Three of `TestPrefetchService` overriding `PrefetchUrl()` are mostly the
same and thus merged into one.
The other one in `prefetch_url_loader_interceptor_unittest.cc` is
different and is renamed to `TestPrefetchServiceForInterceptor`.
Bug: 400761083
Change-Id: I1ed8813f1fa8df3615717bb1a04fa2335bab9805
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6337367
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Auto-Submit: Hiroshige Hayashizaki <hiroshige@chromium.org>
Reviewed-by: Ken Okada <kenoss@chromium.org>
Reviewed-by: Taiyo Mizuhashi <taiyo@chromium.org>
Commit-Queue: Hiroki Nakagawa <nhiroki@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432543}
For the ServiceWorker-controlled prefetches, `ServiceWorkerClient` is
also a part of prefetched result (as the prefetch record's request's
reserved client).
This CL is plumbing only and doesn't change the behavior (especially
with the flag disabled, where the returned `ServiceWorkerClient` is
always null).
Bug: 40947546
Change-Id: Id96d5e2d0a08151e2d6f28a6f481091187ae156a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6250403
Commit-Queue: Hiroshige Hayashizaki <hiroshige@chromium.org>
Reviewed-by: Taiyo Mizuhashi <taiyo@chromium.org>
Reviewed-by: Ken Okada <kenoss@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432522}
This is a reland of commit c6bfb8a155
Nothing changed from the original CL -- we reland the same change we did
originally. The reason for this change of mind is that we have found the
fix to the bug that prompted us to revert the change, and we intend to
land that (very scoped) fix in a separate CL we can then cherry-pick
directly to release branches.
Original change's description:
> [a11y] Fixes move by format for CSS spellcheck highlights
>
> Uses Spelling/Grammar error CSS highlight markers in the
> Accessibility Tree to break format boundaries around those
> errors to help navigate better around these spelling/grammar
> errors.
>
> This CL achieves following:
> 1) Fixes Move by format to accommodate CSS spellcheck highlights.
> 2) Narrator calls out the spelling issues correctly.
>
> Bug: 364795299
> Change-Id: I4842a49449c666100e31f859447f951e9f9eb09e
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6178990
> Auto-Submit: Vinay Singh <vinaysingh@microsoft.com>
> Reviewed-by: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
> Commit-Queue: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
> Cr-Commit-Position: refs/heads/main@{#1415710}
Bug: 364795299
Change-Id: I561aa47ece506a7077d2b7ef08c4ca6a74662a10
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6354023
Auto-Submit: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
Reviewed-by: Jacques Newman <janewman@microsoft.com>
Commit-Queue: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
Commit-Queue: Jacques Newman <janewman@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1432488}
This CL addresses an issue with CSS system colors where the values of
`Field` and `FieldText` were shared with `Canvas` and `CanvasText`
respectively. This occurred after switching to retrieve system colors
from color providers instead of the NativeTheme in Blink
(http://crrev.com/c/5322440).
Previously, for Windows, `Field` and `FieldText` were retrieved using
the NativeTheme in light mode and forced colors mode, sharing the same
colors as `Canvas` and `CanvasText`. However, for Dark mode on Windows
and non-forced colors scenarios for other OSs, we used
`LayoutTheme::DefaultSystemColor()`, which treated these colors
separately.
To fix this discrepancy, new color ids have been added to the color
pipeline for these system colors, allowing them to share the same colors
when appropriate and have distinct colors when necessary.
Bug: 400746852
Change-Id: I47f7c8dd56bbd66bdde04c70847a5274cc431cec
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6349557
Reviewed-by: Alison Maher <almaher@microsoft.com>
Reviewed-by: Kentaro Hara <haraken@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>
Commit-Queue: Sam Davis Omekara <samomekarajr@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1432463}
We refactor several of the static creator methods for
SharedStorageEventParams in preparation for adding several new members
to the class.
Bug: 401011862
Change-Id: Idfafcccce0f8ecd3d530e3223fc0036b79efd654
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6351803
Reviewed-by: Yao Xiao <yaoxia@chromium.org>
Commit-Queue: Cammie Smith Barnes <cammie@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432387}
The Captured Surface Control indicator in the tab-sharing infobar should
trigger for all write-actions. In the past, sendWheel() was the only
gesture-related action; now forwardWheel() replaces it, and it should
also trigger the indicator even before the user actually scrolls.
