Automatic update from web-platform-tests
Fix the `test_tests_affected_null` test
As the comment says, this "will fail if the file we assert is
renamed", and that's what happened:
https://github.com/web-platform-tests/wpt/pull/18688https://github.com/web-platform-tests/wpt/pull/18743
Also change the commit to one which came after these changes, not
because it makes a difference to the outcome but for anyone inspecting
the commit to see the same files.
--
wpt-commits: 253dcf4bc4a5d8bec265a4aa719a22f737777819
wpt-pr: 18819
Automatic update from web-platform-tests
[WPT/common/security-features] Support nested source contexts
To enable `spec.src.json` to specify (possibly nested) source contexts,
this CL adds `source_context_list_schema` that represents
source contexts from where subresource requests are sent,
and how policies are delivered (by source contexts or subresource requests).
This CL adds `source_context_list` field to each `test_expansion` entry,
of which value is the key of `source_context_list_schema` to be used.
This CL adds `source_context_schema` and `subresource_schema`
that represent supported delivery types and subresources
for each source context or subresource type. These are used:
- To filter out test files for unsupported combinations of delivery types,
source contexts and subresources.
A part of the filtering was removed in
https://chromium-review.googlesource.com/c/chromium/src/+/1717535
and this CL re-adds it (in a different form of implementation).
- To determine what delivery types should be used for `anotherPolicy`
placeholder.
After this CL, the generate.py generates `source_context_list` and
`subresource_policy_deliveries` fields that are passed directly
without coversion.
Specifically,
- The conversion from `classic-data-worker-fetch` to
`classic-data-worker` + `fetch-request` in mixed-content-test-case.js
is removed.
- <iframe srcdoc>-related tests for `img-tag` referrer-policy tests
are split into separate tests, and related code in
referrer-policy-test-case.sub.js is removed.
Instead, to support tests inside <iframe srcdoc>, this CL adds
`referrerSource` calculation that can support <iframe srcdoc>s.
This CL also changes the structure of generated file names.
Now all files are generated under a dedicated directory `gen/`.
This CL creates `test-case.sub.js` instead of modifying
`referrer-policy-test-case.sub.js` etc., because:
After this CL, I'll commit new generated test HTML files,
incrementally in order to monitor potential regressions,
Because thousands of files are modified, split, and renamed.
During the incremental commits, `test-case.sub.js` is used
by new generated tests while
`referrer-policy-test-case.sub.js` etc. are still used by
old tests.
After all commits are done, `referrer-policy-test-case.sub.js`
etc. will be removed.
This CL renames subresource names, to make the names of subresources,
source contexts and source context lists consistent.
`test-case.sub.js` reflects renaming of JSON keys and values done
in previous commits.
- opt_in_method -> delivery_type/value
https://chromium-review.googlesource.com/c/chromium/src/+/1717107
- delivery_method -> delivery_type
https://chromium-review.googlesource.com/c/chromium/src/+/1716109
- referrer_policy -> delivery_value
https://chromium-review.googlesource.com/c/chromium/src/+/1717592
- source_protocol -> source_scheme
https://chromium-review.googlesource.com/c/chromium/src/+/1717115https://chromium-review.googlesource.com/c/chromium/src/+/1718991
- target_protocol + origin -> origin
https://chromium-review.googlesource.com/c/chromium/src/+/1717470
- Use shorter values for origin and redirection
https://chromium-review.googlesource.com/c/chromium/src/+/1717475
- referrer_url -> expectation
https://chromium-review.googlesource.com/c/chromium/src/+/1717319
- source context names
https://chromium-review.googlesource.com/c/chromium/src/+/1777110
Bug: 906850
Change-Id: Ia6695081e5a93355c29fb1bdb7e89ba6b673261f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1586548
Commit-Queue: Hiroshige Hayashizaki <hiroshige@chromium.org>
Reviewed-by: Mike West <mkwst@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692920}
--
wpt-commits: 03267447fc0f5a76c8ab2e22b7549798ca7419d8
wpt-pr: 18812
Automatic update from web-platform-tests
Use testdriver Action API in WPT coalesced pointerevent tests
In order to run the web-platform-tests automatically, we will use
testdriver Action API in all the wpt to simulate inputs. Here we are
changing a coalesced pointerevent input tests which sends more than one
mouse events in one frame and coalesces some of them together.
