This is a first step towards having a standalone
mediacapture-stream.idl. It also makes the existing test fail fast in
case the getUserMedia promise is rejected.

…s. (web-platform-tests#7864)

The AXRole, AXSubrole, and AXRoleDescription are already being tested
by the aria-modal="false" tests. In addition, what we really need to
verify for AX API when aria-modal="true" is that the background content
has been removed from the accessibility tree. The role details of other
elements are irrelevant in performing that verification, and their
inclusion introduces noise into the results.

There seems to be a race on whether the count of datachannels
is incremented before a session goes to "connected". This CL
allows both 0 and 1.

The test would fail ~1/4000 times on a Linux workstation.

Bug: 771540
Change-Id: Ic250443c2addef995dbb26e7cd021a516dcd023b
Reviewed-on: https://chromium-review.googlesource.com/721379
Reviewed-by: Henrik Boström <hbos@chromium.org>
Commit-Queue: Harald Alvestrand <hta@chromium.org>
Cr-Commit-Position: refs/heads/master@{#509564}

* dynamic-imports-error.html was converted to dynamic-imports-script-error.html which tests the additional cases of bad module specifiers and that the error objects are memoized appropriately.
* dynamic-imports-fetch-error.sub.html was introduced to test a large variety of possible fetch errors.

Also: fix errorhandling.html (a test for static imports) which was sending X-Content-Type-Options: nosniff. No sniffing should be performed for modules no matter what; the header is not relevant and might be masking a broken implementation of that requirement.

…orm-tests#7680)

Missing coverage noted in w3c/IndexedDB#213

The test cases verify the `source` property of requests returned by operations made against stores, indexes, and cursors.

This makes the origin secure so that getUserMedia() doesn't fail.
Updated expected file too.

Bug: 771540
Change-Id: I2717e195d6fb52cd2dedf45dd5c48b05b1058823
Reviewed-on: https://chromium-review.googlesource.com/716719
Commit-Queue: Philip Jägenstedt <foolip@chromium.org>
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Cr-Commit-Position: refs/heads/master@{#509623}

As it follows from w3c/sensors#313

Change-Id: Ic81e118b2a3c8db6d8b30ee61ef8811b88a0f970
Reviewed-on: https://chromium-review.googlesource.com/721742
Reviewed-by: Reilly Grant <reillyg@chromium.org>
Reviewed-by: Kentaro Hara <haraken@chromium.org>
Commit-Queue: Mikhail Pozdnyakov <mikhail.pozdnyakov@intel.com>
Cr-Commit-Position: refs/heads/master@{#509707}

In some cases, relatedTarget kept its value after event dispatch has
finished even though target is set to null. See test file for the cases.

This change adds a check so that when target is set to null in the
post-processing phase, the change is also reflected in relatedTarget.

Bug: 771580
Change-Id: Idda5d00c1a793b2669be66e76439851d1f66459c
Reviewed-on: https://chromium-review.googlesource.com/720897
Reviewed-by: Hayato Ito <hayato@chromium.org>
Reviewed-by: Takayoshi Kochi <kochi@chromium.org>
Commit-Queue: Rakina Zata Amni <rakina@chromium.org>
Cr-Commit-Position: refs/heads/master@{#509710}

* Remove SpecialPowers from css parsing tests

This is Gecko-specific and doesn't even exist in wpt tests run in
Gecko.

* Update test_font_family_parsing.html

…ts#7872)

…m-tests#7878)

Previously this always looked for changed files as required for
CI. But it's also useful to test an explicit list of files locally, so
check if this is passed before using the changed files.

checkLayout() method defined in resources/check-layout-th.js
had a outputContainer argument that was not used.
This patch removes that argument.

On top of that, some tests that were trying to use that argument
have been modified as it was not needed for them.

Last, the copies of check-layout-th.js in WPT CSS test suites
have been updated with this change and another change introduced
in r377261.

