This PR removes `jsstring_to_str`, which is replaced with
`jsstr_to_string`, and updates `mozjs` to
6f3dcb99a7.
Given that servo now always replaces unpaired surrogate since
https://github.com/servo/servo/pull/35381, the internal conversion
function `jsstring_to_str` is functionally the same as `jsstr_to_string`
from `mozjs`. This PR removes `jsstring_to_str` and replaces with
`jsstr_to_string` with conversions to `DOMString` where necessary.
Testing: Passes all unit test. No regression was found in WPT test (see
try run: https://github.com/minghuaw/servo/actions/runs/16821156583)
---------
Signed-off-by: minghuaw <wuminghua7@huawei.com>
Signed-off-by: sagudev <16504129+sagudev@users.noreply.github.com>
Co-authored-by: sagudev <16504129+sagudev@users.noreply.github.com>
to debug workers in a page with the [SpiderMonkey Debugger
API](https://firefox-source-docs.mozilla.org/js/Debugger/), we need to
pass the worker’s global object to
[Debugger.prototype.**addDebuggee()**](https://firefox-source-docs.mozilla.org/js/Debugger/Debugger.html#adddebuggee-global).
we could pick up the global via the [onNewGlobalObject()
hook](https://firefox-source-docs.mozilla.org/js/Debugger/Debugger.html#onnewglobalobject-global),
but if our script system passes the global to the debugger script
instead, we can later add details like the PipelineId that will help
servo identify debuggees that the API is notifying us about (#38334).
this patch creates a debugger global in worker threads, runs the
debugger script in those new globals, and plumbs new worker globals from
those threads into addDebuggee() via the debugger script. since worker
threads can’t generate PipelineId values, but they only ever run workers
on behalf of one pipeline, we use that pipeline’s PipelineId as the
PipelineId of the debugger global, rather than generating a unique
PipelineId like we do in script threads.
Testing: will undergo many automated tests in #38334
Fixes: part of #36027
Signed-off-by: Delan Azabani <dazabani@igalia.com>
Co-authored-by: atbrakhi <atbrakhi@igalia.com>
- [Dispatch Event](https://dom.spec.whatwg.org/#concept-event-dispatch)
should return a Boolean. This function is used frequently in spec and
the change makes things easier to follow.
- Remove `enum EventStatus` and related functions.
- Update some dead spec link.
- Update some steps.
This is intended as cleanup before working on #38435 and reduces binary
size by 488KB in Release profile.
Testing: No behaviour change.
---------
Signed-off-by: Euclid Ye <euclid.ye@huawei.com>
This is a follow-up to #38532 which simply defaulted to an empty string
as the spec's WebIDL doesn't mark `innerHTML` as `Throws`. However,
since the absence of `Throws` in the spec doesn't imply no-throw, we can
mark this as fallible in our WebIDL. This makes the `ShadowRoot`'s
implementation match `Element`'s `innerHTML`.
Testing: Same as #38532.
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
When clearing input file with WebDriver, we should set input's filelist
with empty list instead of setting it with None.
Testing:
`./tests/wpt/tests/webdriver/tests/classic/element_send_keys/file_upload.py`
Signed-off-by: PotatoCP <Kenzie.Raditya.Tirtarahardja@huawei.com>
Clippy now checks to see if unsafe code is wrapped in unsafe blocks. We
have this lint disabled for `script` and `script_bindings` because of a
lot of legacy code that doesn't do this. The lint is useful though as it
makes it more obvious what code is unsafe. This is an incremental step
toward being able to turn this lint on for `script`.
This has the benefit of silencing warnings that show up in some IDEs
that use rust-analyzer.
Testing: This should not change behavior at all and is thus covered by
existing tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
We were instead stretching to the containing block, which implied that
the behaviors of a `stretch` size and `stretch` alignment weren't
consistent.
As resolved by the CSSWG, the behavior will now be:
- If the cross size of the line is known, stretch to the line.
- Otherwise, stretch to the containing block.
See https://github.com/w3c/csswg-drafts/issues/11784
This aligns us with Blink, which has already shipped this new behavior.
Testing: Improves existing WPT and adds a new test.
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
#37686 fixed an entire class of devtools protocol bugs that led to
protocol desyncs with warnings in the Servo log, but it also removed the
warnings, making it harder to spot where our devtools impl is
incomplete.
this patch reintroduces a warning whenever some Actor::handle_message()
returns Err(ActorError).
