Upgrade from Chromium 92 to Chromium 93 #9062

…0.4530.5.

It's chrome/browser/sync/sync_service_factory.cc now. Chromium change: https://chromium.googlesource.com/chromium/src/+/ff4e5f6e2d335450d95d29be8496d2f65471a3d2 commit ff4e5f6e2d335450d95d29be8496d2f65471a3d2 Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Tue Jun 1 20:12:14 2021 +0000 Rename chrome/'s ProfileSyncServiceFactory to SyncServiceFactory The goal of this factory is to return a SyncService. This follows the naming convention of other keyed services that have an abstract interface. Bug: 1201272

It's ios/chrome/browser/sync/sync_service_factory.cc now. Chromium change: https://chromium.googlesource.com/chromium/src/+/c17aab02c3bc7cb9732708126a5001d15ad6efd3 commit c17aab02c3bc7cb9732708126a5001d15ad6efd3 Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Tue Jun 1 18:44:24 2021 +0000 Rename ios/'s ProfileSyncServiceFactory-s to SyncServiceFactory Same as crrev.com/c/2928734 but for the 2 factories in ios/: ProfileSyncServiceFactory and WebViewProfileSyncServiceFactory. Bug: 2928734

Need to patch the right .ts file now. Chromium change: https://chromium.googlesource.com/chromium/src/+/683b64f39d5890d0938bad29ecd768b8fb6a6aeb commit 683b64f39d5890d0938bad29ecd768b8fb6a6aeb Author: dpapad <dpapad@chromium.org> Date: Thu May 27 12:01:17 2021 +0000 Downloads: Migrate to TypeScript. Leveraging the "allowJs" option of the Typescript compiler, to consume Mojo JS bindings (downloads.mojom-webui.js), until Mojo can generate TypeScript code directly. Bug: 1189595

Need to patch the right .ts files now. Chromium change: https://chromium.googlesource.com/chromium/src/+/f3907c07a8a0c13b1b513f507719fe3e316d8c80 commit f3907c07a8a0c13b1b513f507719fe3e316d8c80 Author: dpapad <dpapad@chromium.org> Date: Wed May 26 17:41:25 2021 +0000 History: Migrate to TypeScript. - Remove js_library(), js_type_check() targets. - Add new ts_library() target, and hook it up to the build. - Migrate all files to TypeScript, and fix all errors. - Override default tsconfig with a few options that seem necessary for Polymer code. - Introduce EventTracker usage in app.ts as part of the migration (easier than adding type info for boundOnKeyDown_) - Rename 'open-menu' event to 'synced-device-card-open-menu' in synced_device_card.ts, to be able to differentiate with the 'open-menu' event in history_card.ts, when declaring the Event type in the global HTMLElementEventMap. Bug: 1189595

Chromium change: https://chromium.googlesource.com/chromium/src/+/c0f9fc515c962497b90c8c8fa3b72e77b8a88647 commit c0f9fc515c962497b90c8c8fa3b72e77b8a88647 Author: Lei Zhang <thestig@chromium.org> Date: Sat May 22 08:00:53 2021 +0000 Remove base/containers/contain.h include from base/stl_util.h. This include is temporary scaffolding to help move base::Contains() to its new location. Now that most callers to base::Contains() include base/containers/contain.h, remove it from base/stl_util.h and fix the remaining callers that need the correct include. Bug: 970209

A different GetNativeTheme() symbol is now used in focus_ring.cc, so we can't just redefine it as we used to do. Chromium change: https://chromium.googlesource.com/chromium/src/+/b62ab23c8d294fceda02edad378327cc787292e2 commit b62ab23c8d294fceda02edad378327cc787292e2 Author: Peter Boström <pbos@chromium.org> Date: Tue May 25 23:29:28 2021 +0000 Add scoped overriding of FocusRing colors This fixes focus-ring contrast when using themes for: * The omnibox. * Page-action icons inside the omnibox. * The tab strip. * New-tab button + tab-search button (buttons outside the tab strip). * Infobars. * Download shelf. This is done without individually fixing all buttons/views/subclasses in those view subhierarchies. It's also able to stay correct for any new subclasses added to either of those view hierarchies. This also adds contrast boosting for platforms where the default FocusRing colors are not fully opaque. Bug: 1068929

Chromium change: https://chromium.googlesource.com/chromium/src/+/62060fb114a2f54030c2c5200982eed77fa52d2c commit 62060fb114a2f54030c2c5200982eed77fa52d2c Author: David Roger <droger@chromium.org> Date: Tue May 25 23:22:41 2021 +0000 [profiles] Delete flag for new profile picker The new profile picker is now enabled by default, remove the flag to disable it. Bug: 1063856

This feature is only enabled/disabled via finch trials now, which we disable via a chromium_src override to origin_trials.cc anyway, so we can safely drop all references to this no longer existing parameter. Chromium change: https://chromium.googlesource.com/chromium/src/+/428b9ab9c67f504a745905974d1a67cdd8739706 commit 428b9ab9c67f504a745905974d1a67cdd8739706 Author: Glen Robertson <glenrob@chromium.org> Date: Tue Jun 1 10:07:42 2021 +0000 Cleanup the kHandwritingRecognitionWebPlatformApi flag. The flag was originally needed to restrict the availability of the API to specific hardware devices that could support it. The API is now available on all devices and the flag is now enabled by default. The flag is not needed even to turn off the API via Finch, as we have a separate flag for that (kHandwritingRecognitionWebPlatformApiFinch). Bug: 1166910

The new profile menu/picker design is now the only one available, so there's no longer any need to deal with this parameter. Chromium change: https://chromium.googlesource.com/chromium/src/+/62060fb114a2f54030c2c5200982eed77fa52d2c commit 62060fb114a2f54030c2c5200982eed77fa52d2c Author: David Roger <droger@chromium.org> Date: Tue May 25 23:22:41 2021 +0000 [profiles] Delete flag for new profile picker The new profile picker is now enabled by default, remove the flag to disable it. Bug: 1063856

DictionaryValue::empty() has been removed upstream. Chromium change: https://chromium.googlesource.com/chromium/src/+/c276866bb1276cf9d39604921e421fb92ac44ad9 commit c276866bb1276cf9d39604921e421fb92ac44ad9 Author: Song Fangzhen <songfangzhen@bytedance.com> Date: Wed May 26 08:32:56 2021 +0000 CodeHealth: Remove DictionaryValue::empty After removing the use of DictionaryValue::empty(), now DictionaryValue::empty() itself should be removed. Bug: 1187020

A new |kWorkspaceDesk| value was added to the ModelTypeForHistograms class enumeration in components/sync/base/model_type.h. Chromium change: https://chromium.googlesource.com/chromium/src/+/f5d50f5d1c2009fb570aa1e4a41dd773c86f425a commit f5d50f5d1c2009fb570aa1e4a41dd773c86f425a Author: Yanzhu Du <yzd@google.com> Date: Fri May 21 16:19:52 2021 +0000 Add workspace_desk field to Chrome Sync EntitySpecifics Bug: 1177413

Define an specialization of content::GetServiceSandboxType<T>() for the bat_ads::mojom::BatAdsService mojo interface that returns the default sandbox type that was being used until now (i.e |kUtility|). Chromium change: https://chromium.googlesource.com/chromium/src/+/4aac4e9b3b64c634dcdd935e989dddaf19df43e3 commit 4aac4e9b3b64c634dcdd935e989dddaf19df43e3 Author: Alex Gough <ajgo@chromium.org> Date: Wed May 26 01:46:47 2021 +0000 Remove default sandbox type for services Deletes content::GetServiceSandboxType<>() (see content/public/browser/service_process_host.h) so that ::Launch cannot compile if a sandbox is not specified by a specialized template. All existing services without a specified sandbox are launched as kUtility, so these now have that specified. No functional changes. Template for recording::mojom::RecordingService moved to chromeos and includes adjusted. Bug: 1210301

A small change upstream in the order in which command line parameters are propagated from the browser process to renderers means we can't inspect the renderer's parameteres in AppendExtraCommandLineSwitches() to search for switches::kTestType, but the browser process' parameters. Otherwise the seed used to instantiate the crypto::HMAC for farbling purposes will be non-deterministic for tests, causing test failures. More specifically, this change fixes the following 12 browser tests: * BraveContentSettingsAgentImplBrowserTest.FarbleGetImageData * BraveContentSettingsAgentImplV2BrowserTest.FarbleGetImageData * BraveDeviceMemoryFarblingBrowserTest.FarbleDeviceMemory * BraveNavigatorHardwareConcurrencyFarblingBrowserTest.FarbleNavigatorHardwareConcurrency * BraveNavigatorPluginsFarblingBrowserTest.FarbleNavigatorPlugins * BraveNavigatorPluginsFarblingBrowserTest.FarbleNavigatorPluginsBuiltin * BraveNavigatorPluginsFarblingBrowserTest.FarbleNavigatorPluginsReset * BraveNavigatorUserAgentFarblingBrowserTest.FarbleNavigatorUserAgent * BraveOffscreenCanvasFarblingBrowserTest.FarbleGetImageData * BraveWebAudioFarblingBrowserTest.FarbleWebAudio * BraveWebGLFarblingBrowserTest.FarbleGetParameterWebGL * BraveWebGLFarblingBrowserTest.FarbleGetParameterWebGL2 Chromium change: https://chromium.googlesource.com/chromium/src/+/f453a14bc7051fd3b90ab21fc460cba1fa9fd703 commit f453a14bc7051fd3b90ab21fc460cba1fa9fd703 Author: Bruce Dawson <brucedawson@chromium.org> Date: Fri May 21 15:36:33 2021 +0000 Put --extension-process earlier on the command line The process type (--type, --utility-sub-type, and --extension-process) flags are useful for human beings looking at chrome command lines. Therefore they should be early, before the non-human-readable switches. Moving them early also avoids problems with them being trimmed by tools (such as ETW tracing) that have limits on the lengths of the command lines that they can record. This change will ensure that UIforETW can reliably identify extension processes and will aid human readers also. Bug: 1211558

Upstream now does single cookie-specific checks in some places to determine whether cookie access should be granted. However, when ephemeral storage is enabled, Brave shouldn't care about whether access is being requested for a specific cookie or not, so we simply return |true| if that's the case. This change fixes the following 13 browser tests: * BraveContentSettingsAgentImplBrowserTest.BlockThirdPartyCookieByDefault * BraveContentSettingsAgentImplBrowserTest.ChromiumCookieBlockOverridesBraveAllowCookiesIframe * BraveContentSettingsAgentImplBrowserTest.ExplicitBlock3PCookies * EphemeralStorageBrowserTest.ClosingTabClearsEphemeralStorage * EphemeralStorageBrowserTest.EphemeralStorageDoesNotLeakBetweenProfiles * EphemeralStorageBrowserTest.FirstPartyNestedInThirdParty * EphemeralStorageBrowserTest.NavigatingClearsEphemeralStorageAfterKeepAlive * EphemeralStorageBrowserTest.ReloadDoesNotClearEphemeralStorage * EphemeralStorageBrowserTest.StorageIsPartitioned * EphemeralStorageKeepAliveDisabledBrowserTest.NavigatingClearsEphemeralStorageWhenKeepAliveDisabled * EphemeralStorageTest.CrossSiteCookiesBlockedInitial * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageDifferentSession * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageSameSession Last, this CL also removes a patch for cookie_settings_base.h that can be avoided entirely via a #define in the chromium_src override file. Chromium change: https://chromium.googlesource.com/chromium/src/+/a50819a49ecf91b62ec9958355915e972eb89abd commit a50819a49ecf91b62ec9958355915e972eb89abd Author: cfredric <cfredric@chromium.org> Date: Tue Jun 1 18:34:53 2021 +0000 Rewrite callsites of IsCookieAccessAllowed that care about a single cookie. Rewriting these callsites allows us to eventually replace the implementation with one that takes per-cookie details into account when determining whether the cookie is accessible re: the user's cookie-blocking settings. Bug: 1203706

In the past, the call to CookieSettings::IsCookieAccessible() run from NetworkServiceNetworkDelegate::OnCanSetCookie() would handle ephemeral storage via the redefinition we have in the chromium_src override for CookieSettingsBase::IsCookieAccessAllowed(). However, CookieSettings::IsCookieAccessible() does not longer call now CookieSettingsBase::IsCookieAccessAllowed() but the recently-added function CookieSettings::IsHypotheticalCookieAllowed(), which follows a different logic and does not hit our chromium_src overrides. We need therefore to make sure that NetworkServiceNetworkDelegate does call our wrapper IsEphemeralCookieAccessible() method instead of IsCookieAccessible(), so that Ephemeral storage is accounted for properly, falling back to IsHypotheticalCookieAllowed() if needed. Chromium change: https://chromium.googlesource.com/chromium/src/+/aba4a46c62f5f3ec198e7b7bd8ac1bbd6c953d1d commit aba4a46c62f5f3ec198e7b7bd8ac1bbd6c953d1d Author: cfredric <cfredric@chromium.org> Date: Tue Jun 22 18:29:13 2021 +0000 Add new --sameparty-considered-first-party flag and logic. This flag will resolve cfredric/sameparty#4, and will allow developers to test the eventual interaction of SameParty cookies and third-party-cookie-blocking. It causes SameParty cookies to not be blocked by the user's third-party-cookie-blocking settings, as they are considered first-party. Bug: 1203706