Bug: 606367
Change-Id: I4f094d01bd0b5ff55dd9df1d97a9ac4a78e41ffd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1775185
Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692783}
--
wpt-commits: adab23aed9f6f1d112a86a51d5aee6132fe813da
wpt-pr: 18802
Automatic update from web-platform-tests
Use promise in wpt tests to end the tests to avoid memory leak - Part 2
When we add default tick duration of two-frame time, we see some tests
memory leak which may end the tests when the page is still active, so
I add promise to each tests, once the actions are all finished, we will
call the test end in the promise callback.
Bug: 606367
Change-Id: Ief54ab266219cbda861fa230a4083c9f77b793b9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1775738
Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692701}
--
wpt-commits: 7199ea486be4ad56c3253b1a5700adc3caa33b27
wpt-pr: 18753
Automatic update from web-platform-tests
Use promise in more wpt pointer event tests to end the tests
Instead of adding a variable to check if the test should be finished,
we can create a promise object and end the test at its last callback.
This changes the rest of pointer event wpt tests.
Bug: 606367
Change-Id: I8cb2c9c074bc376d6d628c775f050184f2af9f8f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1772687
Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692744}
--
wpt-commits: bac3ea5db517f14f8a70ba0fc404ac831d6d40d2
wpt-pr: 18732
Automatic update from web-platform-tests
[LayoutNG] Fix NGLineBreaker::ShouldHangTraillingSpaces
This patch fixes |NGLineBreaker::ShouldHangTraillingSpaces|
to check the style of the given item, rather than the
|auto_wrap_| value which is from the style of the item
currently being handled.
Using the incorrect style has caused an infinite loop under
a specific pattern.
Bug: 988832
Change-Id: I907116dc239073f207c7c607953c98f722071150
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1782214
Reviewed-by: Emil A Eklund <eae@chromium.org>
Commit-Queue: Emil A Eklund <eae@chromium.org>
Commit-Queue: Koji Ishii <kojii@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692727}
--
wpt-commits: 46e85380444eacc080681e66b629114bd3292d5b
wpt-pr: 18801
Automatic update from web-platform-tests
Temporarily send `Sec-Fetch-*` headers via non-secure transport.
After [1], we're stripping `Sec-Fetch-*` headers from non-secure
transport. This is lovely! Except insofar as it also strips the headers
from secure redirect targets. Which is a problem, as it gives attackers
the option of laundering their requests through HTTP to strip the
headers.
This patch removes the secure transport restriction while we figure out
a cleaner way of removing the headers on a per-hop basis.
[1]: https://chromium-review.googlesource.com/c/chromium/src/+/1647354
Bug: 995745, 971938, 964053
Change-Id: Icec4e685902b7be2983bb81b7289ac9b45467782
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1762079
Commit-Queue: Mike West <mkwst@chromium.org>
Reviewed-by: Eric Lawrence [MSFT] <ericlaw@microsoft.com>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#689428}
--
wpt-commits: b4c94c8375815a279e00193abad2a499e105b519
wpt-pr: 18564
Automatic update from web-platform-tests
Load Ahem as web font in vars-font-shorthand-001.html (#18485)
--
wpt-commits: 8f45c16c164e8b8f6aaaae58d85fa3ba628c88d2
wpt-pr: 18485
Automatic update from web-platform-tests
Indexed DB: Fix compound index keys vs. autoincrement stores
Stores with (1) a keyPath (a.k.a. inline keys) and (2) autoIncrement
(a.k.a. a key generator) have primary keys generated by the browser
and inserted into objects lazily, since the renderer doesn't know what
the primary keys will be when the object is serialized.