Change-Id: Iefb649d811e0fd09106bf3d02225539e481fa713
Reviewed-on: https://chromium-review.googlesource.com/721545
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Commit-Queue: Manuel Rego Casasnovas <rego@igalia.com>
Cr-Commit-Position: refs/heads/master@{#509752}

 Fix wdspec new_session create tests

Both create_alwaysMatch.py and create_firstMatch.py were failing to
import support.create where the input data is shared.

E ValueError: Attempted relative import in non-package

Add __init__.py to support directory and remove the leading '.' from the
import command in both test files.

add Jerry Smith (@jdsmith3000) as an OWNER for MS Edge

Test parsing and serialization of the following properties:
- color-interpolation-filters
- filter
- lighting-color

In particular, we test that hue-rotate(0) is accepted.
w3c/fxtf-drafts#228

Some filter functions remain to be added to the tests.

Add tests for drop-shadow() with color before lengths.
w3c/fxtf-drafts#231

Add tests for filter functions with no arguments supplied.

The assertion in this test did not match what the test was doing,
and was probably a copy+paste leftover.

This adjusts and completes the test suite, reflecting the changes
resolved in
w3c/csswg-drafts#1598 (comment)

…latform-tests#7895)

Those tests are not currently run because there's another function
called test_invalid() overriding the previous one. Rename both as
test_invalid_values() and test_invalid_extensions().

This uploads a MANIFEST.json.gz file for every commit to master, so
that we can build tooling that attempts to download that file as an
initial manifest that can then have a rapid update rather than a slow
build.

Updating WebDriver tests session creation to have valid capabilities

Several of the tests in the WebDriver suite are sending capabilities for
the new session that are incomplete or incorrect. This commit corrects the
capabilites payloads for the new session call to be a valid capabilities
dictionary.

…dpate

[css-ui-3] Cursor test udpate

…m-tests#7904)

This means that manifest_upload and lint run on master. Although the
lint is not going to do anything at the moment since it only runs on
changed files. In the future we could make it pass --all in this case.

Several of the pywebsocket handlers used in tests had the problem that they
would generate a hand-crafted header and then pywebsocket would add its own
header immediately afterwards, causing unnecessary errors.

In most cases there was no need to generate a hand-crafted header at all, so the
tests have been modified to use pywebsocket APIs to do the same
thing. simple_handshake_wsh.py still generates a hand-crafted header but now
suppresses pywebsocket's own header.

Several tests closed the WebSocket on the client side but the handler didn't
wait to receive the client-side close. These have been modified to wait for the
close frame.

Several tests had expectations that onclose would never be called but then
called close() themselves, causing onclose to be called. In some cases this was
mitigated by calling t.done() before t.unreached_func() was called, but it has
been fixed by disabling the onclose handler after calling close().

This change only fixes a subset of handlers and their affected tests. It is not
intended to be comprehensive, and many issues remain.

Discussed in
w3c/csswg-drafts#492

Change-Id: I31172814ad0db99b2ede246f92c4611dab07c756
Reviewed-on: https://chromium-review.googlesource.com/759918
Commit-Queue: Eric Willigers <ericwilligers@chromium.org>
Reviewed-by: Darren Shen <shend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515117}

Copy the tests for transform streams from the whatwg/streams repository.

TransformStream has now been added to the Streams Standard so it seems
reasonable to move these tests to the web-platform-tests repository. A
follow-on change will remove them from the whatwg/streams repository.

There are no changes to the test Javascript files, except that some
calls to setTimeout() have been replaced with equivalent calls to
delay(). Also a duplicate test name in lipfuzz.js has been fixed.

All the tests have all been reviewed in their old location. All the HTML
files have been regenerated by generate-test-wrappers.js.

recording-streams.js has an additional "recordingTransformStream"
function which is used in some of the new tests. test-utils.js has an
additional "constructorThrowsForAll" test helper.