The overrides we have in place redefine IsCookieAccessAllowed() but now we need to override IsFullCookieAccessAllowed() instead to achieve the same behaviour, plus include a couple of extra overrides while upstream hasn't completely migrated to IsFullCookieAccessAllowed(). This change fixes the following 3 browser tests: * BraveNetworkDelegateBrowserTest.ThirdPartyCookiesBlockedNestedFirstPartyIframe * BraveNetworkDelegateBrowserTest.ThirdPartyYesCookiesWpComInWordpressCom * BraveNetworkDelegateBrowserTest.ThirdPartyYesCookiesnWordpressComInWpCom Chromium change: https://chromium.googlesource.com/chromium/src/+/160c6fb00757dac09df8545ef7a2ae3bc15ea162 commit 160c6fb00757dac09df8545ef7a2ae3bc15ea162 Author: cfredric <cfredric@chromium.org> Date: Tue Jun 1 16:14:59 2021 +0000 Introduce IsFullCookieAccessAllowed and IsPrivacyModeEnabled. IsCookieAccessAllowed is being renamed to more closely fit its semantics, since there will soon be scenarios in which some cookies are sent, but IsCookieAccessAllowed would return false. IsPrivacyModeEnabled is being introduced because this caller is really concerned about *any* cookie being accessible, rather than having full access to cookies, so IsFullCookieAccessAllowed would give the wrong result in cases when some cookies are accessible and others are not. In particular, in the future we intend to support sending/setting SameParty cookies (in same-party contexts) even when third-party cookies are blocked. In such a situation, we do not want to enable privacy mode even though third-party cookies are blocked, since that would prevent SameParty cookies from being sent/set. This CL rewrites the following IsCookieAccessAllowed callsites: * PrivacySandboxSettings::IsFledgeAllowed * PrivacySandboxSettings::FilterFledgeAllowedParties * PrivacySandboxSettings::IsPrivacySandboxAllowedForContext * DurableStoragePermissionContext::DecidePermission * CookieControlsController::FirstPartyCookiesBlocked * signin::SettingsAllowSigninCookies * NetworkServiceNetworkDelegate::OnForcePrivacyMode * NetworkServiceNetworkDelegate::OnCanQueueReportingReport * NetworkServiceNetworkDelegate::OnCanSetReportingClient * NetworkServiceNetworkDelegate::OnCanUseReportingClient More callsites will be handled in subsequent CLs. Bug: 1203706

…4535.3.

They are in components/sync/driver/sync_service_impl.{cc,h} now. Chromium change: https://chromium.googlesource.com/chromium/src/+/b77854267a62e9208b1ac5ceb4167387ec7c56cc commit b77854267a62e9208b1ac5ceb4167387ec7c56cc Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Fri Jun 4 09:15:21 2021 +0000 Reland "Rename ProfileSyncService to SyncServiceImpl" Reason for reland: Original CL was reverted to unwind other changes. Original change's description: > Rename ProfileSyncService to SyncServiceImpl > > Renaming related classes like ProfileSyncServiceAndroid and > ProfileSyncServiceHarness is left for next CLs. We also use this CL as > an opportunity to hide the implementation class when possible: > - Some includes for sync_service_impl.h are removed if unnecessary, or > replaced with sync_service.h. > - Even comments referring to SyncServiceImpl are updated to refer to > the interface if the layer isn't aware SyncServiceImpl is being used. > > The CL also adds one missing include for <string> in > fake_server_helper_android.cc to address one IWYU presubmit warning. Bug: 1201272

Chromium change: https://chromium.googlesource.com/chromium/src/+/bdac63b4caf1f2eaaecdcd4fa4058755e704537b commit bdac63b4caf1f2eaaecdcd4fa4058755e704537b Author: rbpotter <rbpotter@chromium.org> Date: Thu Jun 3 16:19:08 2021 +0000 Reland Bookmarks: Update to TypeScript Bug: 1189595

Chromium change: https://chromium.googlesource.com/chromium/src/+/27f5525d7368694e0d049975e1fef0049bab1a2e commit 27f5525d7368694e0d049975e1fef0049bab1a2e Author: dpapad <dpapad@chromium.org> Date: Thu Jun 3 18:58:55 2021 +0000 Extensions: Migrate to TypeScript, part1. In this part migrating non Polymer files (11/40 total files). Also adding minimal definition files for chrome.management and chrome.runtime.lastError APIs. Remaining files will be addressed in follow up CL(s). Bug: 1189595

They have moved into the ntp_tiles:: namespace now. Chromium change: https://chromium.googlesource.com/chromium/src/+/b0e2a7e1f4ef8634dc4bc417fdce2f450e6b0315 commit b0e2a7e1f4ef8634dc4bc417fdce2f450e6b0315 Author: Moe Ahmadi <mahmadi@chromium.org> Date: Wed Jun 2 22:35:34 2021 +0000 [MV] Moves MV settings into MostVisitedSites Currently MV settings/preferences are registered and controlled by the InstantService which updates the MostVisitedSites when those preferences change in the NTP. MostVisitedSites has a partial knowledge of whether custom links are enabled via a state it keeps. In order to eliminate the dependency of the 1P NTP to the InstantService this change moves those prefs into the MostVisitedSites. In the future, this will allow the NTP to observe and set those preferences directly in MostVisitedSites without needing the InstantService as an intermediary. This change is not expected to make any changes to the logic around setting or using these preferences. Bug: 1084363

We should use Value::FindIntKey() now. Chromium changes: https://chromium.googlesource.com/chromium/src/+/0284432657804a07a1493bdb025b111851962539 https://chromium.googlesource.com/chromium/src/+/146e9cdbf128d29b0cf72d98659d07c9e1e59df7 commit 0284432657804a07a1493bdb025b111851962539 Author: Song Fangzhen <songfangzhen@bytedance.com> Date: Wed Jun 2 16:28:50 2021 +0000 CodeHealth: Remove use of DictionaryValue::GetIntegerWithoutPathExpansion This change removes the use of DictionaryValue::GetIntegerWithoutPathExpansion and replaces with Value::FindIntKey() Bug: 1187044 commit 146e9cdbf128d29b0cf72d98659d07c9e1e59df7 Author: Song Fangzhen <songfangzhen@bytedance.com> Date: Fri Jun 4 15:26:07 2021 +0000 Remove DictionaryValue::GetIntegerWithoutPathExpansion After removing the use of DictionaryValue::GetIntegerWithoutPathExpansion, now DictionaryValue::GetIntegerWithoutPathExpansion itself should be removed. Bug: 1187044

BraveP3ALogStore inherits from LogStore, so we need to override this newly added method. Chromium change: https://chromium.googlesource.com/chromium/src/+/d7d5fdc161ccfa3fe292dca639a2a0b16b8709de commit d7d5fdc161ccfa3fe292dca639a2a0b16b8709de Author: Jonghyun Ahn <jongahn@google.com> Date: Fri Jun 4 00:15:08 2021 +0000 Reporting service checks for user id and corresponding user consent before uploading. User ID added to log metadata so log doesn't need to be de-compressed to observe user ID. Bug: 1181504

Chromium change: https://chromium.googlesource.com/chromium/src/+/c126d64cab6443252034b1bd4b52e47ed1c50bbb commit c126d64cab6443252034b1bd4b52e47ed1c50bbb Author: Anastasiia N <anastasiian@chromium.org> Date: Wed Jun 2 14:11:49 2021 +0000 CodeHealth: Remove ListValue::AppendStrings() Bug: 1187103

This is already enabled by default and there's no need to explicitly enable the feature for testing purposes. Chromium changes: https://chromium.googlesource.com/chromium/src/+/e2124a1d261f70ffe0048553b0f4642e0444d6a1 https://chromium.googlesource.com/chromium/src/+/83e7d72c507e71db60900a59923e90fd3c4644ee commit e2124a1d261f70ffe0048553b0f4642e0444d6a1 Author: Caroline Rising <corising@chromium.org> Date: Fri Mar 5 19:30:23 2021 +0000 Clean up tab hover card feature flag and references. This feature has been launched on all platforms since M87. Bug: 910739 commit 83e7d72c507e71db60900a59923e90fd3c4644ee Author: Peter Boström <pbos@chromium.org> Date: Wed Jun 2 16:43:11 2021 +0000 Remove stale kTabHoverCards feature flag This experiment was cleaned up and is no longer in use. Bug: 910739

Chromium change: https://chromium.googlesource.com/chromium/src/+/d8ef96e6b8d1de5c4441c8b7df1dab8ef9142290 Remove ProfileSyncServiceAndroid::Init() This method is basically a way to check whether the native ProfileSyncService is null. If so, JNI_ProfileSyncService_Init() destroys ProfileSyncServiceAndroid immediately. Instead, let's check first and then construct the object if PSS is non-null. ProfileSyncServiceAndroid now only has 2 deps (the native and java services), both set on construction. The CL also does some minor readability improvements to profile_sync_service_android.cc. Bug: None

Chromium change: https://chromium.googlesource.com/chromium/src/+/e69282129b23e48d40c953fa7d388e0ebf8c7a5a [GURL] Remove references to Tab.getUrlString() ~50 left Bug: 783819

We previously only had to replace calls to IsCookieAccessAllowed() and IsCookieAccessible() with calls to IsEphemeralCookieAccessAllowed(), but we now need to also include IsFullCookieAccessAllowed() or we will end up wrongly blocking 3P cookies when Ephemeral storage is enabled. This change fixes the following 6 browser tests: * EphemeralStorageTest.CrossSiteCookiesBlockedInitial * EphemeralStorageTest.CrossSiteCookiesBlockedNewPageResetSession * EphemeralStorageTest.CrossSiteCookiesBlockedRemotePageNewSession * EphemeralStorageTest.CrossSiteCookiesBlockedRemotePageSameSession * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageDifferentSession * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageSameSession Chromium change: https://chromium.googlesource.com/chromium/src/+/32dfa4130cd0729f4dd39a59837582e8ff495b9c commit 32dfa4130cd0729f4dd39a59837582e8ff495b9c Author: cfredric <cfredric@chromium.org> Date: Thu Jun 3 14:39:42 2021 +0000 Rename remaining non-read IsCookieAccessAllowed callsites. IsCookieAccessAllowed will be deleted after the remaining "read" callsites have been rewritten in a subsequent CL. (In this context, by "read", I'm referring to callsites that actually read cookies. Some callsites of IsCookieAccessAllowed read cookies, some set cookies, and some do neither.) Bug: 1203706

…4542.2.

Chromium change: https://chromium.googlesource.com/chromium/src/+/8f7c30fadd05304173e12a997e9ed0450ab1aded commit 8f7c30fadd05304173e12a997e9ed0450ab1aded Author: dpapad <dpapad@chromium.org> Date: Mon Jun 7 16:10:48 2021 +0000 Extensions: Migrate to TypeScript, part3. Migrating all remaining files except for the activity_log/ folder, which will be migrated in the next (and last) part. Bug: 1189595

It moved from //components/safe_browsing/core/browser to //components/safe_browsing/content/browser, so we need to adapt the patches and the chromium_src override. Chromium change: https://chromium.googlesource.com/chromium/src/+/3cacd06c63f0c2506f375bf3597d22110efe9647 commit 3cacd06c63f0c2506f375bf3597d22110efe9647 Author: Colin Blundell <blundell@chromium.org> Date: Wed Jun 9 10:11:18 2021 +0000 [Safe Browsing] Move SafeBrowsingNetworkContext from core/ to content/ This CL moves safe_browsing_network_context.*, which depends on //content/public, from its current architecturally-incorrect //components/safe_browsing/core/browser location to //components/safe_browsing/content/browser. Bug: 1216341