Testing: debug logging only, so not really worth testing
---------
Signed-off-by: Delan Azabani <dazabani@igalia.com>
Co-authored-by: atbrakhi <atbrakhi@igalia.com>
Co-authored-by: Martin Robinson <mrobinson@igalia.com>
This reverts commit dcb90bb85e.
This broke scrollable overflow calculation in the following case:
```
<div id="foo" style="width: 200px; height: 200px; outline: solid; overflow: auto;">
<div style="height: 5000px; background: pink;">hello</div>
</div>
```
In this case the overflow is propagating through the `overflow: auto`
`<div>` and into the parents. When dumping the flow tree I see the root
node being 5000 pixels tall. It's unclear why this change didn't break
any tests, so it's likely that we need to add a test for this case.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This is useful to better isolate `unsafe` code. Once all unsafe calls
are wrapped we can enable the Rust warning. This also explicitly
disables the warning for generated code, which is a much more difficult
task. After this change there are 211 warnings left in
`script_bindings`.
Testing: This should not change behavior and is thus covered by existing
tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
in the devtools protocol, [source
forms](https://firefox-source-docs.mozilla.org/devtools/backend/protocol.html#loading-script-sources)
announced in `resources-available-array` messages can include the
`introductionType`, which more or less mirrors the field of the same
name in SpiderMonkey’s CompileOptions.
this patch exposes `introductionType` accordingly, allowing us to check
for the correct values in automated tests.
Testing: new coverage in devtools tests
Fixes: part of #36027
---------
Signed-off-by: Delan Azabani <dazabani@igalia.com>
Co-authored-by: atbrakhi <atbrakhi@igalia.com>
Testing: [Try run][1] did not reveal any test failures. There doesn't
seem to be any straightforward failure scenarios that can be triggered
in `xml5ever` that are not IO errors and the xml_serialize method simply
serializes to a String buffer.
[1]:
https://github.com/servo/servo/actions/runs/16824267959/job/47657275606l
Signed-off-by: Mukilan Thiyagarajan <mukilan@igalia.com>
As titled. Initially I want to include this as part of another PR but
looks so messy..
The motivation is that we are no longer just waiting for script
response.
Testing: No behaviour change.
Signed-off-by: Euclid Ye <euclid.ye@huawei.com>
Instead of using WebRender hit testing to update the cursor, base it on
layout hit tests. This allows removing the majority of WebRender hit
test items and finally opens up the possibility of adding support for
custom cursors. In addition, this change fixes an issue where cursors
were not set properly on areas of the viewport that extended past the
page content.
Testing: This is difficult to test as verifying that the cursor changed
properly is beyond the capabilities of Servo's test harnesses.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
When stretching the cross size of a flex item to its flex line, we were
computing the stretch size by subtracting padding, border and margin
from the line size. However, this could result in a negative amount for
the content-box cross size. Therefore, this floors it by zero.
Testing: Adding new tests
Fixes: #38517
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
This performs various refactorings:
- Turns `content_size` into `content_block_size`, losing the inline
component, since it was the same as `containing_block_inline_size`.
- Merges `containing_block_inline_size` and
`containing_block_block_size` into `containing_block_size`.
- Removes `has_child_which_depends_on_block_constraints` since this
information should already be in `depends_on_block_constraints`.
- `FlexItem::layout()` is no longer responsible for trying to reuse the
previous result. Therefore it no longer returns an `Option`, and no
longer accepts the previous result as a parameter.
- `FlexItemLayoutResult::compatible_with_containing_block_size()` is
removed, and a simplified version is inlined into the relevant caller of
`FlexItem::layout()`.
Testing: Not needed (no change in behavior)
Signed-off-by: Oriol Brufau <obrufau@igalia.com>
Incorporates the updates from https://github.com/servo/mozjs/pull/584.
Testing: Existing WPT coverage is enough.
Fixes: Part of #36258
---------
Signed-off-by: Josh Matthews <josh@joshmatthews.net>
### Changes made
This implements named grid lines (line names in `grid-template-*`),
named grid areas (`grid-template-areas`), and the ability to target
those using `grid-{row,column}-{start,end}`. It also includes a bunch of
miscelaneous fixes for `repeat(auto-fill | auto-fit, ...)` syntax as
that interacts with the specification of line names.
The actual layout implementation is in Taffy. The bulk of this PR is
updating Servo to translate (CSS Grid-related) Stylo types into Taffy
types using a new iterator-based API which uses iterators and lazy
translation for efficiency (which is more important now that we're
dealing with string data, even though they're `Atom`s).