Indexes might have a keyPath that references the same spot in the
object. This is handled by checking if the keypaths match. If so,
the browser synthesizes the index key (same as the primary key). But
Chrome was not handling the case where the index had a compound
key - a keypath that's an array, plucking multiple values out of
the object. An object with unresolved keypaths would normally just
not be indexed, per spec. But since the primary keys should be injected
before the indexing occurs, these should be indexed.
Fix this by sending the index keys from the renderer to the browser
as an array with "holes" that need to be filled in.
This is covered by an existing Web Platform Test, which we now pass.
Bug: 701972
Change-Id: I14940b23cfcbb7f8b673143b402f574220184fd7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728058
Commit-Queue: Joshua Bell <jsbell@chromium.org>
Reviewed-by: Kenichi Ishibashi <bashi@chromium.org>
Reviewed-by: Chase Phillips <cmp@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692313}
--
Mark the test as slow
--
wpt-commits: 980ea8f8e8e2df5ce1694620e5858380c30e02ed, ed6c6286a0273e9ee09406ccffb9d70e123273b4
wpt-pr: 18396
Automatic update from web-platform-tests
[WPT/referrer-policy] Call getRequestURLs() for each subtest
Previously, getRequestURLs(), PolicyDelivery/Subresouce
object creation etc. were done once per test HTML file and
reused among multiple subtests.
This CL introduces invokeScenario() that does these things and
calls invokeScenario() for each subtest, and thus
calls getRequestURLs() for each subtest.
This removes hacky modifications to `subresource.url`
in 4K-referrer-length-related subtests to avoid cache hit.
Also this makes subresource URLs unique for
srcdoc-related subtests (previously the same image URL
was used).
Bug: 906850
Change-Id: I894a902338a90526678b70d01857103f46984189
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1723013
Commit-Queue: Hiroshige Hayashizaki <hiroshige@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Reviewed-by: Mike West <mkwst@chromium.org>
Cr-Commit-Position: refs/heads/master@{#690288}
--
wpt-commits: bda02313aa4ba9975b3efd80acf9cef094dc52c8
wpt-pr: 18683
Automatic update from web-platform-tests
Remove random strings from a failing test.
We don't implement the underlying feature for [1] yet. This patch
ensures that it has consistent error state so that we can land
expectation files instead of skipping it.
[1]: external/wpt/html/cross-origin-opener-policy/coop-navigated-popup.https.html
Bug: 994214
Change-Id: Icfe42f553d35265d83f19f244f0f8bcc2ce3743d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1778615
Commit-Queue: Mike West <mkwst@chromium.org>
Reviewed-by: Kamila Hasanbega <hkamila@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692584}
--
wpt-commits: 1ac959f3e6198767fecf67740d3e1687b184265f
wpt-pr: 18783
Automatic update from web-platform-tests
Add infrastructure test for uncaught exception
To make current behavior clear.
This would change with https://github.com/web-platform-tests/wpt/pull/11024.
--
wpt-commits: 57a40f11e5f549e672188851227b47759c22d133
wpt-pr: 18794
Automatic update from web-platform-tests
Load Ahem as web font in white-space-pre-wrap-trailing-spaces-004.html
--
wpt-commits: 3cc474f1fa825abd7ab33f23c0804695c7bb6e35
wpt-pr: 18484
Automatic update from web-platform-tests
Expect Safari to pass infrastructure/testdriver/actions/pause.html
It appears to pass now, whether it reliably passes is not clear.
--
wpt-commits: 5645f4c61207dd29ffc1e5960caf7f4b204089d4
wpt-pr: 18795
Automatic update from web-platform-tests
LayoutVideo should not mark NeedsLayout in layout
This patch fixes |LayoutVideo| not to mark |NeedsLayout|
even when intrinsic size was changed, if it is during the
layout. Layout clears |NeedsLayout| of its inline ancestors
that doing this can leave them |NeedsLayout| set after the
layout was complete.