Files live under //third_party/blink/renderer/platform/media now. Chromium change: https://chromium.googlesource.com/chromium/src/+/b56c921dc8c64ae3d7b5a3a04b168e7773dc77d6 commit b56c921dc8c64ae3d7b5a3a04b168e7773dc77d6 Author: Henrique Ferreiro <hferreiro@igalia.com> Date: Wed Jun 9 17:53:59 2021 +0000 Move media/blink/ to Blink media/blink/ files are being moved to public/platform/media/ and renderer/platform/media/, depending on the private/public status of the file as mentioned in this table: https://docs.google.com/document/d/1p0OuzArDNZHT_aOm75oOWh3qR-mT6Zq-qKxxKYT-jYI/edit#bookmark=kix.2v0uvunkva0g The move includes renaming files as described in the first table at https://docs.google.com/document/d/1p0OuzArDNZHT_aOm75oOWh3qR-mT6Zq-qKxxKYT-jYI/edit#heading=h.jqrzzxjh40df where underscores are added to be consistent with the camel-case naming of the classes defined in the renamed files. //media/blink:media_blink_unittests is split with all files but one being moved to blink_platform_unittests. web_media_player_impl_unittest.cc uses WebView, so it requires a Platform object and, as such, should be moved into //third_party/blink/renderer/platform:unit_tests which is part of blink_unittests. This makes the android bots fail, as they find no tests to run. Because of this, this CL leaves this test in place, which will be removed in a follow-up, together with all of media/blink/. Only the header guards and the export macros are updated. Namespace changes and other Blink coding conventions (GURL => KURL, Blink mojom includes, etc.) are left for follow-ups. Because of the latter, this CL fails the Blink-variant mojom presubmit check. No-Presubmit: true Bug: 1198341

One of the two blocks our chromium_src override was modifying has now moved up in the file. Chromium change: https://chromium.googlesource.com/chromium/src/+/85d65333b8b9a73c7e2cc3f0db6ec47c9513d418 commit 85d65333b8b9a73c7e2cc3f0db6ec47c9513d418 Author: Peter Boström <pbos@chromium.org> Date: Tue Jun 8 19:34:10 2021 +0000 Paint FocusRing aura with same path as its ring This fixes a rendering bug where the focus aura is incorrectly offset from the focus ring (probably by not being translated to the same coordinate space?). Bug: 1213390, 1210419

There used to be only one definition, but now there are two. Chromium change: https://chromium.googlesource.com/chromium/src/+/98b30255228997bc8cf30f8bfbdbb1fcbd7d5ed5 commit 98b30255228997bc8cf30f8bfbdbb1fcbd7d5ed5 Author: Christopher Cameron <ccameron@chromium.org> Date: Thu Jun 10 19:03:03 2021 +0000 CanvasColorManagement: Move back behind a flag I enabled CanvasColorManagement by un-guarding the stable parts of the spec. This was not the right way to do things. Change the still-very-experimental parts of the spec to be behind a CanvasColorManagementV2 flag, and move the parts intended for shipping behind a CanvasColorManagement flag. This patch is intended to be a no-op, so leave CanvasColorManagement as "stable" to accomplish this. A follow-on patch will change the label to experimental and update other settings that will not be no-ops. Bug: 1218419

We need to adapt the chromium_src overrides for this new method. Chromium change: https://chromium.googlesource.com/chromium/src/+/98b30255228997bc8cf30f8bfbdbb1fcbd7d5ed5 commit 98b30255228997bc8cf30f8bfbdbb1fcbd7d5ed5 Author: Christopher Cameron <ccameron@chromium.org> Date: Thu Jun 10 19:03:03 2021 +0000 CanvasColorManagement: Move back behind a flag I enabled CanvasColorManagement by un-guarding the stable parts of the spec. This was not the right way to do things. Change the still-very-experimental parts of the spec to be behind a CanvasColorManagementV2 flag, and move the parts intended for shipping behind a CanvasColorManagement flag. This patch is intended to be a no-op, so leave CanvasColorManagement as "stable" to accomplish this. A follow-on patch will change the label to experimental and update other settings that will not be no-ops. Bug: 1218419

The PopulateInterstitialStrings() prototype's signature has changed to expect a base::Value instead of a base::DictionaryValue now. Chromium change: https://chromium.googlesource.com/chromium/src/+/6093cddab725b5ba15d790d2b45a640cbb93549d commit 6093cddab725b5ba15d790d2b45a640cbb93549d Author: Sylvain Defresne <sdefresne@chromium.org> Date: Wed Jun 9 11:45:12 2021 +0000 Convert //components/security_interstitials to use base::Value The base::DictionaryValue class is deprecated in favor of simply using base::Value (and the API that internally check the stored value to be a dictionary). Bug: 1187061

We should now use a class method instead of an instance one. Chromium change: https://chromium.googlesource.com/chromium/src/+/a61eda05062d20ec0f592415f3a145cdf1dce095 commit a61eda05062d20ec0f592415f3a145cdf1dce095 Author: Peter Boström <pbos@chromium.org> Date: Mon Jun 7 18:20:22 2021 +0000 Make InkDrop a View class property Removes views::Button::ink_drop() in favor of views::InkDrop::Get(view), which works equally for all Views (not just Buttons). The InkDrop is removed in the destructor for all classes that override Add/RemoveLayerBeneathView() to prevent using the wrong overrides during ~View() when class properties destruct. This also removes the `destroying_` bool from InkDropHost so that layers can be properly removed on InkDropHost destruction, as the View may still live after this point on. Bug: 931964

In this particular case we should use the new SplitOnceCallback utility, so that there's guaranteed that only one of them is called at runtime. Chromium change: https://chromium.googlesource.com/chromium/src/+/c7069e96999fe58a182ea4d339b7ccec9c67dfef commit c7069e96999fe58a182ea4d339b7ccec9c67dfef Author: danakj <danakj@chromium.org> Date: Tue May 18 18:35:55 2021 +0000 Replace AdaptCallbackForRepeating with SplitOnceCallback in chrome/browser chrome/browser/profiles/ chrome/browser/sync_file_system/ chrome/browser/ui/ SplitOnceCallback keeps the types correct, with 2 OnceCallbacks, and ensures only one of the two is called at runtime. Also changes KeyedServiceShutdownNotifier to hold OnceCallbacks since they are called once, at shutdown. R=avi@chromium.org Bug: 730593 Change-Id: Ieac72bab8133233adae6dee8df1358051a133272 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2894709 Reviewed-by: Colin Blundell <blundell@chromium.org> Reviewed-by: Avi Drissman <avi@chromium.org> Commit-Queue: danakj <danakj@chromium.org> Cr-Commit-Position: refs/heads/master@{#884027} commit 06e5791cb03d31285675f45cbb4dcae06b94ef92 Author: Joshua Bell <jsbell@chromium.org> Date: Tue Jun 8 17:24:15 2021 +0000 Remove base::AdaptCallbackForRepeating All usages have been removed from the codebase. \o/ * Remove implementation/tests from base/ * Remove clang rewriter from tools/ Bug: 730593

Chromium change: https://chromium.googlesource.com/chromium/src/+/eed724832ecc7ebb808749dca4be4d3dd948bd44 commit eed724832ecc7ebb808749dca4be4d3dd948bd44 Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Tue Jun 8 07:34:43 2021 +0000 Rename last references to ProfileSyncService in components/sync/ No behavior is changed. This CL does the following renamings: - ProfileSyncServiceBundle -> SyncServiceImplBundle - profile_sync_service_startup_unittest.cc -> sync_service_impl_startup_unittest.cc R=jkrcal@chromium.org Bug: 1201272

We also need to specify a CreateMode now, which should be asynchronous by default. Chromium change: https://chromium.googlesource.com/chromium/src/+/9750cf2e2a24cf7c0dd006ef39c24a29e7f73556 commit 9750cf2e2a24cf7c0dd006ef39c24a29e7f73556 Author: Nicolas Ouellet-Payeur <nicolaso@chromium.org> Date: Mon Jun 7 18:36:21 2021 +0000 Allow ScopedProfileKeepAlive during Profile construction CL crrev.com/c/2867585 creates a ScopedProfileKeepAlive during Profile construction, before the Profile is registered with the ProfileManager. This caused issues in tests, where Profile* is constructed synchronously. Store keepalive states in a separate map, keyed by Profile* pointer. This way we can already increment/decrement the counters, even before the Profile* is added to `profiles_info'. Bug: 1209192

Instead of the now removed CommandIdToTabVectorIndex() function. Chromium change: https://chromium.googlesource.com/chromium/src/+/bc0b2f2f6f640641b9f4ea299b851c79704a63c1 commit bc0b2f2f6f640641b9f4ea299b851c79704a63c1 Author: Emily Shack <emshack@chromium.org> Date: Tue Jun 8 22:31:21 2021 +0000 Allow for unbounded app menu history submenus * Modifies the command id structure used by RecentTabsSubMenuModel to avoid bounded id buckets. Instead, shares the unbounded ids held by the bookmarks menu by each using every other id. This allows for the window and tab group submenus within the app menu history to display all tabs within, regardless of length. * Also addresses any clang-tiny warnings in the touched files. * There is a known bug in RecentTabsSubMenuModel::GetLabelFontListAt due to it only taking in an index, rather than model and index. This causes the second menu item in a group/window submenu to be bold, and causes a crash if group/window submenus are longer than the parent menu. Will address in a follow-up, as the crash trigger is behind a feature flag anyways and this CL is already on the larger side. Bug: 1205646

The signature of ChromeContentBrowserClient::HandleExternalProtocol() has changed the type of its second parameter. Chromium change: https://chromium.googlesource.com/chromium/src/+/111cdbb5ef8c56fc8add40ad3cb805e32051b066 commit 111cdbb5ef8c56fc8add40ad3cb805e32051b066 Author: Jeremy Roman <jbroman@chromium.org> Date: Thu Jun 10 19:42:52 2021 +0000 Use WebContents::Getter instead of RVH routing ID in external protocol handler. The caller has a getter that looks it up by the FrameTreeNode ID. The RenderViewHost process and routing ID can change on navigation (if the navigation ends up changing the renderer hosting the main frame). This is more concise and consistent and, additionally, removes WebContents::GetRenderViewHost calls which are potentially problematic for MPArch (supporting bfcache, prerendering and portals evolution).

Previously, the code generator used for v8_storage_estimate.h would not include the definition of the StorageEstimate class in the generated code, forcing us to include another generated file in the list of includes where it was defined: //t_p/b/r/m/quota/storage_estimate.h. However, recent changes in Chromium has changed this and now the StorageEstimate class is defined in v8_storage_estimate.h already, so we need to drop that other include from Brave-specific code. Chromium changes: https://chromium.googlesource.com/chromium/src/+/64e3501dc66a1465a5784e62deb2a32539d0d26c https://chromium.googlesource.com/chromium/src/+/4dec89481af5e69832b8a4c2f227e2ccd3da2f76 commit 64e3501dc66a1465a5784e62deb2a32539d0d26c Author: Yuki Shiino <yukishiino@chromium.org> Date: Mon Jun 7 12:11:12 2021 +0000 bind-gen: Implement the new IDL dictionary (2nd ver) Implements the new implementation of the IDL dictionaries. The 2nd ver implementations are based on: 1) Reuse the techniques used to implement IDL unions. IDL unions are class objects and similar to IDL dictionaries. So the same implementation techniques can be applied to IDL dictionaries. 2) ToV8 was replaced with ToV8Traits<IDLType>::ToV8, so the implementation of this patch is based on ToV8Traits. This patch is yet behind a flag, and doesn't affect prod. Bug: 839389 commit 4dec89481af5e69832b8a4c2f227e2ccd3da2f76 Author: Yuki Shiino <yukishiino@chromium.org> Date: Thu Jun 10 07:47:36 2021 +0000 bind-gen: Fix IDL enumeration implementation in two points Fixes the following two issues in the IDL enumeration implementation. 1. The default constructor was publicly accessible. This can be a cause of abuse of the construction in the invalid state (Empty state). Moves the default constructor to private:. 2. The string table was defined in *.h as constexpr, which allows compilers to make a copy of the string table in each translation unit. Moves the definition of the string table to *.cc so that there exists no copy of the string table. No copy means the strings in the table are interned strings. We can compare strings by the address of the strings. This optimization has already been used in operator==. Bug: 839389

The method has been removed from RenderView in favour of the newly added one in RenderFrame. Chromium change: https://chromium.googlesource.com/chromium/src/+/a3f2d52bc842c320df2d8d648f7510225e43992a commit a3f2d52bc842c320df2d8d648f7510225e43992a Author: Dave Tapuska <dtapuska@chromium.org> Date: Fri Jun 11 15:30:18 2021 +0000 Remove GetMainRenderFrame from RenderView. The MainFrame can be accessed off of WebView's API so there is no need to store it in RenderViewImpl. Provide a convenience method on RenderFrame to get the MainRenderFrame from the WebView. BUG=1155202

