‘usenewprocess‘ attribute for webview allowing to create new process for every webview. #2

janRucka · 2016-02-23T14:33:45Z

Reasons for pull request:

In our project we need to use multiple webviews and share cookies. Also we need new process for every webview. If webviews share process then if one webview is slow-down (rendering new page or even frozen for whatever reason) the other is slow-down as well. Only way (or best way as far as I know) to avoid this is to use new process.

In nw12 we used different partition ID for every webview. Then every webview is run in new process. But this is not a possibility for us anymore since cookies are not shared (if different partition ID is used). I think the behavior in nw12 was more or less buggy because is different from what is described here: https://developer.chrome.com/apps/tags/webview#partition

Here is bachel complain about running in one process as well:
nwjs/nw.js#3622 (comment)

Solution:

We added ‘usenewprocess’ attribute to webview. If this attribute is used then new process for every webview should be started. If attribute is not used behavior should remain the same as is now (new process only if different partition ID is used).

Hence change will only affect users which for whatever reason need to use new process for every webview (and will use ‘usenewprocess’ attribute). Others will not be affected at all.

Examples:

Working only with our chanches in code.
twoWebviewsOneProcess.zip
twoWebviewsTwoProcesses.zip

… (patchset #2 id:20001 of https://codereview.chromium.org/1730153005/ ) Reason for revert: The reason for reverting is: Speculative revert for Compile failure on Win, Mac & Precise 64 Beta Bots (crbug.com/590286) Original issue's description: > [cherrypick] Fix computation of runtime for throttled tasks > > The computation of the delayed task runtime was computing the (virtual) > TimeDomain's now + delay. That doesn't take into account the skew > between the real time and the virtual time, which meant timers could > run sooner than expected for background tabs. Timers associated with > foreground tabs are unaffected. > > This patch makes the computation of delayed runtime the responsibility > of the TimeDomain and introduces a ThrottledTimeDomain which computes > the delayed runtime as realtime + delay. > > This means timers will no longer run sooner than expected. Under normal > circumstancs this only timers whose delay modulo 1000ms is less than 100 > ms are less likely to be affected. Timers whose whose delay modulo > 1000ms is greater than 900 are highly likely to be affected (there's a > good chance their delay was too small before). > > BUG=587074 > > Review URL: https://codereview.chromium.org/1718233002 > > Cr-Commit-Position: refs/heads/master@{#376993} > (cherry picked from commit b56bbe4) > > [email protected] > > Committed: https://chromium.googlesource.com/chromium/src/+/fd563254490ae93382cb88a1e9a1c1e54815c97b [email protected],[email protected] # Skipping CQ checks because original CL landed less than 1 days ago. NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true BUG=587074 Review URL: https://codereview.chromium.org/1743813002 Cr-Commit-Position: refs/branch-heads/2623@{#526} Cr-Branched-From: 92d7753-refs/heads/master@{#369907}

The computation of the delayed task runtime was computing the (virtual) TimeDomain's now + delay. That doesn't take into account the skew between the real time and the virtual time, which meant timers could run sooner than expected for background tabs. Timers associated with foreground tabs are unaffected. This patch makes the computation of delayed runtime the responsibility of the TimeDomain and introduces a ThrottledTimeDomain which computes the delayed runtime as realtime + delay. This means timers will no longer run sooner than expected. Under normal circumstancs this only timers whose delay modulo 1000ms is less than 100 ms are less likely to be affected. Timers whose whose delay modulo 1000ms is greater than 900 are highly likely to be affected (there's a good chance their delay was too small before). BUG=587074 Review URL: https://codereview.chromium.org/1718233002 Cr-Commit-Position: refs/heads/master@{#376993} (cherry picked from commit b56bbe4) [email protected] Review URL: https://codereview.chromium.org/1751953002 . Cr-Commit-Position: refs/branch-heads/2623@{#550} Cr-Branched-From: 92d7753-refs/heads/master@{#369907}

-Attribute 'usenewprocess' added on webview. If used it ensure that webview is started in new process. If not used chrome will decide whether or not new process will be used.

janRucka · 2016-03-04T13:14:04Z

Did you have time to review this Pull request? It is there already for 10 days without any reaction. What is the most annoying for me is to do constant merging with every new commit in nw13 branch. Your workflow (to create patch with changes on actual chromium version) make merge even more time consuming.