This is probably an old issue, but there were no code path
to hit it, unitl r684192 (crrev.com/c/1703540) changed the
intrinsic size computations.
yosin@ pointed out that |intrinsic_size_| is not used at all
when |ShouldApplySizeContainment()| since r684192, so we can
probably skip |LayoutVideo::UpdateIntrinsicSize()|. This
should fix the problem and also is supposed to improve the
performance a bit. This patch takes more conservative
approach to ensure we never |SetNeedsLayout| of inline
ancestors during the layout.
Bug: 998395
Change-Id: I735780bbc3b65321796d217934b57c6be2d2adbd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1780986
Reviewed-by: Yoshifumi Inoue <yosin@chromium.org>
Commit-Queue: Koji Ishii <kojii@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692541}
--
wpt-commits: b86e07e58522387be6929b72b4c50ba01bc15ce9
wpt-pr: 18784
Automatic update from web-platform-tests
Update wpt test to match the new spec where loadend isn't anymore
--
wpt-commits: 02e1c03ed4e47006acd2d40f0c9c679f7f9db52d
wpt-pr: 18781
Automatic update from web-platform-tests
html, svg: Add a test for 'autofocus' as a global attribute (#18558)
For https://github.com/whatwg/html/pull/4830
--
wpt-commits: cfb69fe84608b80823905e9f257e2ef7700362bc
wpt-pr: 18558
Automatic update from web-platform-tests
Add infrastructure expectations for Epiphany (#18589)
* Tested with WebKitGTK version 2.25.4 on Debian experimental.
--
wpt-commits: d865af1b41bd7fda439930fc19a0389128538edb
wpt-pr: 18589
Automatic update from web-platform-tests
Fix 2dcontext/building-paths ref-tests to use a canvas instead of pre-generated images. (#18280)
* The tests 2dcontext/building-paths/canvas_complexshapes_arcto_001.htm
and 2dcontext/building-paths/canvas_complexshapes_beziercurveto_001.htm
were using a pre-generated image to compare with the rendered canvas.
But this is not reliable as there is no guarantee that the rendering
of the canvas has to be exactly pixel-by-pixel identical across
different Web engines. Things like canvas anti-aliasing features, or
platform differences can make the visual output to be slightly different
(yet still correct).
* So this patch changes the reference file of this tests to render the
expected result using a canvas as well.
* It has been tested that this 2 tests now passes on Chrome, Firefox, and WebKit.
--
wpt-commits: 9eabcfe84fef0281a7335ec526275722b3625b5b
wpt-pr: 18280
Automatic update from web-platform-tests
OOR-CORS: Make the CORS preflight have Accept: */*
The current spec does not require the CORS preflight request to
have an Accept header. But Firefox sends it, and Chrome sent it
until OOR-CORS is enabled.
https://github.com/whatwg/fetch/issues/922
Since there is no reason not to have the header, and no objections
at the spec proposal thread, we'll restore the Accept header to
the CORS preflight request.
Bug: 995740
Change-Id: I50854e65770df593c8bbcff7ee85404051a23f36
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1768503
Commit-Queue: Takashi Toyoshima <toyoshim@chromium.org>
Reviewed-by: Ben Kelly <wanderview@chromium.org>
Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692455}
--
wpt-commits: 9614edb355132c8191c99518d2f83d9751dcf935
wpt-pr: 18739
Automatic update from web-platform-tests
[LayoutNG] Don't allow multicol + custom layout on the same node.
This should be a useless combination, just like it is to combine flex,
grid and tables with multicol.
Multicol should only be enabled for nodes that would otherwise create a
block container (NGBlockLayoutAlgorithm).
