situations for acid1.html are identical.
Instead of just putting in members directly, wrap them up in structs
which represent what a URL blob entry is meant to hold per the spec.
This makes more obvious what this is meant to represent, such as the
ByteBuffer being used to represent the bytes behind a Blob.
This also allows us to use a stronger type for a function that needs
to return a Blob URL entry's object.
Interestingly, the spec has a note saying:
> window.postMessage() performs StructuredSerializeWithTransfer on
> its arguments, but is careful to do so immediately, inside the
> synchronous portion of its algorithm. Thus it is able to use the
> algorithms without needing to prepare to run script and prepare
> to run a callback.
But there is no note about the deserialization steps. In any case, we do
need callbacks enabled here.
...Which doesn't do anything given start() itself doesn't do anything,
but this is a subtle enough point of the spec that it seems worthwhile
to implement now for whenever this does become meaningful.
For example, running `alert(1)` will pause the event loop, during which
time no JavaScript should execute. This patch extends this disruption to
microtasks. This avoids a crash inside the microtask executor, which
asserts the JS execution context stack is empty.
This makes us behave the same as Firefox in the following page:
<script>
queueMicrotask(() => {
console.log("inside microtask");
});
alert("hi");
</script>
Before the aforementioned assertion was added, we would execute that
microtask before showing the alert. Firefox does not do this, and now
we don't either.
The use of this HashMap looks very spooky, but let's at least use
finalize when cleaning them up on destruction to make things slightly
less dangerous looking.
This commit begins to implement the track processing model. When the
`src` attribute is updated, we now fetch the given source file.
Currently, we always fire an `error` event once fetching is completed,
as we don't support processing the fetched data.
Instead of always reporting a colno and lineno of zero try and use the
values from the Error object that may be provided, falling back to the
source location of the invocation if not provided. We can definitely
improve the reporting even more, but this is a start!
Also update this function to latest spec while we're in the area.
This isn't a full fix, as the paint function does not handle this
either. But instead of getting the bitmap from the image source
immediately, follow the spec a bit more closely by creating the
CanvasPatern object with the ImageSource directly.
Fixes a crash for the 5 included WPT tests.
We added these methods to propagate OOM errors at process startup, but
we longer fret about these tiny OOM failures. Requiring that these init
methods be called prohibits using these strings in processes that have
not set up a MainThreadVM. So let's just remove them and initialize the
strings in a sane manner.
In doing so, this also standardizes how we initialize strings whose C++
variable name differs from their string value. Instead of special-casing
these strings, we just include their string value in the x-macro list.
This makes it more convenient to use the 'relvant agent' concept,
instead of the awkward dynamic casts we needed to do for every call
site.
mutation_observers is also changed to hold a GC::Root instead of raw
GC::Ptr. Somehow this was not causing problems before, but trips up CI
after these changes.
URL::basic_parse has a subtle bug where the resulting URL is not set
to valid when StateOveride is provided and the URL parser early returns
a valid URL.
This has not surfaced as a problem so far, as the only users of the
state override API provide an already valid URL buffer and also ignore
the result of basic parsing with a state override.
However, this bug surfaces implementing the URL pattern spec, which as
part of URL canonicalization:
* Provides a dummy URL record
* Basic URL parses that URL with state override
* Checks the result of the URL parser to validate the URL
While we could set URL validity on every early return of the URL parser
during state override, it has been a long standing FIXME around the code
to try and remove the awkward validity state of the URL class. So this
commit makes the first stage of this change by migrating the basic
parser API to return Optional, which also happens to make this subtle
issue not a problem any more.
This matches the behavior of other browsers, which always set the dirty
checkedness flag when setting checkedness, except when setting the
`checked` content attribute.
Previously, <a> elements were frequently invalidated because
`set_the_url()` was called by `reinitialize_url()`, which is a
preparation step in every HTMLHyperlinkElementUtils function. As a
result, styles were unnecessarily invalidated each time any of these
functions were invoked without changing the URL.
Previously, if the NumericCharacterReferenceEnd state was reached when
current_input_character was None, then the
DONT_CONSUME_NEXT_INPUT_CHARACTER macro would restore back before the
EOF, and allow the next state (after the SWITCH_TO_RETURN_STATE) to
proceed with the last digit of the numeric character reference.
For example, with something like `ї`, before this commit the
output would incorrectly be `<code point with the value 1111>1` instead
of just `<code point with the value 1111>`.
Instead of putting the `if (current_input_character.has_value())` check
inside NumericCharacterReferenceEnd directly, it was instead added to
DONT_CONSUME_NEXT_INPUT_CHARACTER, because all usages of the macro
benefit from this check, even if the other existing usage sites don't
exhibit any bugs without it:
- In MarkupDeclarationOpen, if the current_input_character is EOF, then
the previous character is always `!`, so restoring and then checking
forward for strings like `--`, `DOCTYPE`, etc won't match and the
BogusComment state will run one extra time (once for `!` and once
for EOF) with no practical consequences. With the `has_value()` check,
BogusComment will only run once with EOF.