We need to remove the chromium_src overrides for that method, no longer existing, as well as to do some changes into RestrictedCookieManager to make sure that the calls to AnnotateAndMoveUserBlockedCookies() do take into account whether Ephemeral Storage is enabled or not. This change fixes the following 11 browser tests: * EphemeralStorageBrowserTest.ClosingTabClearsEphemeralStorage * EphemeralStorageBrowserTest.EphemeralStorageDoesNotLeakBetweenProfiles * EphemeralStorageBrowserTest.FirstPartyNestedInThirdParty * EphemeralStorageBrowserTest.NavigatingClearsEphemeralStorageAfterKeepAlive * EphemeralStorageBrowserTest.ReloadDoesNotClearEphemeralStorage * EphemeralStorageBrowserTest.StorageIsPartitioned * EphemeralStorageDisabledTest.CrossSiteCookiesBlockedInitial * EphemeralStorageKeepAliveDisabledBrowserTest.NavigatingClearsEphemeralStorageWhenKeepAliveDisabled * EphemeralStorageTest.CrossSiteCookiesBlockedInitial * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageDifferentSession * EphemeralStorageTest.CrossSiteCookiesBlockedThisPageSameSession Chromium change: https://chromium.googlesource.com/chromium/src/+/59f8a845e61de2ebbed367590b009abdf38eb450 commit 59f8a845e61de2ebbed367590b009abdf38eb450 Author: cfredric <cfredric@chromium.org> Date: Tue Jun 8 15:27:11 2021 +0000 Refactor CookieSettings to apply exclusion reasons directly. This clears the way for cookies to be blocked by CookieSettings individually, rather than on a per-header basis for cookie-sends. (Note that cookie-sets already occur on a per-cookie basis, though the cookie itself isn't referenced.) We will need to do this blocking individually in order to support the --sameparty-cookies-considered-first-party flag. Change-Id: I68b8e82c949f70cfe73e141e80e409da97bd27ca Bug: 1203706

Adapt the expected dirname in |preprocessed_folder| to allow JavaScript files generated out of TypeScript files that include our Brave-specific modules to find them when doing the rollup optimization, at build time. This prevents this build error from happening on optimized builds: [!] Error: Could not resolve './brave_overrides/index.js' from gen/chrome/browser/resources/extensions/tsc/extensions.js Error: Could not resolve './brave_overrides/index.js' from gen/chrome/browser/resources/extensions/tsc/extensions.js at error (/home/mario/work/brave-rebases/brave-browser-next/src/third_party/node/node_modules/rollup/dist/rollup.js:10162:30) at ModuleLoader.handleMissingImports (/home/mario/work/brave-rebases/brave-browser-next/src/third_party/node/node_modules/rollup/dist/rollup.js:17237:17) at ModuleLoader.<anonymous> (/home/mario/work/brave-rebases/brave-browser-next/src/third_party/node/node_modules/rollup/dist/rollup.js:17288:26) at Generator.next (<anonymous>) at fulfilled (/home/mario/work/brave-rebases/brave-browser-next/src/third_party/node/node_modules/rollup/dist/rollup.js:40:28) Ideally, Brave's JavaScript files should be migrated to TypeScript, but that should be done separate from the rebase, see GH issue #16553 [1]. [1] brave/brave-browser#16553 the time being this quick hack will keep Brave building.

Chromium change: https://chromium.googlesource.com/chromium/src/+/c84e15f86da8a6f00920d4dd03087cc88b8970f5 More feed launch reliability logging * Log the "feed reloading" start event * Log a catch-all launch end event when the stream is unbound: more specific feed launch end conditions (user navigated away, switched tabs/apps, etc.) will be added in another CL * Use no-op logger in public builds or if reliability logging is disabled Bug: 1195940

Chromium change: https://chromium.googlesource.com/chromium/src/+/21667acc2a3c5fe6738899fb494248c583bcc230 Rename ProfileSyncService to SyncServiceImpl in Android code No behavior is changed. Do the following renamings: - ProfileSyncService.java -> SyncServiceImpl.java - profile_sync_service_android.h -> sync_service_android_bridge.h - FakeProfileSyncService.java -> FakeSyncServiceImpl.java (and corresponding *WithFakeProfileSyncService test files) Moreover - Update the documentation of *sync_service_android_bridge.h Bug: 1201272

This extention is no longer used upstream as Media Router capabilities are now unconditionally enabled, meaning that there's no need to select any setting to have access to enable ChromeCast in the browser. Thus, drop the code from #2836, plus every other reference to this extension in Brave-specific code. Chromium change: https://chromium.googlesource.com/chromium/src/+/e287bab0c15dc688d9604f9e43993ccd3b217adf commit e287bab0c15dc688d9604f9e43993ccd3b217adf Author: Takumi Fujimoto <takumif@chromium.org> Date: Thu Jun 10 06:19:57 2021 +0000 Stop loading the Media Router component extension We no longer use the component extension so it can now be removed. Although this CL does not explicitly uninstall the extension, if the extension is installed, its files get deleted on startup via ExtensionService::CheckExternalUninstall(). Bug: 1179785, b/184392261 fixup! [WIP] Remove Brave-specific code related to the Media Router extension

…4549.4.

From //chrome/android/java to //chrome/browser/ui/android/omnibox/java. Chromium change: https://chromium.googlesource.com/chromium/src/+/385121b715b153153f7626baf728cbd588250ff7 commit 385121b715b153153f7626baf728cbd588250ff7 Author: Jinsuk Kim <jinsukkim@chromium.org> Date: Thu Jun 17 03:25:23 2021 +0000 Android: Modularize omnibox/ Moves the files in chrome/browser/omnibox and related resources to a new modularized target under chrome/browser/ui/android/omnibox. Bug: 1171512

We should use EraseListValue() now. Chromium change: https://chromium.googlesource.com/chromium/src/+/816cc6dcef88491540a7f18522a25f9b3d700501 commit 816cc6dcef88491540a7f18522a25f9b3d700501 Author: Maks Orlovich <morlovich@chromium.org> Date: Sun Jun 20 15:51:05 2021 +0000 Code health: Remove deprecated ListValue::Remove ... After converting the last remaining use in a base/ testcase. Bug: 1187094

Chromium change: https://chromium.googlesource.com/chromium/src/+/96da3e494a8b7e6839298cd653a92dfdffc3f7e3 commit 96da3e494a8b7e6839298cd653a92dfdffc3f7e3 Author: Peter Boström <pbos@chromium.org> Date: Wed Jun 16 03:40:22 2021 +0000 Make FocusRing a View class property This removes Button::focus_ring() in favor of a shared way of using FocusRing for all Views. Bug: None

We need to pass a content::WebContentsFrameReceiverSetPassKey() object in the constructor now to comply with security policies. Chromium change: https://chromium.googlesource.com/chromium/src/+/5b18ca94be3ef387e2aeeb5002e1bd236c2bf04a commit 5b18ca94be3ef387e2aeeb5002e1bd236c2bf04a Author: Robert Sesek <rsesek@chromium.org> Date: Wed Jun 16 14:52:41 2021 +0000 Restrict the use of content::WebContentsFrameReceiverSet with a pass key The WebContentsFrameReceiverSet lets classes register IPC bindings outside of the central registries that are reviewed by the security team. WebContentsFrameReceiverSet bypasses security review. Bug: 1213679

It expects a TabRelativeDirection parameter now instead of a bool. Chromium change: https://chromium.googlesource.com/chromium/src/+/cbe4171877a268d9b430425d51b5e2e76856ea1a commit cbe4171877a268d9b430425d51b5e2e76856ea1a Author: David Pennington <dpenning@chromium.org> Date: Mon Jun 14 19:53:42 2021 +0000 Use a more specific enum for directionality in TabStripModel The first in a series of small cleanup CLs for TabStripModel. Here the methods used for relative selection/movement of tabs is changed to have more specific params. An enum class is added which provided direction information and used in place of "forward". Change-Id: I5e118f293dd1d97880b63b36cd49900bd207f35f Bug: 1219551

Chromium change: https://chromium.googlesource.com/chromium/src/+/5cd115f54c670fafd04972a66725113fd51302ba commit 5cd115f54c670fafd04972a66725113fd51302ba Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Mon Jun 14 17:08:28 2021 +0000 Remove sync_ui_util namespace from c/b/sync/sync_ui_util.h No behavior is changed. Namespaces are discouraged for top-level embedders [1]. Since the namespace is removed: 1) Rename the following so they aren't too generic: - MessageType -> SyncStatusMessageType - ActionType -> SyncStatusActionType - StatusLabels -> SyncStatusLabels - GetStatusLabels() -> GetSyncStatusLabels() - GetStatus() -> GetSyncStatus() 2) Use enum classes rather than raw enums. Besides the above: 3) Convert ENUM_VALUE casing to kEnumValue. 4) Avoid one line over 80 characters and remove one unused function in sync_ui_util_unittest.cc [1] https://source.chromium.org/chromium/chromium/src/+/main:styleguide/c++/c++-dos-and-donts.md;l=294;drc=be05130dfdd2db2b14f73d0b99be3a2b4b118869 Fixed: 1219516

Widget itself is not used in Brave, but we need this new value to fix build. Chromium change: https://chromium.googlesource.com/chromium/src/+/44801e6aebacb2b8e7c16f11deb5fc8b6255e1ae Added a stub implementation of the Quick Action Search Widget The branded strings are defined in this CL: https://crrev.com/i/3903128 + The widget is gated behind a Finch flag and is also gated by Build channel using jinja vars Bug: 1213541

Chromium change: https://chromium.googlesource.com/chromium/src/+/385121b715b153153f7626baf728cbd588250ff7 Android: Modularize omnibox/ Moves the files in chrome/browser/omnibox and related resources to a new modularized target under chrome/browser/ui/android/omnibox. Bug: 1171512

Chromium change: https://chromium.googlesource.com/chromium/src/+/62ed7cdc676b4838bcf620911cb56df13b8430bb Android: Introduce LocaleManagerDelegate LocaleManagerDelegate/Impl is where the public/private behavior diff will go about locale-based operations. This CL introduces the delegate interface, and its impl to be used for the public apks. Also modularize c.b.locale under the directory chrome/browser/locale. In the next patch the private repo will have its own delegate impl in preparation for moving away from AppHooks. This is patch 1/4 for moving the package from AppHooks to its own build rule-based link. Bug: 1201640

Chromium change: https://chromium.googlesource.com/chromium/src/+/333aadcb14d5fe2f9b4f62441b971324cf4d6e14 [Signin][Android] Remove the pre-MICE code in MainSyncSettings This cl removes legacy tests and code paths in MainSyncSettings prior to MOBILE_IDENTITY_CONSISTENCY rollout. Bug: 1206701

…sugaring After D8 desugaring there are warnings (that are treated as errors in Release) on mising classes for code patched in bytecode. This happens because we add deps to general targets (brave_chrome_java_deps, brave_chrome_all_java_deps) rather then to each target separately, which we do to avoid extensive direct patching of gn targets. Example of warning: Warning in /tmp/tmpez828382/tmp_extract_dir/org/chromium/chrome/browser/search_engines/settings/SearchEngineAdapter.class: Type `org.chromium.chrome.browser.search_engines.settings.BraveBaseSearchEngineAdapter` was not found, it is required for default or static interface methods desugaring of `Lorg/chromium/chrome/browser/search_engines/settings/SearchEngineAdapter;refreshData()V` Chromium change: https://chromium.googlesource.com/chromium/src/+/cb2696352af0f051d9a08c3ca60bd99ab7cf35b2 Android: Use D8 for desugaring by default Bug: 1015559 Fixed: 1015559

…ookies() This new method replaces the formerly existing OnCanGetCookies() method. Chromium change: https://chromium.googlesource.com/chromium/src/+/59f8a845e61de2ebbed367590b009abdf38eb450 commit 59f8a845e61de2ebbed367590b009abdf38eb450 Author: cfredric <cfredric@chromium.org> Date: Tue Jun 8 15:27:11 2021 +0000 Refactor CookieSettings to apply exclusion reasons directly. This clears the way for cookies to be blocked by CookieSettings individually, rather than on a per-header basis for cookie-sends. (Note that cookie-sets already occur on a per-cookie basis, though the cookie itself isn't referenced.) We will need to do this blocking individually in order to support the --sameparty-cookies-considered-first-party flag. Change-Id: I68b8e82c949f70cfe73e141e80e409da97bd27ca Bug: 1203706

This reverts commit b2a8762. FIXME: Just removing it to keep things building before upgrading to a newer Cr93 tag that will include that definition.

…4557.4.

This reverts commit 86b16de. This change does undo the temporary build fix introduced for Chromium 93.0.4549.4, that didn't include kSegmentationPlatformStorageDirName.