Bug: 402532195
Change-Id: I1c4f66d3f20a49e9364081e7e2ec21433b328c64
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6347522
Auto-Submit: Elad Alon <eladalon@chromium.org>
Reviewed-by: Guido Urdaneta <guidou@chromium.org>
Commit-Queue: Guido Urdaneta <guidou@chromium.org>
Commit-Queue: Elad Alon <eladalon@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432375}
This just adds basic functionality and doesn't include things like
full compaction support, expiration, manual clearing logic, etc., as
those will come in subsequent CLs.
The mojom type that will eventually be passed to the worklet process is
defined, but the plumbing to actually pass values to the worklet process
will come in a future CL.
Bug: 394108643
Change-Id: If7948579a78c909c0bd372ac4e5751cb8a86549d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6256880
Reviewed-by: Ayu Ishii <ayui@chromium.org>
Reviewed-by: Orr Bernstein <orrb@google.com>
Commit-Queue: Caleb Raitto <caraitto@chromium.org>
Reviewed-by: Dominic Farolino <dom@chromium.org>
Reviewed-by: Cammie Smith Barnes <cammie@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432359}
`loader_callback_` can resume navigation which can reentrantly call
PrefetchURLLoaderInterceptor::OnGetPrefetchComplete() on another
`PrefetchURLLoaderInterceptor` object, at least after
https://chromium-review.googlesource.com/c/chromium/src/+/6021277.
In order to call `GetPrefetchCompleteCallbackForTesting()` in the
correct order (i.e. the order of `OnGetPrefetchComplete()` calls), this
CL moves `GetPrefetchCompleteCallbackForTesting()` before
`loader_callback_`.
Bug: 40947546
Change-Id: Ic9b7d3aeb0d2bb2b63225862c094061b48d9a4b5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6347576
Commit-Queue: Hiroshige Hayashizaki <hiroshige@chromium.org>
Reviewed-by: Ken Okada <kenoss@chromium.org>
Reviewed-by: Taiyo Mizuhashi <taiyo@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432331}
The following `kPrefetchUseContentRefactor` params no longer exist and
thus are removed:
- block_until_head_eager_prefetch
- block_until_head_moderate_prefetch
- block_until_head_conservative_prefetch
- use_streaming_url_loader
Some no-op lines (e.g. enabling `kPrefetchUseContentRefactor` with no
parameters) are removed.
`ExemptSpeculationRulesHeaderFromCSP` is removed from
`VirtualTestSuites` because it is enabled by default.
Tests with identical features after the removal are merged:
- `virtual/prefetch-no-vary-search` -> `virtual/prefetch`.
- `PrefetchServiceStreamingURLLoaderTest`
-> `PrefetchServiceTest`
- `PrefetchServiceAllowRedirectsAndAlwaysBlockUntilHeadTest`
-> `PrefetchServiceTest`
- `PrefetchServiceAlwaysBlockUntilHeadWithTimeoutTest`
-> `PrefetchServiceAlwaysBlockUntilHeadTest`
`PrefetchServiceNeverBlockUntilHeadTest.HeadNotReceived` is removed
because the `{"block_until_head_eager_prefetch", "false"}` param etc.
are actually no-op and navigation is actually blocked-until-head but
timed out, which is already covered by
PrefetchServiceAlwaysBlockUntilHeadWithTimeoutTest.
BlockUntilHeadTimedout.
Bug: 403138272, 40947546
Change-Id: I75ea7738307bdd616fe60b4508d0c08e847888ed
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6343622
Reviewed-by: Taiyo Mizuhashi <taiyo@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Reviewed-by: Ken Okada <kenoss@chromium.org>
Commit-Queue: Hiroshige Hayashizaki <hiroshige@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432329}
This update ensures that a progress updates are normalized to zero
and 1 rounded to the nearest 0x10000th will.
Instead of updating OnDownloadProgressUpdate to a single double and
assuming that it is normalized properly, this ensures that it is
normalized properly by forcing the caller to pass it's total as
0x10000. This forces the caller to be aware that it should be normalized
to 0x10000.