- In AfterDOCTYPEName, ConsumeNextResult::RanOutOfCharacters can only
occur when stopping at the insertion point, and because of how
the code is structured, it is guaranteed that current_input_character
is either `P` or `S`, so the `has_value()` check is irrelevant.
When setting the textContent of an element with no children to null or
the empty string, nothing happens. Even so, we were still invalidating
style, layout and collections, causing pointless churn.
Skipping invalidation in this case also revealed that we were missing
invalidation when changing the selected state of HTMLOptionElement.
This was all caught by existing tests already in-tree. :^)
Instead of storing all storage objects in static memory, we now
follow the the spec by lazily creating a unique Storage object
on each document object.
Each Storage object now holds a 'proxy' to the underlying backing
storage. For now, this proxy is simply a reference to the backing
object. In the future, it will need to be some type of interface
object that stores on a SQLite database or similar.
Session storage is now correctly stored / tracked as part of the
TraversableNavigable object.
Local storage is still stored in a static map, but eventually this
should be factored into something that is stored at the user agent
level.
This is consistent with other functions such as
HTMLElement::offset_width and fixes a crash for the included test.
Returning an offset of zero is not correct for this case, but this is
still an improvement to not crash.
Previously, the`HTMLInputElement.selectinStart` and
`HTMLInputElement.selectionEnd` IDL setters, and the
`setRangeText()` IDL method were used when updating an input's value
on keyboard input. These methods can't be used for this purpose,
since selection doesn't apply to email type inputs. Therefore, this
change introduces internal-use only methods that don't check whether
selection applies to the given input.
This change updates the bindings generator for the case defined at
https://html.spec.whatwg.org/#reflecting-content-attributes-in-idl-attributes:element;
that is, the case “If a reflected IDL attribute has the type T?, where T
is either Element or an interface that inherits from Element”.
The change “normalizes” the generator behavior for that case — such that
the generated code expects a getter with a name of the form used in
other cases; e.g., popover_target_element().
Otherwise, without this change, the generator expects a name of the form
get_popover_target_element() for that case.
1. Stop using GC::Root in member variables, since that usually creates
a realm leak.
2. Stop putting OrderedHashMap<FlyString, GC::Ptr> on the stack while
setting these up, since that won't protect the objects from GC.
Many times, attribute mutation doesn't necessitate a full style
invalidation on the element. However, the conditions are pretty
elaborate, so this first version has a lot of false positives.
We only need to invalidate style when any of these things apply:
1. The change may affect the match state of a selector somewhere.
2. The change may affect presentational hints applied to the element.
For (1) in this first version, we have a fixed list of attribute names
that may affect selectors. We also collect all names referenced by
attribute selectors anywhere in the document.
For (2), we add a new Element::is_presentational_hint() virtual that
tells us whether a given attribute name is a presentational hint.
This drastically reduces style work on many websites. As an example,
https://cnn.com/ is once again browseable.
The previous VERIFY statement incorrectly asserted that the
interception state was not "committed" or "scrolled". Updated
the condition to ensure the interception state is either
"committed" or "scrolled" as intended.
Before this change, StyleComputer would essentially take a DOM element,
find all the CSS rules that apply to it, and resolve the computed value
for each CSS property for that element.
This worked great, but it meant we had to do all the work of selector
matching and cascading every time.
To enable new optimizations, this change introduces a break in the
middle of this process where we've produced a "CascadedProperties".
This object contains the result of the cascade, before we've begun
turning cascaded values into computed values.
The cascaded properties are now stored with each element, which will
later allow us to do partial updates without re-running the full
StyleComputer machine. This will be particularly valuable for
re-implementing CSS inheritance, which is extremely heavy today.
Note that CSS animations and CSS transitions operate entirely on the
computed values, even though the cascade order would have you believe
they happen earlier. I'm not confident we have the right architecture
for this, but that's a separate issue.
Auto popovers now correctly establish a close watcher when shown.
This means popovers now correctly close with an escape key press.
Also correctly hide open popovers when removed from the document.
An AK::String works fine for a USVString as a USVString is just a more
strict version of DOMString. Maybe we will have a different String type
for it in the future, but for now using an AK::String is fine and we do
not need this FIXME.
In conformance with the requirements of the spec PR at
https://github.com/whatwg/html/pull/9546, this change adds support for
the “switch” attribute for type=checkbox “input” elements — which is
shipping in Safari (since Safari 17.4). This change also implements
support for exposing it to AT users with role=switch.
These were converted to lambdas in
6b921e91d4ce4791c67bf467a2ba519b8e3ca88b
But I merged fcf6cc27f2 without checking
that the code had responded to the change.
The popoverTargetElement seems to be one of the only cases of a
reflected Element? attribute in the HTML spec, the behaviour of which
is specified in section 2.6.1.
Buttons can't actually toggle popovers yet because showing/hiding
popovers is not implemented yet.
Additionally: For “img” elements with empty “alt” attributes, change the
default role to the newer, preferred “none” synonym for the older
“presentation” role; import https://wpt.fyi/results/html-aam/roles.html
(which provides test/regression coverage for these changes).