### Testing
This functionality has lots of WPT tests. It fixes some seemingly random
CSS Grid tests that use named lines/areas even though that's not what
they're testing.
### Screenshots
wikipedia.org
<img width="1624" height="1056" alt="Screenshot 2025-07-27 at 20 03 16"
src="https://github.com/user-attachments/assets/2c50b96f-ae36-4405-ac48-b771bfdcb515"
/>
bbc.co.uk:
<img width="1624" height="1056" alt="Screenshot 2025-07-27 at 20 32 57"
src="https://github.com/user-attachments/assets/ba84e211-65d2-4411-95fb-7b9b91bea31c"
/>
theguardian.com:
<img width="1624" height="1056" alt="Screenshot 2025-07-27 at 20 33 29"
src="https://github.com/user-attachments/assets/e85daaa6-5fb0-45d4-b9ec-b22b38b087ec"
/>
---------
Signed-off-by: Nico Burns <nico@nicoburns.com>
The fetch thread is currently not shut down, meaning it is one of the
threads counted in the "threads are still running after shutdown (bad)"
counter shown when Servo has shut-down.
This PR introduces a mechanism to start, and shut-down, one fetch thread
per process that requires one.
Testing: WPT tests and unit tests(for font context). Also manually
tested loading and closing "about:blank": this change indeed brings down
the count of threads still running after shutdown by one.
Fixes: https://github.com/servo/servo/issues/34887
---------
Signed-off-by: gterzian <2792687+gterzian@users.noreply.github.com>
Before, the compositor was responsible for doing the hit testing during
input events within a page. This change moves that hit testing to
layout. With this change, epoch mismatches are no longer a bit deal and
we can simply ignore them, as the Constellation and Script will take
care of ignoring hit tests against scroll nodes and browsing contexts
that no longer exist. This means that hit testing retry support can be
removed.
Add the concept of a Script `HitTest` that transforms the coarse-grained
renderer hit test into one that hit tests against the actual layout
items.
Testing: Currently we do not have good tests for verifying the behavior
of
input events, but WebDriver tests should cover this.
Fixes: This is part of #37932.
Fixes: #26608.
Fixes: #25282.
Fixes: #38090.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Co-authored-by: kongbai1996 <1782765876@qq.com>
When restoring context/state we need to pop all clips from current
state, before we just poped one (even if there was none).
Testing: Added new WPT tests
---------
Signed-off-by: sagudev <16504129+sagudev@users.noreply.github.com>
We currently send exit signals to script-threads, and we also join on
the BHM worker. This ensure the constellation shuts-down only after
script has dropped it's sender to the BHM worker. By joining on the
script-threads, we have a guarantee that they have exited(which is
stronger than having dropped their senders) by the time the
constellation exits.
Testing: Manually opened many tabs and closed the window, both in
single- and multi-process modes.
Fixes: Part of - https://github.com/servo/servo/issues/30849
---------
Signed-off-by: gterzian <2792687+gterzian@users.noreply.github.com>
This makes grid layout make use of Servo's layout caching system.
Testing: No behavior change. Existing WPT coverage
---------
Signed-off-by: Nico Burns <nico@nicoburns.com>
IntersectionObserver needs to be able to query node geometry without
forcing a layout. A previous layout could have run without needing a
`StackingContextTree`. In that case the layout-less query should finish
building the `StackingContextTree` before doing the query. Add a new
type of layout API which requests that layout finishes building the
StackingContextTree.
This change also slightly simplifies and corrects the naming of
`Element` APIs around client box queries.
Testing: This should fix intermittent failures in WPT tests.
Fixes: #38380.
Fixes: #38390.
Closes: #38400.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
This is an implementation for `scrollIntoView`. For now, it is called
when a certain element gains focus.
Testing: Existing WPT tests
Fixes: #24059
Signed-off-by: abdelrahman1234567 <abdelrahman.hossameldin.awadalla@huawei.com>
Include the scrollable overflow of a child box if either its parent or
child has `overflow: visible`
**Issue**: For the blocks having property `overflow:hidden`, their
scroll overflow is not added to parent's scroll overflow.
Causing unable to scroll the parent block aka `Root` block in our Issue
#38248 .
**Testing**: css/cssom-view/scrolling-quirks-vs-nonquirks.html
**Fixes**: #38248
Signed-off-by: Shubham Gupta <shubham13297@gmail.com>
*Describe the changes that this pull request makes here. This will be
the commit message.*
Testing: *Describe how this pull request is tested or why it doesn't
require tests*
Fixes: *Link to an issue this pull requests fixes or remove this line if
there is no issue*
---------
Signed-off-by: Jerens Lensun <jerensslensun@gmail.com>
Add step 1 and simplify step 2 - 5 to remove duplicate `GetClientRects`
call.