Bug: 391715395
Change-Id: I3c25b7a737b9077d597c9f1d003471299858ced7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6300721
Reviewed-by: Christine Hollingsworth <christinesm@chromium.org>
Commit-Queue: Nathan Memmott <memmott@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432316}
At construction, we assign a monotonically increasing ID to each
SharedStorageWorkletHost, which will be reported to DevTools for
each worklet event.
The current CL only partially plumbs the worklet ID to DevTools by
setting it in the event params for addModule, createWorklet, run, and
selectURL.
Followups will have to send the ID for the remaining worklet events,
as well as update the DevTools protocol and forward to the DevTools
frontend.
Bug: 401011862
Change-Id: I554d3c2ac3b6ce9f239dddcebe71a7b52359517f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6348125
Reviewed-by: Yao Xiao <yaoxia@chromium.org>
Commit-Queue: Cammie Smith Barnes <cammie@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432308}
The network traffic annotation suggests the intent here is to
not include cookies in requests. However, the ResourcRequest's
credentials_mode is still kInclude.
If this is intentional, let me know and I will not change
credentials_mode, but we should also probably add a comment
to avoid future confusion.
Bug: 402803381
Change-Id: I657e8bb323827b95b18396340dc68cedf658af5a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6351143
Reviewed-by: Ken Buchanan <kenrb@chromium.org>
Commit-Queue: Dylan Cutler <dylancutler@google.com>
Cr-Commit-Position: refs/heads/main@{#1432305}
From a grt@'s review comment in CL:6336932 where we were creating ui/accessibility/assistive_tech.h:
"i don't think //ui/accessibility is the right place for this (nor for ax_mode_observer.h -- it belongs in //ui/accessibility/platform since it's part of the browser-side platform integration, not part of the codebase shared with blink/renderers).
how do you feel about moving this and the observer into //ui/accessibility/platform?"
Bug: none
Change-Id: Ie284ff3b3ccb7d0bc714da222b1054e3f8ae0d39
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6351802
Reviewed-by: Aaron Leventhal <aleventhal@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@google.com>
Auto-Submit: Aaron Leventhal <aleventhal@chromium.org>
Reviewed-by: Abigail Klein <abigailbklein@google.com>
Commit-Queue: Aaron Leventhal <aleventhal@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432302}
GURL lacks a single-argument constructor accepting a std::string;
rather, it accepts a std::string_view, so there is no need to create a
temporary string; sql::Statement::ColumnStringView can be used instead.
Bug: 403218784
Change-Id: I899cee54e4eb35ff6efaedb18152cdacf9d308a0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6351720
Owners-Override: Daniel Cheng <dcheng@chromium.org>
Commit-Queue: Andrew Paseltiner <apaseltiner@chromium.org>
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432241}
This change addresses the flakiness in the tests caused by improper
initialization of BucketLocator.
Key Points:
The expected count for various assertions of notify_list_changed_count
has been decreased. This adjustment is necessary because there are two
definitions in indexed_db_context_impl.cc that call
OnIndexedDBListChanged internally (OnFilesWritten and
DidForceCloseForDeleteBucketData).
When OnFilesWritten is called, incrementing the
notify_list_changed_count for each observer. During the Teardown phase
of IndexedDBTest, all observers are cleared.
Consequently, when DidForceCloseForDeleteBucketData is called, no
observers remain to trigger OnIndexedDBListChanged and increment the
count. Therefore, the expected count has been decreased in the new
implementation.
Bug: 41348374, 324111895
Change-Id: Idac145cfe04a79434b9e5d541511aaeb97386e0c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6275787
Auto-Submit: Garima Chadha <garimachadha@microsoft.com>
Commit-Queue: Abhishek Shanthkumar <abhishek.shanthkumar@microsoft.com>
Reviewed-by: Steve Becker <stevebe@microsoft.com>
Reviewed-by: Abhishek Shanthkumar <abhishek.shanthkumar@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1432173}
Also for net/cookies/cookie_monster.h,
net/cookies/cookie_partition_key.h and
services/network/restricted_cookie_manager.h
Removed unused includes, added forward declarations where possible.
Saves 150MB of compile size, according to the compile-size bot.
