Automatic update from web-platform-tests[LayoutNG] Avoid false positives for clearance.
Clearance inhibits margin collapsing and thus makes it possible to
resolve the BFC offset of its parent. In that regard we sometimes
incorrectly assumed that a child would get clearance. The only case
where we can tell for sure up front, is when there are adjoining floats
that match the value of the 'clear' property. In all other cases, we
first need to calculate the block-start border edge of the child.
Added a few new tests. Only two of them failed in NG prior to this CL,
but coverage seemed low, so I added them anyway. Two of them fail in the
legacy engine.
Move the code that detects pending child floats a bit further down, to
give HandleInflow() an opportunity to re-layout the child without
aborting container layout. Otherwise clearance on an empty block will
be lost when positioning the float. Added two tests for this
(second-float-inside-empty-cleared-block*), since nothing except a
multicol test (that happened to do complicated clearance stuff)
regressed.
Removed now unused bits from NGExclusionSpace.
Cq-Include-Trybots: master.tryserver.chromium.linux:linux_layout_tests_layout_ng
Change-Id: I73b4aa02e5e88d1063f345db271cde1ad31ea9f5
Reviewed-on: https://chromium-review.googlesource.com/1049974
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#557282}
--
wpt-commits: 857126e062ef13c28981a9a1a01df4695f74715d
wpt-pr: 10897
Automatic update from web-platform-testsAdd WorkerNavigatorUSB supplement
This change adds a WorkerNavigatorUSB supplement for WorkerNavigator.
The WorkerNavigatorUSB interface contains a usb attribute that returns
a USB object. The constructor for this USB object was also changed to
use an ExecutionContext* parameter to allow a worker to properly
initialize the object.
This change also updates the global interface listing tests to reflect
the new attribute that WorkerNavigator contains and adds web platform
tests to check that the WorkerNavigator does indeed have the usb
attribute defined in the supported contexts.
Due to a bug (https://crbug.com/839117) with the WebIDL bindings
generator, the global-interface-listing-service-worker.html test
exposes the usb attribute of WorkerNavigator. Therefore, a check is
performed in WorkerNavigatorUSB to return a null pointer if the
execution context is a ServiceWorkerGlobalScope. Since the intended
behavior for navigator.usb in a service worker is to be undefined, the
web platform test for this checks that condition instead.
Bug: 837414
Change-Id: I3ca63c9ae0b85e5425134024d2fdd9b29ef7d8c3
Reviewed-on: https://chromium-review.googlesource.com/1024724
Commit-Queue: Ovidio Henriquez <odejesush@chromium.org>
Reviewed-by: Kentaro Hara <haraken@chromium.org>
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Chong Zhang <chongz@chromium.org>
Reviewed-by: Reilly Grant <reillyg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#557272}
--
wpt-commits: ede61140b37b59259a3faf48478e7badae90016e
wpt-pr: 10835
Automatic update from web-platform-tests[infra] Correct integration test
By default, "single-page tests" must include an explicit call to the
global `done` function [1]. Insert this invocation to signal completion
and avoid a test timeout when the test's expectation is satisfied.
[1] http://web-platform-tests.org/writing-tests/testharness-api.html
--
[ci] Correct syntax in build script
--
[wptrunner] Allow test failures to be ignored
In a previous commit [1] the WPT CLI was modified to unconditionally
ignore the return value of the `wptrunner.start` method. Although this
behavior is convenient in cases where there is no expectation data, it
obscures errors that are relevant in any context.
Specifically, the `start` method returns a non-zero value when no tests
are run, but this is a reliable indicator of an erroneous configuration.
Because the `wpt run` command is used to validate the infrastructure of
the web-platform-tests project, ignoring this case allows faulty
configurations to go unnoticed [2].
Implement a new CLI option named `--no-fail-on-unexpected` to allow
users to allow test failure while still being alerted to errors that
cause zero tests to be executed. Update the script designed for the
TaskCluster service to enable this option.
[1] 5a1b0365267b540c7f4feac3e0a85e9146997fcd
[2] https://github.com/w3c/web-platform-tests/pull/10721
--
wpt-commits: 28e519569775b44a600ce09bd50320e0d9679db1, 8412dad42a7ca4c963aaa6dba771772dbe8d8c5c, f7a0aa07816a2f5a9af6c569f61a5d877baa3a08
wpt-pr: 10742
Automatic update from web-platform-tests[infra] Validate changes to infra tests in CI
The `infrastructure/` directory contains integration tests for the
project's infrastructure. Ensure that any changes to the files in that
directory are validated by the continuous integration system at the time
they are submitted.
--
[infra] Remove reference to non-existent directory
--
wpt-commits: 0ed30960df05eb1fcf590aa747db0c616b298e61, ef1cdebc46f6a49632a6015bc36b0ca9115b6770
wpt-pr: 10803
Automatic update from web-platform-testsFetch: cleanup api/cors/cors-filtering
I forgot about this one last time around somehow. No .html left in this directory after this one though.
--
wpt-commits: 4f38f7d34855b1cac6de739fa3c32cfdb9a7b7de
wpt-pr: 10932
Automatic update from web-platform-testsSend boundary events rightaway after pointerup
If pointerup implicitly releases the pointer
capture this change sends the boundary events
rightaway after that without waiting for the
next event.
