Automatic update from web-platform-tests
[CSP] Fix outdated comment in WPT
The WPT about inheritance of CSP for srcdoc iframes on history
navigations had been changed: first we wanted to always inherit CSP
from the parent, then we agreed to actually store the CSPs in history.
The comments on the test are still outdated and confusing, so let's
fix them.
Bug: 1329190
Change-Id: I366be21a1b845f2b22c1555874ae4433b57e6069
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3779760
Reviewed-by: Arthur Sonzogni <arthursonzogni@chromium.org>
Auto-Submit: Antonio Sartori <antoniosartori@chromium.org>
Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1028186}
--
wpt-commits: a0058bfc42ac18dd866188c2d1949b7aa0fadde0
wpt-pr: 35231
Automatic update from web-platform-tests
HighlightInheritance: fix implicit defaulting of ‘color’ at :root
Highlight pseudos redefine the inherited value of ‘color’ for :root to
be ‘currentColor’ (which is special for highlights) rather than the
initial value, such that the implicit default and ‘inherit’ and
‘unset’ all mean ‘currentColor’, while ‘initial’ still refers to the
initial value (‘CanvasText’).
The special concept of ‘currentColor’ for highlights is encoded in the
ColorIsCurrentColor extra field, with the help of css_longhand::Color
methods (for explicit defaults) and the default value of the field
itself (for implicit defaulting at :root).
This works if we actually resolve highlight styles at :root, such as
with getComputedStyle(document.documentElement) or if there are rules
for the :root highlight setting other properties. But if the :root
highlight is completely unstyled, the StyleResolverState ctor will
default its parent style to the ElementResolveContext’s parent style,
that is, the originating(!) parent element style.
This patch fixes that “initial-value bug” (CL:3173223), by making the
StyleResolverState ctor leave the parent and layout parent unset when
resolving highlight styles with highlight inheritance.
To keep tests passing, we also need to fix a bug where highlighting a
visited link yields a currentColor foreground when the visited style
isn’t currentColor but the unvisited style is currentColor. This patch
does so by splitting ColorIsCurrentColor (CL:3506351, CL:3563099) into
a pair of flags, one visited and one unvisited.
Fixed: 1341621
Change-Id: I2ac356ef8c69756d5d95cd09e639ad32980d41bd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3763095
Commit-Queue: Delan Azabani <dazabani@igalia.com>
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1028184}
--
wpt-commits: 48cf93337096b093fa94c13c484b5ad87a33b3f7
wpt-pr: 34860
Automatic update from web-platform-tests
Set contents_opaque(false) if the background color is transparent
Previously if a layer's background was transparent and the layer is
opaque, we used the layer_tree_host's background as
SafeOpaqueBackgroundColor(). One of the usages of
SafeOpaqueBackgroundColor() is to fill the otherwise translucent
antialiased pixels along the edge of rasterization, when the layer
has non-integral raster transform, to ensure the layer is still
fully opaque. However, the color might be dramatically
different from the normal color in the layer, causing edge artifacts.
Now set contents_opaque(false) in the case.
The case of opaque layer not having a background color only happens
in cases such as the layer has an opaque background image. In such
cases, using any heuristic color as the
SafeOpaqueBackgroundColor() may be inappropriate because the color
may be visibly different from the background image.
This also prevents seams between adjacent image layers when the layers
layer have non-integral to-screen transforms.
Bug: 1330509
Change-Id: I91edeb1c02da5eaa38ea0431f0559775b273f418
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3779000
Reviewed-by: Stefan Zager <szager@chromium.org>
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1028113}
--
wpt-commits: f7dc9fa74af1107a161755d004e4ae31a54e756b
wpt-pr: 34930
Automatic update from web-platform-tests
[@container] Clear ContainerQueryData on detach
If not re-attaching, there should be no reason to keep the container
query data. Also fixes a DCHECK/crash that if the element had a
container query evaluator the last lifecycle update, and still has the
same one, the old ComputedStyle should be non-null.
Bug: 1342750
Change-Id: I743abe62ec7597eccf1f9b146d60668d861b62d5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3782882
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Commit-Queue: Rune Lillesveen <futhark@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1027777}
--
wpt-commits: 6c4231c465fc3ed7b928ad662b8f29a5ed29e1fb
wpt-pr: 35223
Automatic update from web-platform-tests
Create dedicated tests for script.callFunction exceptionDetails
--
Update script.evaluate exceptionDetails test to match the callFunction test
--
wpt-commits: db107635bedbe38a7de852321b29eeccb4a5eced, afebdf7858e2e60232efc2e6156c6d173e78f669
wpt-pr: 34933
Automatic update from web-platform-tests
Fix DCHECK with pathLength/-0.0 being -infinity
This comment did not account for the behavior of negative zero:
A value of zero is valid and must be treated as a scaling factor of
infinity.