Bug: 40318405
Change-Id: I88ab337c0c660894a7f0a662d3d67b00e3bd1984
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6349077
Auto-Submit: Aldo Culquicondor <acondor@chromium.org>
Reviewed-by: Andrew Grieve <agrieve@chromium.org>
Commit-Queue: Andrew Grieve <agrieve@chromium.org>
Owners-Override: Andrew Grieve <agrieve@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432148}
This is a reland of commit 0996144caf
Changes:
* move accent color tests from chrome/browser to content/browser
* add accent color for Mac OS
Notes:
The implementation of Accent Color on Mac OS is mainly in
content/browser/*. To ensure the maintainability of future tests,
I have also relocated the Windows/ChromeOS tests to
content/browser/* and consolidated the code accordingly.
Original change's description:
> Enable accent color testing on MacOS
>
> I have verified on main that Accent Color related features
> are now working correctly on Mac, thanks to Priya's series
> of CLs:CL:5763388,CL:5785551,CL:5789206,CL:5793044,CL:5766135,
> CL:5909355 and Joey's CL:4523388.
>
> Bug: 40229450
> Change-Id: Ib51c4bdfc35e5f3bffbc2751198c8be03e5bfce4
> Reviewed-on: CL:6170547
> Reviewed-by: Joey Arhar <jarhar@chromium.org>
> Reviewed-by: Sam Davis Omekara <samomekarajr@microsoft.com>
> Reviewed-by: Alison Maher <almaher@microsoft.com>
> Commit-Queue: Huiting Chen <huitingchen@microsoft.com>
> Cr-Commit-Position: refs/heads/main@{#1409428}
Bug: 40229450
Change-Id: I63b0202d750d31e93d78d171e15f9565208dc590
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6265044
Reviewed-by: Alison Maher <almaher@microsoft.com>
Commit-Queue: Huiting Chen <huitingchen@microsoft.com>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: Sam Davis Omekara <samomekarajr@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1432136}
This CL updates the zoom system to enable PdfOopif PDFs to zoom
properly, i.e. to replicate the zooming behavior of PDFs rendered
with MimeHandlerViewGuest.
When a PDF is viewed, the viewer's UI should not be affected by
the default browser zoom, nor should it change size if the user
changes either the default browser zoom or the zoom of the page
displaying the PDF.
If the PDF is viewed in the mainframe, this involves (i) setting
the zoom mode of the main frame to manual (to prevent changes to
the page's zoom being persisted in the HostZoomMap under the url
of the PDF's content), and (ii) setting the scheme+host zoom level
for the PDF extension to be zoom factor 1.0 (zoom level 0). This
second step is what allows the UI to be rendered at a consistent
size regardless of the page's zoom.
If the PDF is is embedded in a page, only step (ii) is necessary,
as changes to the page's zoom will be persisted under the main
frame's URL.
In both cases, the PDF extension monitors the page's zoom level
so it can be applied to the content without applying it to the
viewer's UI.
There are two major changes involved in this CL:
1) The PdfViewerStreamManager is modified to a) set up a separate
ZoomController for the RenderFrameHost that contains the Pdf
extension (this will allow the PDF viewer's frame to keep its
zoom factor at a different level than the page zoom), and b) a
call to HostZoomMap to set the desired (independent) zoom factor
for the PDF extension frame at 1.0.
2) Modifications to WebContentsImpl::GetPendingZoomLevel() to allow
it to identify a RenderWidgetHost/RenderFrameHost that has
independent zoom, and which isn't a main frame. This frame hosts
the PDF extension.
Note that although with PdfOopif the PDF extension frame will create
its own ZoomController (1), any calls it makes to the TabsAPI zoom
functions still use the ZoomController for the main frame. This allows
it to prevent zoom levels for full-page PDFs from being persisted.
Note also that no PDF extension JS code has been modified.
Bug: 372932834
Change-Id: Iaf418d03227817ec18fe1623d1f77f93be483608
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6281436
Reviewed-by: Kevin McNee <mcnee@chromium.org>
Commit-Queue: James Maclean <wjmaclean@chromium.org>
Reviewed-by: Alex Moshchuk <alexmos@chromium.org>
Reviewed-by: Andy Phan <andyphan@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432129}
In https://codereview.chromium.org/48623006/ Skia added an optimization
for blurring rounded rectangles (treating them as a ninepatch to
do less blurring and make the results cacheable). This saw up to an
80% reduction in time for rendering with the CPU backend in Skia's
benchmarks. However, due to differences in rendering between this
fast path and the existing path, it was not landed in Chromium and sat
unused here for 11+ years. It's been used in Android and all other
clients since then.