This expands and upstreams:
* chromium/frame-detached-by-navigation.html
* chromium/frame-removed.html

Chromium behavior seems a bit weird but for now upload the WPT
to get the test visible, and we can refine the expectations later.

Bug: 713732, 522791, 688116
Change-Id: I4af7fb668f9ef2e4aaff580ba129e6ae353f0e18
Reviewed-on: https://chromium-review.googlesource.com/760061
Commit-Queue: Matt Falkenhagen <falken@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515134}

In particular, this catches a bug in libsoup (used by WebKitGTK), where the
casing of headers depends on global state.

The exact value is not important for this test so it might as well be correct.

Fixes web-platform-tests#6471.

Convert some tabs to spaces in css/css-flexbox/

This CL ensures that a child iframe does not report paint timing entries
when only the parent does all of the painting.

Bug: chromium:736114
Change-Id: Idcba2bb9e93eeb7168e604840df4497df4073472
Reviewed-on: https://chromium-review.googlesource.com/755857
Commit-Queue: Nicolás Peña Moreno <npm@chromium.org>
Reviewed-by: Timothy Dresser <tdresser@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515253}

Remove the old stale copy of the spec and remove js/lodash.js which
is loaded in several places but not actually used for anything.

Bug: 782901
Change-Id: Idfdbc9f1cf2d9edf190d649168074f4d4ec66c73
Reviewed-on: https://chromium-review.googlesource.com/759225
Reviewed-by: Hongchan Choi <hongchan@chromium.org>
Commit-Queue: Raymond Toy <rtoy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515256}

The tests for element state were asserting a success after handling a user
prompt. However, the way the URL end point was formatted in the test, an
error response was being returned. This commit corrects that.

Tests change
https://svgwg.org/svg2-draft/changes.html#extend
"Made ‘foreignObject’ a graphics element."

…ests#8137)

Make testharness.js output something in SVG documents

SVG2: Test foreignObject is a graphics element

…rm-tests#8139)

This is testing web-exposed behavior. The test had a comment that
Response.body is still being specced, and now it is specced.

Bug: 446188, 688116
Change-Id: I4d725b9172389ad675404a817d81c321cdef13d0
Reviewed-on: https://chromium-review.googlesource.com/760496
Commit-Queue: Matt Falkenhagen <falken@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Reviewed-by: Tsuyoshi Horo <horo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515336}

We are changing how we compute the grid area's size of positioned
items, trying to make the logic as independent as possible of the
grid container's direction.

The new way of placing absolute positioned elements in their grid area
is based on the previously computed area's size. This provide a
mechanism to place an item in its area respecting the offsets derived
from the Content Alignment CSS properties and, which is more important,
ignoring the grid container's direction.

Additionally, we use the already implemented alignment logic that's
used for regular grid items. This patch provides almost full alignment
capabilities for positioned objects, with the exception of 'stretch'.

It's worth mentioning that some of the test cases using vertical-rl
writing mode will fail because of the bug 779105.

Bug: 603423, 777978

Change-Id: I8fb121418795244d13589d603007c70f58e1f2df
Reviewed-on: https://chromium-review.googlesource.com/672423
Reviewed-by: Manuel Rego Casasnovas <rego@igalia.com>
Commit-Queue: Javier Fernandez <jfernandez@igalia.com>
Cr-Commit-Position: refs/heads/master@{#515391}

This also fixes the IgnorableCodePoint test to use inputs which have
nontrivial ISO-2022-JP shift states; previously this test did not
actually ensure correct behavior when non-ASCII characters
representable in ISO-2022-JP and non-ASCII characters not
representable in ISO-2022-JP occur in sequence.

This also adds WPT coverage.