The switch statement in PageInfoViewFactory::GetPermissionIcon() to be patched is now in page_info_view_factory.cc. Chromium change: https://chromium.googlesource.com/chromium/src/+/0429a54efa5e04cffe61634107206abcaa7313d3 commit 0429a54efa5e04cffe61634107206abcaa7313d3 Author: Olesia Marukhno <olesiamarukhno@google.com> Date: Mon Jun 21 09:43:20 2021 +0000 Reland "Update icons in page info" This is a reland of 242829fbd44212a65e30dadcaf383c52b0a0eb65 Original change's description: > Update icons in page info > > Updated all icons to size 16px. Most of them are 16px by default, some > of them require size to set (permissions, certificate). There was a > comment about why permission icons were 18px: some of them have internal > padding which might make them look smaller. I checked with a UX > designer, it looks fine in 16px (screenshots attached in the bug). > > Added layout constant for page info icon size instead of constant field > in page_info_ui. Moved icons methods from page_info_ui to > PageInfoViewFactory since they are for desktop only and it makes sense > to keep them closer to usages. Also, this makes using layout constant > easier. > > Icons have Get*Icon methods if they: > 1. are used in multiple views > 2. require a size to set > 3. require logic to determine the right icon (permissions) > > Added new icon - secondary back arrow. kBackArrowIcon doesn't fit here > and looks too big compared to other icons. > > Screenshots: https://bugs.chromium.org/p/chromium/issues/detail?id=1188101#c25 > > Bug: 1188101 > Change-Id: Id72aa6055275d4b12d986fb39e34369aa163a5a3 > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2949468 > Reviewed-by: Peter Kasting <pkasting@chromium.org> > Reviewed-by: Evan Stade <estade@chromium.org> > Reviewed-by: Emily Stark <estark@chromium.org> > Commit-Queue: Olesia Marukhno <olesiamarukhno@google.com> > Cr-Commit-Position: refs/heads/master@{#893515} Bug: 1188101

No need to explicitly set ui::NativeTheme::kColorId_FocusAuraColor anymore now, since there's no focus aura at all. Chromium change: https://chromium.googlesource.com/chromium/src/+/18b030938672b327ff05cb7e019ec772b22c230c commit 18b030938672b327ff05cb7e019ec772b22c230c Author: Peter Boström <pbos@chromium.org> Date: Tue Jun 22 16:41:38 2021 +0000 Remove FocusRing focus-aura functionality This was specifically used for ToggleButton, but the aura would paint on top of the control as part of the FocusRing. Instead this makes sure that the InkDrop only triggers when the ToggleButton isn't focused, that way the InkDrop highlight can remain through keyboard triggering. A TODO is left to investigate making sure that the InkDrop highlight returns to the control as it's still focused, this would let us remove this special case. Bug: 1221137

Also needed to adapt the constructor of BraveMenuButtonCoordinator to pass the two newly-added parameters to the base class. Chromium change: https://chromium.googlesource.com/chromium/src/+/4a6f627495111794c56c46873349521bcc9b5b22 commit 4a6f627495111794c56c46873349521bcc9b5b22 Author: Jinsuk Kim <jinsukkim@chromium.org> Date: Wed Jun 23 01:35:50 2021 +0000 Toolbar: Remove deps on chrome/browser/omaha This CL applies some refactoring around toolbar/omaha in preparation for toolbar modularization: - Moves menu button/ui/item state definitions from c.b.omaha.UpdateMenuItemHelper to toolbar/menu_button - Inject MenuButtonState supplier to MenuButton MVC instead of the components calling UpdateMenuItemHelper directly Bug: 1127732

Chromium change: https://chromium.googlesource.com/chromium/src/+/641e4e46425a0325b3f2841277a78986025534ad commit 641e4e46425a0325b3f2841277a78986025534ad Author: Colin Blundell <blundell@chromium.org> Date: Thu Jun 24 16:09:53 2021 +0000 [//components/safe_browsing] Move content/web_ui/ subdir This subdirectory is browser-side code and thus should be located nested under //components/safe_browsing/content/browser. This CL completes the reorganization of //components/safe_browsing/content into the canonical structure of having only {browser, common, renderer} subdirectories. Bug: 1216341

Chromium change: https://chromium.googlesource.com/chromium/src/+/aba4a46c62f5f3ec198e7b7bd8ac1bbd6c953d1d commit aba4a46c62f5f3ec198e7b7bd8ac1bbd6c953d1d Author: cfredric <cfredric@chromium.org> Date: Tue Jun 22 18:29:13 2021 +0000 Add new --sameparty-considered-first-party flag and logic. This flag will resolve cfredric/sameparty#4, and will allow developers to test the eventual interaction of SameParty cookies and third-party-cookie-blocking. It causes SameParty cookies to not be blocked by the user's third-party-cookie-blocking settings, as they are considered first-party. Bug: 1203706

This time from ntp_tiles::prefs:: to ntp_prefs:: namespace. Chromium change: https://chromium.googlesource.com/chromium/src/+/dca3350f9cdba0d34d80dfe22869487a73bda8a4 commit dca3350f9cdba0d34d80dfe22869487a73bda8a4 Author: Tibor Goldschwendt <tiborg@chromium.org> Date: Thu Jun 24 01:01:45 2021 +0000 [ntp] Move custom links and visibility MV tile prefs to 1P NTP Previously, these prefs where owned by the MostVisitedSites, which is also used for 3P NTPs. However, the 1P NTP is the only surface where these prefs should be applied. This CL makes the 1P NTP the owner of the prefs and isolates them from other NTPs. Fixed: 1219971

This is now defined in components/permissions/constants.cc instead of chrome/common/url_constants.cc, and relocated from the chrome:: to the permissions:: namespace. Chromium change: https://chromium.googlesource.com/chromium/src/+/c5c08e7b157f1c2534098701fb5fc3c98e7ca737 commit c5c08e7b157f1c2534098701fb5fc3c98e7ca737 Author: Michael van Ouwerkerk <mvanouwerkerk@chromium.org> Date: Thu Jun 24 08:53:27 2021 +0000 Move BluetoothChooserAndroid to //components/permissions/android This depends on the following which are also moved: * DeviceItemAdapter * DeviceItemRow * ItemChooserDialog * resources for strings, layouts, colors, drawables, and styles Bug: 1076462

Chromium change: https://chromium.googlesource.com/chromium/src/+/9e64611bf897a38c939ab40c1ab7d4c3c5a243fc commit 9e64611bf897a38c939ab40c1ab7d4c3c5a243fc Author: Mikel Astiz <mastiz@chromium.org> Date: Fri Jun 25 11:17:03 2021 +0000 [sync] Configure URL for trusted vault opt in flow This patch adjusts the URL used for the opt-in entrypoint for the sync trusted vault passphrase type, in settings. Change-Id: I9c4a0c52674e92775a564980f8ac696df2b985cb Bug: 1202088

Since now this flag is being passed to the compiler, we can't use the list-initializing syntax to build types when any of the parameters specified for constructing the new object would need to be implicitly converted into a different type of smaller size than the original one. For instance, we can't now list initialize kPageTransitionQualifierMask in the |PageTransitionType| enumeration to 0xFFFFFF00: enum PageTransitionType : int32_t { [...] // General mask defining the bits used for the qualifiers. kPageTransitionQualifierMask = 0xFFFFFF00, // THIS WILL FAIL NOW }; ...which would give us the following error at build time: enumerator value evaluates to 4294967040, which cannot be narrowed to type 'int32_t' (aka 'int') [-Wc++11-narrowing] This is because 0xFFFFFF00 can't be fitted inside a 32-bit unsigned int as the highest bit is reserved for the sign. Instead we could either define the enumeration as of type uint32_t or use 0x7FFFFF00 instead: enum PageTransitionType : int32_t { [...] // General mask defining the bits used for the qualifiers. kPageTransitionQualifierMask = 0x7FFFFF00, // THIS WORKS }; Since in this case this is used for defining types, we just define the enumeration of type uint8, so that we can use the more intuitive 0xFFFFFFFF mask instead. The right solution depends on each case, which is why this patch fixes the few problems found differently, in some cases by simply silencing the compilers using a static_cast<T>(value), in others by avoiding to use the list-initializing syntax, calling the constructor instead, and in some others adjusting the type used to a more appropriate one.

Mostly related to changes for various c-tor signatures.

Chromium change: https://chromium.googlesource.com/chromium/src/+/4a6f627495111794c56c46873349521bcc9b5b22 Toolbar: Remove deps on chrome/browser/omaha This CL applies some refactoring around toolbar/omaha in preparation for toolbar modularization: - Moves menu button/ui/item state definitions from c.b.omaha.UpdateMenuItemHelper to toolbar/menu_button - Inject MenuButtonState supplier to MenuButton MVC instead of the components calling UpdateMenuItemHelper directly Bug: 1127732

Chromium change: https://chromium.googlesource.com/chromium/src/+/b452ca696d154bb46945947cf2964f4955f3ad07 commit b452ca696d154bb46945947cf2964f4955f3ad07 Author: David Jean <djean@chromium.org> Date: Wed Jun 16 09:56:21 2021 +0000 [ios] Move ios deployment target to 13.0 To be submitted after https://chromium-review.googlesource.com/c/chromium/src/+/2944943 Bug: 1131852

Chromium change: https://source.chromium.org/chromium/chromium/src/+/b77854267a62e9208b1ac5ceb4167387ec7c56cc commit b77854267a62e9208b1ac5ceb4167387ec7c56cc Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Fri Jun 4 09:15:21 2021 +0000 Reland "Rename ProfileSyncService to SyncServiceImpl" Reason for reland: Original CL was reverted to unwind other changes. Original change's description: > Rename ProfileSyncService to SyncServiceImpl > > Renaming related classes like ProfileSyncServiceAndroid and > ProfileSyncServiceHarness is left for next CLs. We also use this CL as > an opportunity to hide the implementation class when possible: > - Some includes for sync_service_impl.h are removed if unnecessary, or > replaced with sync_service.h. > - Even comments referring to SyncServiceImpl are updated to refer to > the interface if the layer isn't aware SyncServiceImpl is being used. > > The CL also adds one missing include for <string> in > fake_server_helper_android.cc to address one IWYU presubmit warning. Bug: 1201272

Chromium change: https://source.chromium.org/chromium/chromium/src/+/a5d08cca10836dd75af3118827746bbabf10dc86 commit a5d08cca10836dd75af3118827746bbabf10dc86 Author: Nohemi Fernandez <fernandex@chromium.org> Date: Mon Jun 14 12:36:36 2021 +0000 [iOS] Remove unused API isSyncEnabled from SyncSetupService. Bug: 1215925

Chromium change: https://source.chromium.org/chromium/chromium/src/+/c17aab02c3bc7cb9732708126a5001d15ad6efd3 commit c17aab02c3bc7cb9732708126a5001d15ad6efd3 Author: Victor Hugo Vianna Silva <victorvianna@google.com> Date: Tue Jun 1 18:44:24 2021 +0000 Rename ios/'s ProfileSyncServiceFactory-s to SyncServiceFactory Same as crrev.com/c/2928734 but for the 2 factories in ios/: ProfileSyncServiceFactory and WebViewProfileSyncServiceFactory. Bug: 2928734

Not sure if we need to make additional changes yet. Chromium change: https://chromium.googlesource.com/chromium/src.git/+/7fdd29f commit 7fdd29f0f1782e917bfd1c8ff68824f02d85a926 Author: Tibor Goldschwendt <tiborg@chromium.org> Date: Thu Jun 24 14:24:03 2021 +0000 [ntp] Remove obsolete MV tile customization code from Instant NTPs http://crrev.com/c/2765626 removed 1P support from the local NTP and, thus, obsoleted support of MV tile customizations for Instant NTPs. This CL removes the obsolete code. Bug: 1041125

The new profile picker is unconditionally being enabled now upstream. Chromium change: https://source.chromium.org/chromium/chromium/src/+/62060fb114a2f54030c2c5200982eed77fa52d2c commit 62060fb114a2f54030c2c5200982eed77fa52d2c Author: David Roger <droger@chromium.org> Date: Tue May 25 23:22:41 2021 +0000 [profiles] Delete flag for new profile picker The new profile picker is now enabled by default, remove the flag to disable it. Bug: 1063856

…573.0.