I addressed those rendering diffs [1][2][3] and this CL makes Chromium
use that fast path. This required setting up Chromium's discardable
memory cache (which Skia uses to cache masks and other things) because
in the new code path, Skia caches those blurred masks to save time
if there are, for example, multiple blurry rrects being drawn with 5px
border and the same sigma values.
There are still some very minor differences in appearance between the
two implementations, which I've addressed by rebasing and adding some
slight fuzzy tolerances, to make the tests less brittle for future
blur changes.
[1] http://review.skia.org/953496
[2] http://review.skia.org/953516
[3] https://crrev.com/c/6328778
Change-Id: I140844d24a9e31cb2bde4e25b6cb1cce1b812f67
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6335820
Commit-Queue: Kaylee Lubick <kjlubick@chromium.org>
Reviewed-by: Benoit Lize <lizeb@chromium.org>
Reviewed-by: Nico Weber <thakis@chromium.org>
Reviewed-by: Ashley Prasad <ashleydp@google.com>
Cr-Commit-Position: refs/heads/main@{#1432101}
This change introduces a mechanism to track and reuse free and empty
renderer processes that have not yet been terminated, reducing the
overhead of creating new ones.
To achieve this:
- A new 'SiteProcessCountTracker, keyed by
'kEmptySiteProcessCountTrackerKey'. This tracker specifically monitors
processes that become free and empty.
- The 'RenderProcessHostImpl::RemoveFrameWithSite' method now adds a
process to this tracker when its last frame is removed, indicating it is
now free and empty.
- Conversely, 'RenderProcessHostImpl::AddFrameWithSite' removes a
process from the tracker if a frame is added to it.
When assigning a process to a new 'SiteInstance', the code prioritizes
reusing a suitable, free and empty process identified by this tracker
(currently only background hosts are considered). If found, the
'SiteInstance' is marked with `REUSED_EXISTING_PROCESS`. Finally,
'RenderProcessHostImpl' destruction clears the process from the tracker
via 'ClearProcessForAllSites()'.
Bug:365958798
Change-Id: Icc1fb48b75b78fe25f83d179f4989124e929785b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6217969
Reviewed-by: Alex Moshchuk <alexmos@chromium.org>
Commit-Queue: Alex Attar <aattar@google.com>
Cr-Commit-Position: refs/heads/main@{#1432081}
When ProgressiveAccessibility is configured with mode=disable_on_hide,
web accessibility is ordinariliy turned off for tabs when they are
hidden. This CL exempts known screen readers from this behavior.
Bug: 40925658
Change-Id: I791fff91a967d0e9464ce3f8e22c23225ffd8d18
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6347049
Commit-Queue: Avi Drissman <avi@chromium.org>
Reviewed-by: Aaron Leventhal <aleventhal@chromium.org>
Auto-Submit: Greg Thompson <grt@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432075}
client bounce
The UMA client bounce metrics were added in https://crrev.com/c/6012217,
and as indicated we're only interested in non-navigation tied
attribution data hosts.
We would like to measure the DWA metrics, i.e. reporting origin level
metrics, for these metrics as well. We only record metrics for the
reporting origins that initiate the data host registration, and
reporting origins registered in the redirect chain are not considered
for simplification as they're unknown during data host registration.
Therefore, reporting origins need to be sent over mojo when registering
data hosts.
Fixed: b/383579282
Change-Id: I4482d3134fa3aff085bab4a391f2836dbb7964ea
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6331143
Reviewed-by: Joe Mason <joenotcharles@google.com>
Commit-Queue: Nan Lin <linnan@chromium.org>
Reviewed-by: Dominic Farolino <dom@chromium.org>
Reviewed-by: Nate Chapin <japhet@chromium.org>
Reviewed-by: Robert Kaplow <rkaplow@chromium.org>
Reviewed-by: Andrew Paseltiner <apaseltiner@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432065}
If we turn on LNA warnings in M136, we need to keep the PNA
PrivateNetworkAccessRestrictionsEnabled enterprise policy working as per
agreement with enterprise team (can't break the enterprise policy until
M137).
This keeps the override in place for M136 even if we turn on the LNA
warnings in M136.