Bug: 834824
Change-Id: I538e0b35d1420574f964666aac394c0bde00f4fb
Reviewed-on: https://chromium-review.googlesource.com/1019347
Commit-Queue: Navid Zolghadr <nzolghadr@chromium.org>
Reviewed-by: Dave Tapuska <dtapuska@chromium.org>
Reviewed-by: Mustaq Ahmed <mustaq@chromium.org>
Reviewed-by: Ella Ge <eirage@chromium.org>
Cr-Commit-Position: refs/heads/master@{#557184}
--
wpt-commits: 13bef404744b9a16919a5e508c72adfb00b41b04
wpt-pr: 10537
Automatic update from web-platform-testsFix #10875: run parsed lints for visual tests
CSS tests, due to their metadata requirement, default to visual tests,
hence it's important we run all of these lints (to in this case catch
../testharness.js)
--
Extend the previous commit to also check manual tests
Plenty of manual tests use testharness.js so we should check them too,
and indeed there's plenty of broken tests so fix them too
--
wpt-commits: b54c11b055959abeefafcde601853ea4cb247e0b, c327c2747db6b71c8c45f61e0a97785a9be622c2
wpt-pr: 10876
Automatic update from web-platform-testsXHR: reduce try/catch usage
This also improves checking for the correct exception by a bit.
--
wpt-commits: 9b96b146321b9f6cf2391225e3b936e4a68ca5a8
wpt-pr: 10903
Automatic update from web-platform-testsPaymentHandler: Change topLevelOrigin to topOrigin (#10872)
The related spec was changed:
https://github.com/w3c/payment-handler/pull/301
Bug: 840041
Change-Id: Ic76fffe0b974210edc3888378ded197d06e035b7
--
wpt-commits: da28291ff0b1b6ee6a36004e8deef9c2f098935d
wpt-pr: 10872
Automatic update from web-platform-testsPaymentRequest: Apply [Exposed=Window] to the PaymentRequest API (#8641)
There is no behavior change but add a test to check it.
--
wpt-commits: ce44721696b37c0d9c4629435f2d4ca692d8a272
wpt-pr: 8641
We want to have some class names with spaces in them, but everything assumes
that an interface identifier is in fact an identifier (e.g. uses it in C++
identifiers like namespace names).
Compared to the spec, we had the following differences:
* Date was a JSON type in our implementation. It doesn't even exist as a type
in the spec. It stops being a JSON type.
* Annotated types are not supported yet. Nothing changes here.
* Typedef types were not JSON types in our implementation. They become JSON
types if the type it's a typedef for is one.
* Frozen arrays are not supported yet. nothing changes here.
* Records were not JSON types in our implementation. They become JSON types
when the value type is a JSON type.
* Object was not a JSON type in our implementation. It becomes a JSON type.
* Interface types were only JSON types in our implementation if they had a
jsonifier. We change to treating them as JSON types if there is a jsonifier
anywhere on the inheritance chain.
In terms of observable behavior, the following properties now get included by
toJSON methods that didn't use to be included:
PaymentResponse.details
Performance.mozMemory
both because they're of type "object".
Without this, we will start including mozMemory in performance.toJSON() even if
the pref for it is not set, once 'object' becomes a JSON type.
This changes behavior in the following observable ways:
1) We stop exposing PerformanceResourceTiming's .serverTiming in the JSON
serialization in insecure contexts.
2) We stop exposing PerformanceTiming's timeToNonBlankPaint and
timeToDOMContentFlushed in the JSON serialization unless the relevant
preferences are turned on.
1. Set A11Y_INSTANTIATORS to never expire.
2. A11Y_CONSUMERS was already set to never expire, but only because expiry was never explicitly declared (when this was allowed). Explicitly set it as never expiring now. This also means it can be removed from the associated whitelist.
3. Change the email for both so it gets to the team and me (since policy requires a responsible person and that's me now for these).
MozReview-Commit-ID: BgjMoVnVp6U
--HG--
extra : rebase_source : f5fbe6f1cec985d762523b32eb2d23a03d824168
This is possibly a temporary solution to avoid sending the user's address to the merchant until they've interacted with the shipping address picker.
MozReview-Commit-ID: 5q8BYr1rLwP
--HG--
extra : rebase_source : e6c5ecee96acc4aaa25c2082b7d6a654bb82e135
These tests use setTimeout to wait for shape-outside layout to hopefully
complete. This is non-deterministic race condition and leads to intermittent
failures. This change makes those tests instead trigger on the onload event
and removes the setTimeout infrastructure and replaces it with an assert.
Without this change, these tests have no guarantee about whether the rendering
will be complete by the time the test's JS fires.
MozReview-Commit-ID: CHrXUC3yCCT
--HG--
extra : rebase_source : b5642bf24abcbfacb7b94e95d4ca9c8d6486b98b
Doesn't look like it's strictly necessary, but suppresses the warning.
MozReview-Commit-ID: EQBar71yNVg
--HG--
extra : rebase_source : 063639000fd47b32343d47a8ca428c134fdce43b
moz-wptsync-bot
Dominic Farolino
Morten Stenshorne
Ovidio Henriquez
Mike Pennisi
Anne van Kesteren
Navid Zolghadr
Shachar Zohar
Geoffrey Sneddon
Philip Jägenstedt
pyup.io bot
Blink WPT Bot
Dorel Luca
Jon Coppeard
Boris Zbarsky
Csoregi Natalia
Gabriel Luong
prathiksha
Ed Lee
James Teh
Matthew Noorenberghe
Paolo Amadini
Brad Werth
Jonathan Kingston
Sebastian Hengst
Chris Pearce
Bogdan Tara
Jonathan Kew
Tim Taubert
Luke Wagner
Jan de Mooij