Chromium change: https://source.chromium.org/chromium/chromium/src/+/2c6cbee37831bffa83494c98eca09a8ca9def489 commit 2c6cbee37831bffa83494c98eca09a8ca9def489 Author: Jinsuk Kim <jinsukkim@chromium.org> Date: Tue Jun 29 14:39:50 2021 +0000 Modularize toolbar/ Moves the files in chrome/browser/toolbar (except some glue code) and related resources to chrome/browser/ui/android/toolbar. Bug: 1127732

Chromium change: https://source.chromium.org/chromium/chromium/src/+/a56ecb8f054f9485c7340702a316647845d49346 commit a56ecb8f054f9485c7340702a316647845d49346 Author: Rushan Suleymanov <rushans@google.com> Date: Fri Jul 2 13:26:52 2021 +0000 [Sync] Reupload local device info on incoming tombstone The reupload takes place at most once per session when the data type is started. This is needed to prevent uncontrolled reuploads which base on incoming updates. Fixed: 1225961

This file has been removed from upstream now along with the old WebIDL compiler GN rules, which haven't been used for a while in Brave anyway due to the switch to V8's new IDL interface implementation. Besides the .idl files referenced from the .gni files listed in the brave_idl_imports variable are still referenced from the other variable brave_static_idl_files_in_modules, so we can remove the patch for third_party/blink/renderer/modules/modules_idl_files.gni now, along with the definition of brave_idl_imports and the .gni files listed from there. Chromium change: https://source.chromium.org/chromium/chromium/src/+/cc09d02098c2a513e9e53861fe5bf82990dad29b commit cc09d02098c2a513e9e53861fe5bf82990dad29b Author: Yuki Shiino <yukishiino@chromium.org> Date: Wed Jul 7 20:22:37 2021 +0000 Reland "bind-gen: Remove the old Web IDL compiler GN rules" This is a reland of 5a88d6a79aee536a09da0139631886318b7dec73 The original patch should have landed after https://crrev.com/c/3004452 , but actually landed before it by my misunderstanding. The change itself is correct, so there is no change since the original patch. Original change's description: > bind-gen: Remove the old Web IDL compiler GN rules > > Removes no longer used old Web IDL compiler GN rules. > > Bug: 839389 > Change-Id: I15955ab341c548592df10f385dbf5107b7f5e646 > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3004463 > Commit-Queue: Yuki Shiino <yukishiino@chromium.org> > Reviewed-by: Kentaro Hara <haraken@chromium.org> > Cr-Commit-Position: refs/heads/master@{#898536} Bug: 839389

This target has been removed upstream and merged into the pre-existing one at //components/safe_browsing/core/common, so we'd need to adapt and depend now on this other target. Chromium change: https://source.chromium.org/chromium/chromium/src/+/6292006aa24ad076f9e796e34627abc02ba56349 commit 6292006aa24ad076f9e796e34627abc02ba56349 Author: Colin Blundell <blundell@chromium.org> Date: Wed Jun 30 09:22:58 2021 +0000 [//components/safe_browsing] Move features* into core/common This code is cross-process code used on all platforms, and so it belongs in //components/safe_browsing/core/common. As part of this change, //components/safe_browsing/core/common:unit_tests is changed to run on iOS (this required no source code changes). Bug: 1216341

The //chrome/browser/search/drive and chrome/browser/search/task_module directories have been moved to //chrome/browser/new_tab_page/modules upstream, so we need to adapt to that. Chromium change: https://source.chromium.org/chromium/chromium/src/+/83b26a03aaacad7853096fb3257d3390f5e0c454 commit 83b26a03aaacad7853096fb3257d3390f5e0c454 Author: Tibor Goldschwendt <tiborg@chromium.org> Date: Fri Jul 9 23:13:18 2021 +0000 [ntp] Move 1P services to //chrome/browser/new_tab_page Before the deprecation of the local NTP, Instant NTPs and the 1P WebUI shared functionality implemented by services in //chrome/browser/search. Since, those NTPs have distinct functionality and sharing of services is no longer necessary. However, the code is still entangled; and NTPs instantiate services they don't need. We regard this as technical debt and are in the process of removing it. To aide this process this CL moves services only required and depended on by the 1P WebUI NTP from //chrome/browser/search to //chrome/browser/new_tab_page. After this CL, some 1P services, such as for backgrounds or promos, will still reside in //chrome/browser/search. These services are not technically used by Instant NTPs anymore but are still entangled with Instant NTP code. We will detangle and move those services in follow-up CLs. Change-Id: Ie66ea17eca89a5556785e1398510e6e0acc6a65c Bug: 1084363

Required for base::OptionalOrNullptr().

Required for network::ResourceRequest.

FrameBorderThickness() and FrameTopThickness() do no longer exist, we need to use FrameBorderInsets() and FrameEdgeInsets() now. Chromium change: https://source.chromium.org/chromium/chromium/src/+/8e48353818f27c27ed55c40f8f9931acce0c120a commit 8e48353818f27c27ed55c40f8f9931acce0c120a Author: Tom Anderson <thomasanderson@chromium.org> Date: Tue Jun 29 18:21:13 2021 +0000 Use gfx::Insets to describe the frame border size OpaqueBrowserFrameViewLayout previously used an int to describe all 4 sides of the frame border. It also used 2 ints to describe the 3D portion of the frame border (one int for the top and another for the remaining sides). When using client side decorations, the shadow is typically offset such that it's thinner at the top and thicker at the bottom. Rather than adding another int for the bottom border, this CL switches to using gfx::Insets. This will also be required when we eventually allow styling the window frame with GTK, which can in general have a different decoration size of each edge. Also, make the gfx::Insets getters virtual so that OpaqueBrowserFrameViewLayoutLinux can override them. R=sky Change-Id: I37e9aff72ca8bc49123b265c7ef2bab3f56311a7 Bug: 650494, 811515

We should use GetList().empty() now. Chromium change: https://source.chromium.org/chromium/chromium/src/+/ffeeed8d45af58f284a8b3624491745e7b10d2b1 commit ffeeed8d45af58f284a8b3624491745e7b10d2b1 Author: Morten Stenshorne <mstensho@chromium.org> Date: Thu Jul 8 22:58:38 2021 +0000 CodeHealth: Remove ListValue::empty() All usage should be replaced with Value::GetList()::empty() by now. Bug: 1187065

Adapt to this change upstream, and add the new include required. Chromium change: https://source.chromium.org/chromium/chromium/src/+/362c4a00017b8ed537014db3896c98552368a6b0 commit 362c4a00017b8ed537014db3896c98552368a6b0 Author: cfredric <cfredric@chromium.org> Date: Fri Jul 9 22:40:40 2021 +0000 Add UseCounters and UMA to measure incidence of SameSite=None cookies in certain contexts. Specifically, SameSite=None cookies that could be replaced by: * Nothing (have to remain a SameSite=None cookie) * a SameParty cookie, using the top frame and resource URL to determine sameparty-ness * a SameParty cookie, using the whole frame tree to determine sameparty-ness * a SameSite=Lax cookie * a SameSite=Strict cookie These metrics fire for both reads and writes.

It used to return a HostPortPair but it does now return an Endpoint, which is an absl::variant<url::SchemeHostPort, HostPortPair> instead. Chromium change: https://source.chromium.org/chromium/chromium/src/+/c98a3e6b1d4e3927bbf0785b190773c8460665be commit c98a3e6b1d4e3927bbf0785b190773c8460665be Author: Eric Orth <ericorth@chromium.org> Date: Fri Jul 2 17:46:37 2021 +0000 Maintain destination scheme in ConnectJobs As ConnectJobs are sometimes created when a scheme is unknown or non- standard, use absl::variants to continue allowing creation without scheme when necessary. Leaving proxy ConnectJobs for a subsequent CL. For now, they do not accept scheme for the eventual destination or the proxy endpoint. Bug: 1206799

This definitions lives under Blink's media namespace now. Chromium change: https://source.chromium.org/chromium/chromium/src/+/d7db8f85c0d7f9df87afea0fa62da78ae188fb5c commit d7db8f85c0d7f9df87afea0fa62da78ae188fb5c Author: Henrique Ferreiro <hferreiro@igalia.com> Date: Thu Jul 8 21:18:05 2021 +0000 Migrate t/b/{public,renderer}/platform/media to blink ns [2/6] This CL is a follow-up to https://crrev.com/c/2847359 and https://crrev.com/c/2925475. The code moved from media/blink/ to Blink should use the blink namespace. This CL just adds needed media:: prefixes to code in third_party/blink/renderer/platform/media/. Bug: 1198341

FocusRingTheme is a class with a trivial destructor. Chromium change: https://source.chromium.org/chromium/chromium/src/+/ded7717f5252a0c9d9de56bc22483b195d1801e0 commit ded7717f5252a0c9d9de56bc22483b195d1801e0 Author: Avi Drissman <avi@chromium.org> Date: Fri Jul 2 18:23:00 2021 +0000 Prevent use of base::NoDestructor for trivially-destructible types base::NoDestructor prevents calling a destructor of the templated type. It's not needed if there is no destructor call to prevent. This relands b961c7632f268f853a88869ad6bc5f33c3a4d549 with a small fix to code that was behind an official build flag that failed to compile. Bug: 1223520

A recent change converted //components/permissions:permissions_common from being a source set to a componnent, so we need to annotate this declaration now for the symbol to get properly exported. Chromium change: https://source.chromium.org/chromium/chromium/src/+/7709f0ce1f555922ab538a6861205f3ee5a485fc commit 7709f0ce1f555922ab538a6861205f3ee5a485fc Author: Illia Klimov <elklm@google.com> Date: Mon Jul 5 09:42:39 2021 +0000 Reland "Use GetLastCommittedOrigin for permissions." This reverts commit 2fb676a482e822fdd2bd9449e0dcd637c88f28e7. Reason for revert: The crash happens due to double base::Feature initialization. This CL fixes that. Original change's description: > Revert "Use GetLastCommittedOrigin for permissions." > > This reverts commit 57faf740e23368163ecbed24f76da11f9fe8875d. > > Reason for revert: speculative revert as this seems a prime candidate for the breakage in content_browsertests on Linux CrOS (https://ci.chromium.org/ui/p/chromium/builders/ci/linux-chromeos-dbg/24420/test-results) > > Original change's description: > > Use GetLastCommittedOrigin for permissions. > > > > Refactor call sites that work with origins for the purpose of permission > > state management to go through a central utility function that handles > > all edge cases. > > > > Bug: 698985,1210669,1220660 > > Change-Id: If76fd05c831dd3d3ac6d1274a6f4d1404ee17957 > > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2944834 > > Commit-Queue: Illia Klimov <elklm@google.com> > > Reviewed-by: Avi Drissman <avi@chromium.org> > > Reviewed-by: danakj <danakj@chromium.org> > > Reviewed-by: Mike Wasserman <msw@chromium.org> > > Reviewed-by: Balazs Engedy <engedy@chromium.org> > > Reviewed-by: Guido Urdaneta <guidou@chromium.org> > > Reviewed-by: Lei Zhang <thestig@chromium.org> > > Reviewed-by: Wez <wez@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#896916} > > Bug: 698985,1210669,1220660 > Change-Id: I3222c21d13f29f5365d283f0e1d3b730f3774409 > No-Presubmit: true > No-Tree-Checks: true > No-Try: true > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2995064 > Auto-Submit: W. James MacLean <wjmaclean@chromium.org> > Reviewed-by: Roman Arora <romanarora@google.com> > Reviewed-by: W. James MacLean <wjmaclean@chromium.org> > Owners-Override: Roman Arora <romanarora@google.com> > Owners-Override: W. James MacLean <wjmaclean@chromium.org> > Commit-Queue: W. James MacLean <wjmaclean@chromium.org> > Cr-Commit-Position: refs/heads/master@{#897033} Bug: 698985,1210669,1220660

Tested on Linux and it preserves the original behaviour from PR #9424. [1] #9424 Chromium change: https://source.chromium.org/chromium/chromium/src/+/2a4b9bfb525d77dfeec91be45068f9c95b00a8e9 commit 2a4b9bfb525d77dfeec91be45068f9c95b00a8e9 Author: Peter Kasting <pkasting@chromium.org> Date: Sat Jul 10 01:41:59 2021 +0000 Avoid accessing NativeTheme too early from bookmarks-related functions. This must only be accessed when the caller is in a Widget. Fixing this required significantly reworking how bookmark folder images are handled. This also modifies the color of bookmark folders on non-Win, non-Mac platforms: they are grey 700 by default (in light mode), like other icons, instead of being something closer to grey 600; and when a custom theme modifies the bookmark bar text color, folder icons in the overflow menu do not change similarly (but keep their standard menu colors). The menu does not necessarily have the same background color as the bookmark bar, so matching the bar's foreground color doesn't make sense, doubly so when done just for the folder icons and nothing else. Bug: 1211091

Upstream re-introduced a cr_toolbar.js import in history_toolbar.ts, which collides with our br_toolbar.m.js, which also defines cr-toolbar, causing a Javascript error when opening the history page like this: Uncaught NotSupportedError: Failed to execute 'define' on 'CustomElementRegistry': the name "cr-toolbar" has already been used with this registry", source: chrome://history/ (0) Remove that import to get the history page working again. Chromium change: https://source.chromium.org/chromium/chromium/src/+/f3907c07a8a0c13b1b513f507719fe3e316d8c80 commit f3907c07a8a0c13b1b513f507719fe3e316d8c80 Author: dpapad <dpapad@chromium.org> Date: Wed May 26 17:41:25 2021 +0000 History: Migrate to TypeScript. - Remove js_library(), js_type_check() targets. - Add new ts_library() target, and hook it up to the build. - Migrate all files to TypeScript, and fix all errors. - Override default tsconfig with a few options that seem necessary for Polymer code. - Introduce EventTracker usage in app.ts as part of the migration (easier than adding type info for boundOnKeyDown_) - Rename 'open-menu' event to 'synced-device-card-open-menu' in synced_device_card.ts, to be able to differentiate with the 'open-menu' event in history_card.ts, when declaring the Event type in the global HTMLElementEventMap. Bug: 1189595

Upstream re-introduced a cr_toolbar.js import in bookmarks/toolbar.ts, which collides with our br_toolbar.m.js, which also defines cr-toolbar, causing a Javascript error when opening the bookmarks page like this: Uncaught NotSupportedError: Failed to execute 'define' on 'CustomElementRegistry': the name "cr-toolbar" has already been used with this registry", source: chrome://bookmarks/ (0) Remove that import to get the bookmarks page working again. Chromium change: https://source.chromium.org/chromium/chromium/src/+/

