Fix absolute descendents being replaced when retreiving root
<!-- Please describe your changes on the following line: -->
The cause of #16410 is because calling `try_get_layout_root` will replace its `abs_descendants` even itself isn't absolute positioned. So I pushed them instead if the root isn't absolute. Also add a few docs and comments to help understand how block size of absolute flow is being calculated.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#16410 (GitHub issue number if applicable)
<!-- Either: -->
- [x] There are tests for these changes:
Reduced testcase from the issue.
```html
<style>
html {
min-height: 100%;
position: relative;
}
div {
position: absolute;
bottom: 0;
width: 100%;
height: 25px;
border: 1px solid black;
}
</style>
<div>
</div>
```
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Allow displaying content with "application/json" mime type
For me this allows the WPT test
performance-timeline/tentative/include-frames-one-remote-child.sub.html
to match expected results. The wptserver is sending a 404 JSON response
because the URL that the test requests is not found.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#29136.
- [x] There are tests for these changes.
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Stringify unknown JavaScript objects in global exception handlers
When turning DOM exceptions into `ErrorInfo` always try to stringify
the JavaScript value, even if it's an object that isn't a `DOMException`
or native exception. This means that exceptions that extend the `Error`
prototype are now stringified. The result is that test output for WPT
global assertion failures is more useful. For instance for the test
include-frames-from-child-same-origin-grandchild.sub.html:
Before:
```
uncaught exception: unknown (can't convert to string)
```
After:
```
uncaught exception: Error: assert_equals: expected 4 but got 3
```
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] There are tests for these changes.
For me this allows the WPT test
performance-timeline/tentative/include-frames-one-remote-child.sub.html
to match expected results. The wptserver is sending a 404 JSON response
because the URL that the test requests is not found.
Combine `run.py` and `update.py` into `servowpt.py` in order to allow
them to share code. Import them directly into the mach script to avoid
having to call `compile` and `exec` on the code. This makes it clearer
how they are executed. In addition, move all of the setup into
`setupwpt.py` to avoid differences between tests executed via mach and
not. Finally, be more ambitious when detecting the build to use. If none
was specified, try to use the one that exists between "release" and
"debug."
When turning DOM exceptions into `ErrorInfo` always try to stringify
the JavaScript value, even if it's an object that isn't a `DOMException`
or native exception. This means that exceptions that extend the `Error`
prototype are now stringified. The result is that test output for WPT
global assertion failures is more useful. For instance for the test
include-frames-from-child-same-origin-grandchild.sub.html:
Before:
```
uncaught exception: unknown (can't convert to string)
```
After:
```
uncaught exception: Error: assert_equals: expected 4 but got 3
```
wpt: update README instructions about running tests manually
I wanted to run a single test many times in parallel to smoke out some timeouts, but this was surprisingly tricky. This involved running multiple instances of `mach test-wpt`, which in turn involved convincing wptrunner not to start a WPT server for each instance, which is not currently exposed as an option. I also eventually realised I needed to install a custom testharnessreport.js, but the instructions were very outdated, so most of the paths and commands etc had changed.
This patch adds a section about running `mach test-wpt` with an external WPT server, and updates the section about running tests manually.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [ ] ~~`./mach build -d` does not report any errors~~
- [ ] ~~`./mach test-tidy` does not report any errors~~
- [ ] These changes fix #___ (GitHub issue number if applicable)
<!-- Either: -->
- [ ] There are tests for these changes OR
- [x] These changes do not require tests because they affect the wpt readme only
Wait until a poster image is cached to in order to unblock document load. If
not, the document may continue loading before the image is ready to use,
leading to intermittency in test output. Now load is unblocked when
getting the ImageResponse from the cache, which allows the code to
properly unblock the load when the entire load fails or succeeds.
This reveals several false passes in the `object-view-box` test suite
which were very flaky.
Fixes#29204.
Fixes#29188.
Fixes#29179.
Signed-off-by: Martin Robinson <mrobinson@igalia.com>