Currently, the following JS snippet will hang indefinitely:
new DOMParser().parseFromString("<object>", "text/html");
Because the document into which the object is inserted is not active. So
the task queued to run the representation steps will never run.
This patch implements the spec steps to rerun the representation steps
when the active state changes, and avoid the hang when the object is
created in an inactive document.
This change aligns the default roles for “th” and “td” elements with the
requirements in the HTML-AAM spec, and with the corresponding WPT tests
at https://wpt.fyi/results/html-aam/table-roles.html, and with the
behavior in other engines.
Otherwise, without this change, the default role values for “th” and
“td” elements in some cases don’t match the behavior in other engines,
and don’t match the expected results for the corresponding WPT tests.
This change makes Ladybird conform to the current requirements at
https://w3c.github.io/core-aam/#roleMappingComputedRole in the “Core
Accessibility API Mappings” spec for the case of “orphaned” li elements;
that is, any li element which doesn’t have a role=list ancestor.
The core-aam spec requires that in such cases, the li element must not
be assigned the “listitem” role but instead must be treated as if it had
no role at all.
This eliminates the use of ResourceLoader in HTMLObjectElement. The spec
steps around fetching have been slightly updated since we've last looked
at this, so those are updated here.
Regarding the text test change: we cannot rely on the data: URL being
fetched synchronously. It will occur on a deferred task now. This does
match the behavior of other browsers, as they also will not have run the
fallback representation steps as of DOMContentLoaded.
There are essentially 3 URL parsing AOs defined by the spec:
1. Parse a URL
2. Encoding parse a URL
3. Encoding parse a URL and serialize the result
Further, these are replicated between the Document and the ESO.
This patch defines these methods in accordance with the spec and updates
existing users to invoke the correct method. In places where the correct
method is ambiguous, we use the encoding parser to preserve existing ad-
hoc behavior.
The DOM spec defines what it means for an element to be an "editing
host", and the Editing spec does the same for the "editable" concept.
Replace our `Node::is_editable()` implementation with these
spec-compliant algorithms.
An editing host is an element that has the properties to make its
contents effectively editable. Editable elements are descendants of an
editing host. Concepts like the inheritable contenteditable attribute
are propagated through the editable algorithm.
Using a default reference capture for these kinds of tasks is dangerous
and prone to error. Some of the variables should for sure be captured
by value so that we can keep a GC object alive rather than trying to
refer to stack objects.
These variables are all captured in queued events or other event loop
tasks, but are all guarded by event loop spins later in the function.
The IGNORE_USE_IN_ESCAPING_LAMBDA will soon be required for all locals
that are captured by ref in GC::Function as well as AK::Function.
Add support for shared memory creation in WebAssembly memory API.
This API is needed for WPT tests that use shared array buffers.
Import related WPT tests.
This change implements spec-conformant computation of default ARIA roles
for elements whose expected default role depends on the element’s
context — specifically, either on the element’s ancestry, or on whether
the element has an accessible name, or both. This affects the “aside”,
“footer”, “header”, and “section” elements.
Otherwise, without this change, “aside”, “footer”, “header”, and
“section” elements may unexpectedly end up with the wrong default roles.
Details' contents matches a new details-content pseudo element.
Further work is required to make this pseudo-element behave per spec.
This pseudo should be element-backed per
https://drafts.csswg.org/css-pseudo/#element-backed
Introduced in bd932858, the code would try to move `result.color_space`
into the ImmutableBitmap created for each single frame of an image. For
animated images, this would result in a use-after-move from the second
frame.
This corresponds to https://github.com/whatwg/html/pull/10753
WHile I was at it, I've also moved the checks inside the spin callback,
and reformatted the spec comments to match our style.
This patch introduces the `Gfx::ColorSpace` class, this is basically a
serializable wrapper for skia's SkColorSpace. Creation of the instances
of this class (and thus ICC profiles parsing) is performed in the
ImageDecoder process. Then the object is serialized and sent through
IPC, to finally be handed to skia for rendering.
However, to make sure that we're not making all LibGfx's users dependent
on Skia as well, we need to ensure the `Gfx::ColorSpace` object has no
dependency on objects from Skia. To that end, the only member of the
`ColorSpace` class is the opaque `ColorSpaceImpl` struct. Though, there
is on issue with that design, the code in `DisplayListPlayer.cpp` needs
access to the underlying `sk_sp<SkColorSpace>`. It is provided by a
template function, that is only specialized for this type.
Doing this work allows us to pass the following WPT tests:
- https://wpt.live/css/css-color/tagged-images-001.html
- https://wpt.live/css/css-color/tagged-images-003.html
- https://wpt.live/css/css-color/tagged-images-004.html
- https://wpt.live/css/css-color/untagged-images-001.html
Other test cases can also be found here:
- https://github.com/svgeesus/PNG-ICC-tests
Note that SkColorSpace support quite a limited amount of color spaces,
so color profiles like the ones in [1] or the v4 profiles in [2] are not
supported yet. In fact, SkColorSpace only accepts skcms_ICCProfile with
a linear conversion to XYZ D50.
[1] https://www.color.org/browsertest.xalter
[2] https://www.color.org/version4html.xalter