Prior to this fix, encoder state was not respected, leading to
incorrect interpretation of the replacements and sometimes following
bytes too, depending on whether the replacement lengths were even or
odd, and on whether the active state of the ISO-2022-JP G0 character
set was one-byte or two-byte. An example, with results transcribed in
Unicode for readability:

Input: ABC~¤•★星🌟星★•¤~XYZ

Old output:
Bytes: ABC~&web-platform-tests#164;&web-platform-tests#8226;␛$B!z@1🌟@1!z&web-platform-tests#8226;&web-platform-tests#164;␛(B~XYZ
Meaning: ABC~&web-platform-tests#164;&web-platform-tests#8226;★星Γ渦祁卦酸院惕８臆胸Γ蔚柑~XYZ

New output:
Bytes: ABC~&web-platform-tests#164;&web-platform-tests#8226;␛$B!z@1␛(B🌟␛$B@1!z␛(B&web-platform-tests#8226;&web-platform-tests#164;~XYZ
Meaning: ABC~&web-platform-tests#164;&web-platform-tests#8226;★星🌟星★&web-platform-tests#8226;&web-platform-tests#164;~XYZ

Bug: 782565
Change-Id: If2a7b76b99ce77cbec433af5384ed5c4d2e3c581
Reviewed-on: https://chromium-review.googlesource.com/758405
Commit-Queue: Benjamin Wiley Sittler <bsittler@chromium.org>
Reviewed-by: Jungshik Shin <jungshik@google.com>
Reviewed-by: Joshua Bell <jsbell@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515425}

… next src

This fixes a bug where @font-face with font-display: optional causes
unnecessary font downloads of subsequent sources when the first source
fails to download by the "short time limit" (100ms).

Bug: 770978
Change-Id: I24ad7c803d04b9eb310b6b916c999660a23c61b5
Reviewed-on: https://chromium-review.googlesource.com/760137
Reviewed-by: Takashi Toyoshima <toyoshim@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Commit-Queue: Kunihiko Sakamoto <ksakamoto@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515453}

This implements link rel=modulepreload without submodules fetch,
behind the Experimental Web Platform feature flag.

LinkLoader fetches module script using Modulator::FetchSingle, so it
populates the module map but does not fetch descendant modules.
Unlike the preload fetches by link rel=preload, this doesn't set
FetchParameters::SetLinkPreload flag, so it works like a regular
script fetch.

HTMLPreloadScanner fetches module scripts in a similar way to
speculative preload of <script type=module>, i.e. just fetch a
ScriptResource and not populate the module map.

Spec PR: whatwg/html#2383
Intent to implement:
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/ynkrM70KDD4

Bug: 740886
Change-Id: I5f4420e305f4962b3fbe3f703ce95a5a43e4f7a9
Reviewed-on: https://chromium-review.googlesource.com/662697
Commit-Queue: Kunihiko Sakamoto <ksakamoto@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Reviewed-by: Kouhei Ueno <kouhei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515490}

Currently, EventTarget cannot be constructed or be subclassed (in JavaScript).
The spec for EventTarget was updated to allow constructing and
subclassing EventTarget.

A new class, EventTargetImpl, is created as the user-constructable version of
EventTarget. Even though the name is EventTargetImpl, it will be exposed as
EventTarget to JavaScript. We are using a hacky approach by making
EventTarget::Create return an instance of EventTargetImpl, which means an
instance of EventTargetImpl will be returned when the EventTarget object is
constructed in JavaScript.

Spec:
https://dom.spec.whatwg.org/#dom-eventtarget-eventtarget

Design doc:
https://docs.google.com/a/google.com/document/d/1TXMAkLeaNpbbETeqTyJZ57YnmqzgchyhFsANVZ2Ow3Y/edit?usp=sharing

Bug: 740576
Change-Id: Ia0569ee99c4d856120c387ca6aceed5fbe1491b5
Reviewed-on: https://chromium-review.googlesource.com/754248
Reviewed-by: Hayato Ito <hayato@chromium.org>
Reviewed-by: Yuki Shiino <yukishiino@chromium.org>
Commit-Queue: Rakina Zata Amni <rakina@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515495}