There is some issue with spec to be updated later.
Testing: No behaviour change.
---------
Signed-off-by: Euclid Ye <euclid.ye@huawei.com>
Using `_` on such long bit patterns makes them easier to read.
Testing: Just formatting.
Fixes: #38474
Signed-off-by: sagudev <16504129+sagudev@users.noreply.github.com>
to use the [SpiderMonkey Debugger
API](https://firefox-source-docs.mozilla.org/js/Debugger/) as the single
source of truth about scripts and their sources for devtools purposes
(servo/servo#38334), the debugger script needs to be able to distinguish
inline scripts from other scripts, because inline scripts are a special
case where the source contents need to come from the Servo parser.
the mechanism for this is
[Debugger.Script.prototype.**introductionType**](https://firefox-source-docs.mozilla.org/js/Debugger/Debugger.Source.html#introductiontype),
which is `inlineScript` for inline scripts or a variety of other values
for other kinds of scripts, but only the embedder can provide this
information.
this patch bumps mozjs to servo/mozjs#603, which expands on
CompileOptionsWrapper, making it a safe wrapper around CompileOptions.
to construct one from safe code, use Runtime::new_compile_options().
then you can call `set_introduction_type(&'static CStr)` on the new
instance. we also make Runtime::evaluate_script() take a
CompileOptionsWrapper from the caller, instead of constructing one
internally.
in this patch, we set the introductionType to `c"inlineScript"` when
calling run_a_classic_script() and compile_module_script() for inline
scripts, and leave it unset all other cases.
Testing: will undergo automated tests in #38334
Fixes: part of #36027, part of servo/servo#38378
---------
Signed-off-by: Delan Azabani <dazabani@igalia.com>
Co-authored-by: atbrakhi <atbrakhi@igalia.com>
Implement a simple one element cache for projected points in spatial
nodes. This should reduce the amount matrix math done during hit
testing.
Testing: This should not change test results, but should improve
performance a bit. Thus, tests are not necessary beyond existing
performance tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
In #18933, hit testing was moved from layout to WebRender. This presents
some issues. For instance, the DOM can change at the same time that hit
test is happening. This can mean that hit test returns references to
defunct DOM nodes, introducing memory safety issues. Currently, Servo
will try to ensure that the epochs used for testing and those recorded
in the DOM match, but this is not very reliable and has led to code that
retries failed hit tests.
This change reintroduces (8 years later) a layout hit tester and turns
it on for `document.elementFromPoint` and `document.elementsFromPoint`.
The idea is that this hit tester will gradually replace the majority of
the WebRender hit testing happening in the renderer.
Testing: This shouldn't really change the behavior hit testing, but it
seems to improve one WPT test.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Co-authored-by: kongbai1996 <1782765876@qq.com>
There were various booleans on `ReflowResults` that represented various
actions that might have been taken during a reflow request. Replace
those with a bitflags that better represents what reflow phases have
actually been run. Update variable names to reflect what they mean.
In addition, run some post-layout tasks unconditionally. They are
already contingent on the results returned from layout.
This simplifies and clarifies the code a good deal.
Testing: This should not change observable behavior and thus is covered
by existing WPT tests.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
to debug the scripts in a page with the [SpiderMonkey Debugger
API](https://firefox-source-docs.mozilla.org/js/Debugger/), we need to
pass the page’s global object to
[Debugger.prototype.**addDebuggee()**](https://firefox-source-docs.mozilla.org/js/Debugger/Debugger.html#adddebuggee-global).
we could pick up the global via the [onNewGlobalObject()
hook](https://firefox-source-docs.mozilla.org/js/Debugger/Debugger.html#onnewglobalobject-global),
but if our script system passes the global to the debugger script
instead, we can later add details like the PipelineId that will help
servo identify debuggees that the API is notifying us about (#38334).
this patch plumbs new Window globals from script into addDebuggee() via
the debugger script. to call into the debugger script with structured
input, we create a new DOM event type, DebuggerEvent, that the debugger
script listens for as the “addDebuggee” event.
Testing: no testable effects yet, but will be used in #37667
Fixes: part of #36027
---------
Signed-off-by: Delan Azabani <dazabani@igalia.com>
Co-authored-by: atbrakhi <atbrakhi@igalia.com>