Move the custom layout option higher up in the algorithm picker, so that
it trumps multicol. Also avoid creating the multicol flow thread if we
have custom layout, or we'd confuse other pieces of the engine, when we
in fact use custom layout and not multicol. As we move forward with the
NG block fragmentation implementation, we'll eventually stop creating
the flow thread, but for now, it's still in use, because we use the
legacy engine for painting.
Bug: 997403, 998423
Change-Id: Id8577da0b129d9182fb9195f508dd57195290ee4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1774277
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: Christian Biesinger <cbiesinger@chromium.org>
Reviewed-by: Koji Ishii <kojii@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692447}
--
wpt-commits: c05529a9dfbda874a8f1f48eb20248a794c52e71
wpt-pr: 18741
Automatic update from web-platform-tests
No need to invalidate style in an empty document.
InvalidationRootForTreeScope() assumes we have documentElement() for the
document scope. Exit early from @keyframes invalidation if it's null.
Bug: 999522, 999521, 999524, 999656
Change-Id: Iec34f4871087c50fd2d067d955f26e3eb02536b5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1780724
Reviewed-by: Morten Stenshorne <mstensho@chromium.org>
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Commit-Queue: Rune Lillesveen <futhark@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692443}
--
wpt-commits: 580bc2dba627bfe2e9b81c254d32b76e90a9c24f
wpt-pr: 18786
Automatic update from web-platform-tests
Screen Enumeration: Basic implementation.
This CL exposes a subset of the screen properties outlined in the Screen
Enumeration API explainer:
https://github.com/spark008/screen-enumeration/blob/master/EXPLAINER.md.
The API is gated behind the experimental flag "ScreenEnumeration".
Bug: 994889
Change-Id: I3cd52f0bbfd4df11cd999bf1d21059bf298109f2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1759890
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Michael Wasserman <msw@chromium.org>
Reviewed-by: Victor Costan <pwnall@chromium.org>
Commit-Queue: Staphany Park <staphany@chromium.org>
Cr-Commit-Position: refs/heads/master@{#692420}
--
wpt-commits: 11eb4006e31e9c870bd6a8e26372d24c80d7d04d
wpt-pr: 18571
Automatic update from web-platform-tests
[css-inline] Inheritance of line-height (#18667)
line-height inherits, and has initial value 'normal'.
https://www.w3.org/TR/CSS2/visudet.html#propdef-line-height
--
wpt-commits: 5c7962550bccfaa68412585deb60f0b648a1dd7c
wpt-pr: 18667
`HTMLEditRules::MakeSureElemStartsOrEndsOnCR()` is unclear what it does.
And fortunately, its job is simple if we split it to each mode. Therefore,
this patch splits it to `EnsureHardLineBeginsWithFirstChildOf()` and
`EnsureHardLineEndsWithLastChildOf()`. Finally, the batch caller of them,
`HTMLEditRules::MakeSureElemStartsAndEndsOnCR()` is removed since directly
calling both of them is clearer.
Differential Revision: https://phabricator.services.mozilla.com/D44787
--HG--
extra : moz-landing-system : lando
FuncTypeWithId::args returns a Vec<ir::Type>, most of which are pretty
small. Replacing it with a SmallVec::<[ir::Type; 4]> reduces the number of
allocation (calls) for compiling wasm via CL by about 2.3% and does not
increase the instruction count.
Differential Revision: https://phabricator.services.mozilla.com/D45132
--HG--
extra : moz-landing-system : lando
Philip Jägenstedt
moz-wptsync-bot
fantasai
Hiroshige Hayashizaki
TAMURA, Kent
Lan Wei
Koji Ishii
jugglinmike
Mike West
Joshua Bell
Ms2ger
autofoolip
Yutaka Hirano
Bastien Orivel
Carlos Alberto Lopez Perez
Takashi Toyoshima
Morten Stenshorne
Rune Lillesveen
Staphany Park
Eric Willigers
L. David Baron
Masayuki Nakano
Julian Seward
Itiel
Johann Hofmann