Bug: 1345806
Change-Id: Ib780c7201a6edffdc897ec63f79c1ff53e276ab5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3781420
Auto-Submit: Philip Rogers <pdr@chromium.org>
Commit-Queue: Philip Rogers <pdr@chromium.org>
Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1027491}
--
wpt-commits: 75ce9176fdca196bcc9d57fc15b5ea72288e75a0
wpt-pr: 34975
Automatic update from web-platform-tests
Add WPT-based <slot> fallback reference tests
During release M89, a bug was noticed after release and affected users
in production. Postmortem [1] was created to resolve this bug quickly.
This shows holes in the existing slots fallback tests, especially
when it comes to the dynamically rendered output.
To remediate this, we are adding reference tests for slot
fallback contents that cover most common use cases.
[1] https://docs.google.com/document/d/1IsypzwRE3-MLLFhz6ldaxxIYsn_vLRGPzvv9ftZSoqU/
Change-Id: Ib76d516178ac055591fb51bd04f9be4cd96addc6
Fixed: 1185209
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3778444
Reviewed-by: Mason Freed <masonf@chromium.org>
Commit-Queue: Di Zhang <dizhangg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1027425}
--
wpt-commits: 492ebc31d077bcfdd20dfb22f784d20f14a0e529
wpt-pr: 34923
Automatic update from web-platform-tests
Fix#34926: set SYSTEM_VERSION_COMPAT=0 when installing STP
Presumably Azure Pipelines (temporarily?) had an image which didn't
set SYSTEM_VERSION_COMPAT=1 by default, hence this previously appeared
to work without this.
--
wpt-commits: a2871e42d099de3bc40fe6eec937a8fc99eebb53
wpt-pr: 34942
Automatic update from web-platform-tests
Add "arguments" deserialization tests for "script.callFunction" (#34830)
* Add deserialization tests
* Use `recursive_compare` to assert result
Co-authored-by: Maksim Sadym <sadym@google.com>
--
wpt-commits: 073bed005f5f5a7f2772c83db772a04108c7925d
wpt-pr: 34830
Automatic update from web-platform-tests
[Private Network Access] Skip mixed content check for private access requests.
In this patch, we skip the mixed-content check for HTTPS websites
requesting to HTTP private/local devices if there's a local/private
targetAddressSpace provided as a fetch option.
This patch also combines the blink feature with chrome feature flag for
PrivateNetworkAccessPermissionPrompt.
In a following patch, we would check if the targetAddressSpace meets
the real IPAddressSpace in network services. If false, the request would
be denied in the network services.
This is part of the private network access permission prompt project.
See: https://docs.google.com/document/d/1Q18g4fZoDIYQ9IuxlZTaItgkzfiz_tCqaEAI8J3Y1WY/edit#
Bug: 1338439
Change-Id: I908ca9ee01a0415e58541b094416e22558e05a5e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3766742
Reviewed-by: Yifan Luo <lyf@chromium.org>
Reviewed-by: Camille Lamy <clamy@chromium.org>
Commit-Queue: Yifan Luo <lyf@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1027158}
--
wpt-commits: 5351921c7f494293594629624193592d7a01fbff
wpt-pr: 34914
Automatic update from web-platform-tests
Add a CEReaction test for Animation.prototype.commitStyles. (#34869)
* Add a CEReaction test for Animation.prototype.commitStyles.
Co-authored-by: Marcos Cáceres <marcos@marcosc.com>
--
wpt-commits: 907be0e72eef2687b3c320c1cb1acebf525648b2
wpt-pr: 34869
Automatic update from web-platform-tests
Implement support for :modal pseudo class
Currently :modal pseudo class is not working.
This CL fully implements it.
This CL also fixes WPT tests that are failling.
Bug: 1325861
Change-Id: I3097830c576cb71b2505db9aabc89d57a5bc608b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3655618
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Reviewed-by: Mason Freed <masonf@chromium.org>
Commit-Queue: Jihwan Kim <bluewhale.marc@gmail.com>
Cr-Commit-Position: refs/heads/main@{#1026764}
--
wpt-commits: 6cc880d5283965ca9c9b6eb2a8af4ba0c024f872
wpt-pr: 34676
The back/forward menu shows tooltips when hovering menuitems. For the
current page, the tooltip says Stay on this page, but the effect of the
menuitem is actually to reload the current page. So change the tooltip
string to reflect that the page will reload when the item is clicked.
Differential Revision: https://phabricator.services.mozilla.com/D153468
These comments come from bug 1219469 which had an implementation for QueryInterface that included the following:
> if (aIID.Equals(NS_GET_IID(HttpChannelParentListener))) {
> foundInterface = static_cast<nsIInterfaceRequestor*>(this);
> } else
That has been replaced by
> NS_INTERFACE_MAP_ENTRY_CONCRETE(HttpChannelParent)
so the comments are no longer correct, and nsIInterfaceRequestor doesn't need to be the first class we extend.
Differential Revision: https://phabricator.services.mozilla.com/D153568