Chromium change: https://source.chromium.org/chromium/chromium/src/+/d4a2f8e1ed461d8fc05ed88d1ae2dc94c9773825 commit d4a2f8e1ed461d8fc05ed88d1ae2dc94c9773825 Author: Lei Zhang <thestig@chromium.org> Date: Thu Jul 8 03:43:28 2021 +0000 Swap base/stl_util.h to base/cxx17_backports.h in many files. Files that use base::size() and base::data() should use cxx17_backports.h directly, instead of getting it transitively through stl_util.h. Bug: 1210983

Constness for this method has been recently added upstream, so we need to do this on our override, plus some small adaptation since we can't now return a reference to the internal TestingPrefServiceSimple member. Chromium change: https://source.chromium.org/chromium/chromium/src/+/91476e0ceeda870b0a7e2883330093114eb82272 commit 91476e0ceeda870b0a7e2883330093114eb82272 Author: Orin Jaworski <orinj@google.com> Date: Fri Jul 2 22:24:32 2021 +0000 [omnibox] Add and implement PLAY_CHROME_DINO_GAME Pedal This CL creates the second Pedal for batch 3, which can be triggered with minimal input "chrome dino". Bug: 893183

The format changed from "Cannot set property '%' of %" to "Cannot read properties of % (reading '%')") recently in V8. This fixes the following browser test: * SerialAPIBrowserTest.SerialAPIDisabled V8 change: https://chromium.googlesource.com/v8/v8/+/c0fd89c3c089e888c4f4e8582e56db7066fa779b commit c0fd89c3c089e888c4f4e8582e56db7066fa779b Author: Patrick Thier <pthier@chromium.org> Date: Thu Jun 17 11:15:41 2021 +0000 Reland "Reland "Reland "Improve error messages for property access on null/undefined""" This is a reland of 819c3ae2f87fc7ad634fe4356de6653bb3cb3b87 Original change's description: > Reland "Reland "Improve error messages for property access on null/undefined"" > > This is a reland of 8b18c5e6a534d7c459a16aebdd4fe05133db6dea > > Original change's description: > > Reland "Improve error messages for property access on null/undefined" > > > > This is a reland of 24c626c1f7809ce725e56152ab2928696a2f8e9e > > > > Original change's description: > > > Improve error messages for property access on null/undefined > > > > > > Only print the property name when accessing null/undefined if we can > > > convert it to a string without causing side effects. > > > If we can't, omit the property name in the error message. > > > This should avoid confusion when the key is an object with toString(). > > > E.g. undefined[{toString:()=>'a'}] doesn't print 'read property [object > > > Object]' anymore, which was misleading since the property accessed would > > > be 'a', but we can't evaluate the key without side effects. > > > > > > Bug: v8:11365 > > > Change-Id: If82d1adb42561d4851e2bd2ca297a1c71738aee8 > > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2960211 > > > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > > > Commit-Queue: Patrick Thier <pthier@chromium.org> > > > Cr-Commit-Position: refs/heads/master@{#75250} > > > > Bug: v8:11365 > > Change-Id: Ie2312337f4f1915faa31528a728d90833d80dbd1 > > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2979599 > > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > > Commit-Queue: Patrick Thier <pthier@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#75571} > > Bug: v8:11365 > Change-Id: I90360641ecd870bd93247aa6d91dfb0ad049cfb8 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3008219 > Auto-Submit: Patrick Thier <pthier@chromium.org> > Commit-Queue: Toon Verwaest <verwaest@chromium.org> > Reviewed-by: Toon Verwaest <verwaest@chromium.org> > Cr-Commit-Position: refs/heads/master@{#75604} Bug: v8:11365 Change-Id: I002b537144f328ccbbdcd655e26e5dc87c49c6f5 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3013935 Reviewed-by: Toon Verwaest <verwaest@chromium.org> Commit-Queue: Patrick Thier <pthier@chromium.org> Cr-Commit-Position: refs/heads/master@{#75645}

Now uses signin::BuildTestSigninClient, so need to include test_signin_client_builder.h/cc directly into the unit tests build target like it's done upstream. Chromium change: https://chromium.googlesource.com/chromium/src.git/+/579a983 commit 579a9833c2c5ae8d633fe4e775e86a455037494d Author: David Roger <droger@chromium.org> Date: Wed Jul 7 11:46:17 2021 +0000 [sync] Add missing call to UpdateAccountsInCookieJarInfoIfNotStale This CL adds the missing call in components/sync/trusted_vault/standalone_trusted_vault_client.cc and updates the unit tests on ChromeOS. Bug: 1148328

Additionaly fixed: - c-tor signatures for various classes - explicitly added header for network::ResourceRequest Chromuim change: https://chromium.googlesource.com/chromium/src/+/2c6cbee37831bffa83494c98eca09a8ca9def489 Modularize toolbar/ Moves the files in chrome/browser/toolbar (except some glue code) and related resources to chrome/browser/ui/android/toolbar. Bug: 1127732

Chromium change: https://chromium.googlesource.com/chromium/src/+/296e58b25a68baf8bf044234555c7bd157d5d761 [GMNext][TabSwitcher] Change selection style to material next Disabled: Selected tab will have a hairline drawable foreground; Enabled: Selected tab will be tinted into primary color. Color in incognito mode will be using the ones for dark mode, and will not support dynamic colors. (The update for thumbnail is not included in this CL) Spec see original bug attached. Bug: 1217141

…ut patching

This was breaking the following 37 tests because of not being able to find br_toolbar.m.js in the Brave resources map used for testing: * BraveContentBrowserClientTest.CanLoadCustomBravePages * RewardsBrowserTest.BackupRestoreModalHasNoNotice * RewardsBrowserTest.ResetRewards * RewardsBrowserTest.ShowACPercentInThePanel * RewardsContributionBrowserTest.AutoContributeWhenACOff * RewardsContributionBrowserTest.AutoContribution * RewardsContributionBrowserTest.AutoContributionMultiplePublishers * RewardsContributionBrowserTest.AutoContributionMultiplePublishersUphold * RewardsContributionBrowserTest.CheckIfReconcileWasReset * RewardsContributionBrowserTest.CheckIfReconcileWasResetACOff * RewardsContributionBrowserTest.MultipleRecurringOverBudgetAndPartialAutoContribution * RewardsContributionBrowserTest.PanelMonthlyTipActions * RewardsContributionBrowserTest.PanelMonthlyTipAmount * RewardsContributionBrowserTest.PendingContributionTip * RewardsContributionBrowserTest.ProcessPendingContributions * RewardsContributionBrowserTest.RecurringAndPartialAutoContribution * RewardsContributionBrowserTest.SplitProcessOneTimeTip * RewardsContributionBrowserTest.SplitProcessorAutoContribution * RewardsContributionBrowserTest.TipConnectedPublisherAnon * RewardsContributionBrowserTest.TipConnectedPublisherAnonAndConnected * RewardsContributionBrowserTest.TipConnectedPublisherConnected * RewardsContributionBrowserTest.TipNonIntegralAmount * RewardsContributionBrowserTest.TipVerifiedPublisher * RewardsContributionBrowserTest.TipWithVerifiedWallet * RewardsContributionBrowserTest.RecurringTipNonIntegralAmount * RewardsNotificationBrowserTest.InsufficientNotificationForACNotEnoughFunds * RewardsNotificationBrowserTest.InsufficientNotificationForInsufficientAmount * RewardsNotificationBrowserTest.InsufficientNotificationForVerifiedInsufficientAmount * RewardsP3ABrowserTest.WalletBalanceLessThan10BAT * RewardsP3ABrowserTest.WalletBalanceLessThan50BAT * RewardsP3ABrowserTest.WalletBalanceMoreThan50BAT * RewardsP3ABrowserTest.WalletStateWalletCreatedGrantsClaimedFundsAdded * RewardsP3ABrowserTest.WalletStateWalletCreatedGrantsClaimedNoFundsAdded * RewardsP3ABrowserTest.WalletStateWalletCreatedNoGrantsClaimedFundsAdded * RewardsPublisherBrowserTest.VisitRegisteredPublisher * RewardsPublisherBrowserTest.VisitUnverifiedPublisher * RewardsPublisherBrowserTest.VisitVerifiedPublisher

These fixes 3 tests failing due to producing excessive output: * BraveDarkModeFingerprintProtection.SettingsPagesCheck * SidebarBrowserTest.BasicTest * SidebarBrowserTest.WebTypePanelTest

Previously routes were being defined after the url was parsed to check for a matching route

Chromium change: https://chromium.googlesource.com/chromium/src/+/ca63ae8c5f7d1ee23cb865ffe0a51159950ff8cc [GURL] Remove Tab.getUrlString calls from Favicon code. This starts the process of threading GURL into favicon code but there's a lot of it. :/ Also removes a bunch of dead code I came across in FaviconHelper. Bug: 783819

…4577.8.

We can't fix the test by simply inserting an additional line anymore now that the |num_items| variable has been removed, but it would still be good why we need to account for one more item, so let's keep using the BRAVE_MAX_TABS_PER_SESSION_AND_RECENCY for that purpose, even if it does replace a line from upstream now (vs just adding one). Chromium change: https://chromium.googlesource.com/chromium/src/+/48a117a8da6c45f1ef8a30af74c40c97185d12ad commit 48a117a8da6c45f1ef8a30af74c40c97185d12ad Author: Emily Shack <emshack@chromium.org> Date: Mon Jul 12 23:25:26 2021 +0000 Add additional testing for history sub menu contents Bug: 1205646

Chromium change: https://chromium.googlesource.com/chromium/src/+/a46cb2796033780fee5a9a73a3645c2555e896c3 commit a46cb2796033780fee5a9a73a3645c2555e896c3 Author: Morten Stenshorne <mstensho@chromium.org> Date: Tue Jul 13 13:44:11 2021 +0000 CodeHealth: Remove ListValue::Clear() All usage should be replaced with Value::ClearList() by now. Bug: 1187063

kEphemeralGuestProfileDirPrefix is now gone. Chromium change: https://chromium.googlesource.com/chromium/src/+/35dccf0786c4df7b4f9859a1604dd5be76e7df4f commit 35dccf0786c4df7b4f9859a1604dd5be76e7df4f Author: Ramin Halavati <rhalavati@chromium.org> Date: Mon Jul 12 06:16:06 2021 +0000 Remove all remaining traces of Ephemeral Guest profiles. Ephemeral Guest profiles are deprecated. The last traces of them are removed. Good bye! Bug: 1225156

Chromium change: https://chromium.googlesource.com/chromium/src/+/35dccf0786c4df7b4f9859a1604dd5be76e7df4f commit 35dccf0786c4df7b4f9859a1604dd5be76e7df4f Author: Ramin Halavati <rhalavati@chromium.org> Date: Mon Jul 12 06:16:06 2021 +0000 Remove all remaining traces of Ephemeral Guest profiles. Ephemeral Guest profiles are deprecated. The last traces of them are removed. Good bye! Bug: 1225156

Chromium change: https://chromium.googlesource.com/chromium/src/+/97d3f8150aa5e9dc1603b4ea8f6c7d003ec75b07 commit 97d3f8150aa5e9dc1603b4ea8f6c7d003ec75b07 Author: Alex Rudenko <alexrudenko@chromium.org> Date: Wed Jul 14 10:00:38 2021 +0000 Code health: remove Value::GetAsInteger Fixed: chromium:1187004

Chromium change: https://chromium.googlesource.com/chromium/src/+/a538cc5712e01175153e0dd374028de6dcdd644a commit a538cc5712e01175153e0dd374028de6dcdd644a Author: Marc Treib <treib@chromium.org> Date: Thu Jul 15 16:31:33 2021 +0000 Cleanup: Get rid of SyncServiceImpl::SyncStopDataFate This enum was basically a subset of syncer::ShutdownReason (despite the different naming); no good reason to have both. StopImpl() received SyncStopDataFate as a parameter, but in one of the two cases, it just called ShutdownImpl(). Matching callers are updated to instead call ShutdownImpl() directly, so the param isn't needed anymore. Bug: 1229171

The signin promo has been entirely deleted from upstream, so we neither need nor can keep this code around, as it won't do anything other than breaking the build. This is essentially a revert of github.com//pull/3566 Chromium change: https://chromium.googlesource.com/chromium/src/+/41d144e7c8dbe2b5a59ff87b0d45566fb3a0c468 commit 41d144e7c8dbe2b5a59ff87b0d45566fb3a0c468 Author: Marc Treib <treib@chromium.org> Date: Mon Jul 12 09:47:19 2021 +0000 B4P cleanup, part 3: Delete Passwords Signin promo This promo is not used anymore and so can be deleted. This CL also deletes the related helper ShouldShowChromeSignInPasswordPromo and deprecates the related prefs. Bug: 1108738