Bug: 395895368
Change-Id: Ia07912d551ed5f7ca9dfd2f20ef289563ec729dc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6344890
Commit-Queue: Hubert Chao <hchao@chromium.org>
Reviewed-by: Chris Thompson <cthomp@chromium.org>
Reviewed-by: Camille Lamy <clamy@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432048}
In http://crrev.com/c/2838627, pages not receiving BFCache ACK are not
counted when enforcing the limit, and the limit enforcement will be
triggered when receiving the ACK. However, during pruning (due to memory
pressure), those pages in the entries should be counted and probably
evicted. This CL adds the reason check and a browser test for the case.
Bug: 402313221
Change-Id: Idffce799b843462a3f4eee459b4dfbd9cea3c8f5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6347142
Reviewed-by: Fergal Daly <fergal@chromium.org>
Commit-Queue: Mingyu Lei <leimy@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1432012}
This reverts commit db6d48dfdd.
Reason for revert: Caused Linux UBSan test failure, e.g. https://chromium-swarm.appspot.com/task?id=6fa5b90aa914e711&o=true&w=true
Original change's description:
> [views-ax] Create BrowserViewsAXManager instance in the correct place
>
> Previously, I introduced BrowserViewsAXManager (behind a flag) and
> instantiated it from the ViewsDelegate constructor. This was incorrect,
> as we rely on the AXPlatform::GetInstance that only gets created a few
> moments later in the BrowserMainLoop. The design was invalid.
>
> The challenge here is that we must initialize the BrowserViewsAXManager
> class, which lives in the ui/views layer, from the browser_main_loop.
> Because content/browser cannot depend directly on ui/views, we must
> instead leverage the ChromeBrowserMainExtraPartViews to initialize the
> shared instance and store it.
>
> Bug: 40672441
> Change-Id: I62432f9e547172bbe357040db29428e41db7d82c
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6293824
> Reviewed-by: Robert Liao <robliao@chromium.org>
> Reviewed-by: Avi Drissman <avi@chromium.org>
> Reviewed-by: Dirk Pranke <dpranke@google.com>
> Auto-Submit: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
> Commit-Queue: Avi Drissman <avi@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#1431815}
Bug: 40672441
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Change-Id: I178c4f10b9e31c62cef3ec606ceca6f8d15a8dc8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6349100
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Owners-Override: Kenichi Ishibashi <bashi@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#1431927}
The old bug was about the SW static routing API itself.
It has been updated to the bug specific to this issue.
i.e. make fetches from SharedWorker and DedicatedWorker
to be handled by SW static routing API.
Bug: 402806160
Change-Id: Ib1addb5270ff91b022bc09739bcb9798b71fd7e5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6350282
Reviewed-by: Rakina Zata Amni <rakina@chromium.org>
Reviewed-by: Keita Suzuki <suzukikeita@chromium.org>
Commit-Queue: Yoshisato Yanagisawa <yyanagisawa@google.com>
Reviewed-by: Shunya Shishido <sisidovski@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1431926}
This reverts commit c6bfb8a155.
Reason for revert: This change broke Narrator's move-by-headings
feature. Heading navigation is a critical feature of screen readers that
facilites navigating web pages, like the google/bing results page. It
impacts all headings throughout the web, whereas the original CL was
fixing a scoped problem: spelling errors unannounced when exposed using
CSS highlights. We will also revert this change from 134.
Original change's description:
> [a11y] Fixes move by format for CSS spellcheck highlights
>
> Uses Spelling/Grammar error CSS highlight markers in the
> Accessibility Tree to break format boundaries around those
> errors to help navigate better around these spelling/grammar
> errors.
>
> This CL achieves following:
> 1) Fixes Move by format to accommodate CSS spellcheck highlights.
> 2) Narrator calls out the spelling issues correctly.
>
> Bug: 364795299
> Change-Id: I4842a49449c666100e31f859447f951e9f9eb09e
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6178990
> Auto-Submit: Vinay Singh <vinaysingh@microsoft.com>
> Reviewed-by: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
> Commit-Queue: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
> Cr-Commit-Position: refs/heads/main@{#1415710}
Bug: 402800783, 364795299
Change-Id: I1ee7bd394f18fba21e97e5cc8998de1a03501931
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6349562
Reviewed-by: Jacques Newman <janewman@microsoft.com>
Auto-Submit: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
Commit-Queue: Benjamin Beaudry <benjamin.beaudry@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1431891}