The sql::Statement parameter passed to DecryptedCardFromColumn() can't be const as the method will call ColumnByteLength() over it, which is no longer declared a const method. Chromium change: https://chromium.googlesource.com/chromium/src/+/081d534f16b6b1523ca9a87a081285ab01d30265 commit 081d534f16b6b1523ca9a87a081285ab01d30265 Author: Victor Costan <pwnall@chromium.org> Date: Thu Jul 15 14:23:59 2021 +0000 sql: Remove const qualifier from Statement::Column*() methods. The sql::Statement::Column*() methods are currently const, which suggests no internal state changes. This is incorrect for the following reasons. 1. GetColumnType()'s comment indicates that Column*() perform SQLite type conversion. So, the methods appear to change the underlying SQLite state. 2. In general, the methods call into SQLite, and we can't guarantee that SQLite state doesn't change. This CL fixes the problem by removing the const qualifier from the impacted methods. Bug: 1229420

Adapt the override in Brave subclass BraveDefaultBrowserDialogView. Chromium change: https://chromium.googlesource.com/chromium/src/+/f789e9cb09a4f4c0924b7564967084e1b2a54bc4 commit f789e9cb09a4f4c0924b7564967084e1b2a54bc4 Author: Peter Boström <pbos@chromium.org> Date: Wed Jul 14 19:31:59 2021 +0000 Remove DialogDelegate::OnDialogInitialized Existing uses are replaced with OnWidgetInitialized() overrides. They were synonymous as DialogDelegate::OnWidgetInitialized() was a final override that just called OnDialogInitialized(). Bug: 1144829

The SELECT statement in the Migration_27_UnblindedTokens unit test from LedgerDatabaseMigrationTest queries exactly 7 columns per result from the unblinded_tokens table, so we can't query for sql.ColumnInt64(10) as the index will be out of bounds, resulting on a crash on DCHECK-enabled builds after a recent change upstream. Chromium change: https://chromium.googlesource.com/chromium/src/+/45cfbb09b4d27ff712a1b4169b8ea69fb371472e commit 45cfbb09b4d27ff712a1b4169b8ea69fb371472e Author: Victor Costan <pwnall@chromium.org> Date: Wed Jul 14 07:23:26 2021 +0000 sql: Add index checks to sql::Statement::Column*() methods. Change-Id: I7c119e57710c7ba986875b2a63d9deca3d267264 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3021342 Commit-Queue: Victor Costan <pwnall@chromium.org> Reviewed-by: Marijn Kruisselbrink <mek@chromium.org> Cr-Commit-Position: refs/heads/master@{#901398}

Moved bytecode patching as process_java_prebuilt templated was reorganized. Chromium change: https://chromium.googlesource.com/chromium/src/+/723dd9797f2e767ba348043c8a57964576445f69 [Build] Avoid filter_jar() in java_library_impl() part 1/2 This CL: - Moves proguard dependencies out of GN process_java_prebuilt() template. - Renames process_java_prebuilt() template to process_java_library() because the template is not prebuilt-specific. - Changes process_java_library() template to handle either host jars or device jars in order to remove duplicated code. BUG=1176743

They should be pointing to br_toolbar_search_field.js now, otherwise this will break the build.

Chromium change: https://source.chromium.org/chromium/chromium/src/+/c58ebb0736e6ea806ac700bcbb6786d3cc7e71db commit c58ebb0736e6ea806ac700bcbb6786d3cc7e71db Author: Lily Chen <chlily@chromium.org> Date: Tue May 25 20:48:32 2021 +0000 Remove localStorage legacy database migration code This removes the code associated with migrating localStorage data from legacy SQLite databases to LevelDB. The migration first landed in M61/62, in late 2017, and we can typically remove migration code after 2 years. Bug: 1206834

… Linux This test currently fails on Linux platforms due to an upstream bug when SwANGLE is used, see upstream bug at http://crbug.com/1192632.

…4577.15.

Chromium change: https://source.chromium.org/chromium/chromium/src/+/286f0238ba16dcb5598a40551a634f853dd49b0a commit 286f0238ba16dcb5598a40551a634f853dd49b0a Author: Sylvain Defresne <sdefresne@chromium.org> Date: Fri Jul 9 12:38:54 2021 +0000 [ios] Preparation for exporting a free function based provider API Add //ios/chrome/browser/providers:null_provider_api that exposes a no-op version of the API (based on ChromiumBrowserProvider) and also //ios/public/provider/chrome/browser:test_provider_api that exposes a test version of the API (uses TestChromeBrowserProvider). Both those targets also depends on MaterialDesignComponents since this is an indirect dependency that the internal provider exposes in a similar way. This allow not adding checks in the rest of the code on whether to link with MaterialDesignComponents. Change all the target to only link with a single version of the provider API. Since all of those API exposes the same functions, it will causes linker errors if more than one is linked at the same time. Remove the class TestChromeProviderInitializer and instead use the CreateChromeBrowserProvider() factory function in the Chrome test suite class (thus ensuring that tests can choose which provider is used). Bug: 1201182

Chromium change: https://source.chromium.org/chromium/chromium/src/+/25e69d2197832ba6b08fbc613efa3361affba55c commit 25e69d2197832ba6b08fbc613efa3361affba55c Author: Sylvain Defresne <sdefresne@chromium.org> Date: Thu Jul 8 13:40:49 2021 +0000 [ios] Change return type of ios::GetChromeBrowserProvider() It is a programming error to call ios::GetChromeBrowserProvider() before ios::SetChromeBrowserProvider(...) has been called. Encode this invariant by changing ios::GetChromeBrowserProvider() return type to `ios::ChromeBrowserProvider&` (as a reference can't be null). Bug: none

Chromium change: https://chromium.googlesource.com/chromium/src.git/+/66ff85613ad162ab35051399d8fc7bc335fa7f87 commit 66ff85613ad162ab35051399d8fc7bc335fa7f87 Author: Morten Stenshorne <mstensho@chromium.org> Date: Tue Jul 13 08:53:19 2021 +0000 CodeHealth: Modernize base::Value code in components/bookmarks/ Avoid DictionaryValue and ListValue. Get rid of some temporary heap allocation. Bug: 1187062, 1187061

….4577.18.

Chromium change: https://source.chromium.org/chromium/chromium/src/+/8690530ca5d0a06f5708b3891f25aaae308f01f7 commit 8690530ca5d0a06f5708b3891f25aaae308f01f7 Author: Alexander Timin <altimin@chromium.org> Date: Sat Jun 19 00:34:32 2021 +0000 Rename GlobalFrameRoutingId to GlobalRenderFrameHostId. This id points to a RenderFrameHost, not frame so rename it to avoid confusion with frame tree node id. RenderFrameHost::GetGlobalFrameRoutingId is renamed to GetGlobalId as the "RenderFrameHost" is obvious from the class and the return type. BUG=1216208

Use OnWidgetInitialized instead. Chromium change: https://source.chromium.org/chromium/chromium/src/+/f789e9cb09a4f4c0924b7564967084e1b2a54bc4 commit f789e9cb09a4f4c0924b7564967084e1b2a54bc4 Author: Peter Boström <pbos@chromium.org> Date: Wed Jul 14 19:31:59 2021 +0000 Remove DialogDelegate::OnDialogInitialized Existing uses are replaced with OnWidgetInitialized() overrides. They were synonymous as DialogDelegate::OnWidgetInitialized() was a final override that just called OnDialogInitialized(). Bug: 1144829

….4577.25.

Chromium change: https://source.chromium.org/chromium/chromium/src/+/eb24c8644e3297ac1e5411724fb1c4c8c4950aa2 commit eb24c8644e3297ac1e5411724fb1c4c8c4950aa2 Author: Rushan Suleymanov <rushans@google.com> Date: Tue Jul 13 14:25:02 2021 +0000 [Sync] Prevent additional GetUpdates on browser startup For the new invalidations there should be an additional GetUpdates request after subscribing to new data types. An interested data types manager does it. However, while it's not initialized after browser startup, the manager may trigger refresh for all data types (since all data types are "new"). It doesn't happen always and may be triggered only in case if the device info has been changed (or its pulse timer has been elapsed). This CL prevents an additional GetUpdates request when the sync engine is initialized after browser restart. It is done by comparing the new data type list with the previous one which has been committed to the server (based on DeviceInfo stored specifics). The comparison is performed in DeviceInfoSyncBridge which also triggers refresh directly using repeating callback. Bug: 1192576

This was causing an error: Failed to load module script: The server responded with a non-JavaScript MIME type of "text/html". Strict MIME type checking is enforced for module scripts per HTML spec.

…moval Chromium removal of InstantService usage tracked with https://bugs.chromium.org/p/chromium/issues/detail?id=1084363 This commit: - Renames from `InstantServiceMessageHandler` to `TopSitesMessageHandler` - Borrows logic from `MostVisitedHandler` - InstantServiceObserver removed in favor of ntp_tiles::MostVisitedSites::Observer - Preserves our Super Referral logic - Preserves logic where editing a site in "most visited" mode switches you to "custom links" mode. This is possible in our UI by editing a tile when in most visited mode.

Chromium change: https://chromium.googlesource.com/chromium/src/+/296e58b25a68baf8bf044234555c7bd157d5d761 [GMNext][TabSwitcher] Change selection style to material next Disabled: Selected tab will have a hairline drawable foreground; Enabled: Selected tab will be tinted into primary color. Color in incognito mode will be using the ones for dark mode, and will not support dynamic colors. (The update for thumbnail is not included in this CL) Spec see original bug attached. Bug: 1217141

No includes needing update since this fixes some indirectly added dependencies for the //brave/browser/ui/bookmark GN target. Chromium change: https://source.chromium.org/chromium/chromium/src/+/83b26a03aaacad7853096fb3257d3390f5e0c454 commit 83b26a03aaacad7853096fb3257d3390f5e0c454 Author: Tibor Goldschwendt <tiborg@chromium.org> Date: Fri Jul 9 23:13:18 2021 +0000 [ntp] Move 1P services to //chrome/browser/new_tab_page Before the deprecation of the local NTP, Instant NTPs and the 1P WebUI shared functionality implemented by services in //chrome/browser/search. Since, those NTPs have distinct functionality and sharing of services is no longer necessary. However, the code is still entangled; and NTPs instantiate services they don't need. We regard this as technical debt and are in the process of removing it. To aide this process this CL moves services only required and depended on by the 1P WebUI NTP from //chrome/browser/search to //chrome/browser/new_tab_page. After this CL, some 1P services, such as for backgrounds or promos, will still reside in //chrome/browser/search. These services are not technically used by Instant NTPs anymore but are still entangled with Instant NTP code. We will detangle and move those services in follow-up CLs. Change-Id: Ie66ea17eca89a5556785e1398510e6e0acc6a65c Bug: 1084363

…nt in polymer_overriding

… strings 'rel' is a disallowed attribute for translated strings as per ui/webui/resources/js/parse_html_subset.js

…roperty binding is setup Previously we were creating the elements directly on the document. That led to the constructor being called immediately. When we create as a template element first then we avoid that issue.

…the actual element. We need to: - Make sure chromium's component does not get registered, so we add a dep to a JS module which makes sure the element name is on an ignore list - Define our own element which subclasses chromium's one and accesses and modifies the original static property

…efore brave's overrides have run

….4577.42.

Browser tests error out on Windows CI with content/public/test/test_utils.cc(137): error: Failed RunLoop::Run() timed out. Timeout set at ProxyRunTestOnMainThreadLoop@content/public/test/browser_test_base.cc:800. Running with --disable-gpu avoids the error, but causes various WebGL tests to fail. Running with --no-sandbox avoids the error. Running with --disable-gpu-sandbox also avoids the error - picked this one is the least bad from the 2 above. It is not clear ATM what's causing the error. Only adds the flag on Windows in Release config.

…ponent

Followup to a prior commit that follows upstream and renames ProfileSyncService to SyncServiceImpl in various class names.

…ser#10643 This browser tests got moved into //brave/browser/brave_shields but got removed from its original location, so let's do it now.

This migrates away from WebContentsReceiverSet and avoids having to friend WebContentsFrameReceiverSetPassKey with BraveDRMTabHelper.

…rSet This migrates away from WebContentsReceiverSet and avoids having to friend WebContentsFrameReceiverSetPassKey with BraveShieldsWebContentsObserver.

….cc.patch I initially thought we'd be better off for now with a one-line patch to avoid complicating things too much on the verge of the merge, but turns out that we already subclass ChromeContentBrowserClient, and so we can avoid patching by overriding BindAssociatedReceiverFromFrame() from there to extend it with the binders for our own mojo interfaces.

The patch was introduced here #2514 but upstream added the same functionality in Cr81 here chromium/chromium@432be5c

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Upgrade from Chromium 92 to Chromium 93 #9062

Upgrade from Chromium 92 to Chromium 93 #9062

Commits on Aug 18, 2021