rogerwang · 2016-03-04T14:17:03Z

We was in the middle of rebasing to Chromium 49. Will review soon.

janRucka · 2016-03-04T14:17:58Z

Thanks

Cr-Commit-Position: refs/branch-heads/2564@{#2} Cr-Branched-From: 1283eca-refs/heads/master@{#359700}

…rd navigation.' Cherry-picked patchset #1 and #2 from cl/2691693002 BUG=688047, 677356 Review-Url: https://codereview.chromium.org/2685803002 NOTRY=true NOPRESUBMIT=true Review-Url: https://codereview.chromium.org/2711123004 Cr-Commit-Position: refs/branch-heads/2987@{#666} Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…chset #2 id:20001 of https://codereview.chromium.org/2666893002/ ) Reason for revert: First, this has a race: CONTEXT_KEY_HAS_GAIA_ACCOUNT can (and actually does) come after CONTEXT_KEY_PROFILE_PICTURE_DATA_URL thus triggering exception: login.js:1208 Uncaught TypeError: Cannot read property 'title' of undefined at HTMLUnknownElement.updateItem (login.js:1208) at Array.<anonymous> (login.js:9504) at ScreenContext.applyChanges (login.js:538) at HTMLDivElement.contextChanged_ (login.js:329) at Object.api.contextChanged (login.js:424) at <anonymous>:1:23 Second, it breaks supervised user creation screen (and probably other screens that share user-image-grid object). BUG=697842,682236 Original issue's description: > Delete "Google Profile photo" for the Active Directory > > Delete "Google Profile photo" choice from the image choice screen for > the Active Directory users. > > BUG=682236 > TEST=manual > CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation > > Review-Url: https://codereview.chromium.org/2666893002 > Cr-Commit-Position: refs/heads/master@{#447513} > Committed: https://chromium.googlesource.com/chromium/src/+/33f901b2ff76e1d08d80e3ba5d1d079178cb4978 [email protected],[email protected] BUG=682236 CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation Review-Url: https://codereview.chromium.org/2727953004 Cr-Commit-Position: refs/heads/master@{#454525} (cherry picked from commit ed71ca7) Review-Url: https://codereview.chromium.org/2785593002 . Cr-Commit-Position: refs/branch-heads/3029@{#466} Cr-Branched-From: 939b32e-refs/heads/master@{#454471}

Cr-Commit-Position: refs/branch-heads/3071@{#2} Cr-Branched-From: a106f0a-refs/heads/master@{#464641}

… (patchset #4 id:170001 of https://codereview.chromium.org/2809043004/ ) Reason for revert: Causes http://crbug.com/714386 Original issue's description: > Reland of Initialize default audio device ID with explicit device ID. (patchset #1 id:1 of https://codereview.chromium.org/2813543005/ ) > > Reason for revert: > Will attempt to reland by reverting changes in MSM and updating extensions test. > > Original issue's description: > > Revert of Initialize default audio device ID with explicit device ID. (patchset #2 id:20001 of https://codereview.chromium.org/2812903002/ ) > > > > Reason for revert: > > Patchset 2 restores the behavior I wanted to eliminate. > > > > Original issue's description: > > > Initialize default audio device ID with explicit device ID. > > > > > > This is more consistent with how video device IDs are specified and is > > > also a small first step towards implementing the standard constraints > > > algorithm for audio. > > > This CL disables a misleading selector that allows switching the > > > user-preferred device in the middle of a getUserMedia call. > > > Since that selector does not affect the current getUserMedia call, it > > > is better to have it disabled, which is already the case with video > > > devices. > > > > > > BUG=708081 > > > > > > Review-Url: https://codereview.chromium.org/2812903002 > > > Cr-Commit-Position: refs/heads/master@{#463624} > > > Committed: https://chromium.googlesource.com/chromium/src/+/82382cca8c9aa804acb7fd9cfaa6e4478f92cd7d > > > > [email protected] > > # Skipping CQ checks because original CL landed less than 1 days ago. > > NOPRESUBMIT=true > > NOTREECHECKS=true > > NOTRY=true > > BUG=708081 > > > > Review-Url: https://codereview.chromium.org/2813543005 > > Cr-Commit-Position: refs/heads/master@{#463634} > > Committed: https://chromium.googlesource.com/chromium/src/+/cf8ea8d590f1c89c2633c793a41bd31a85b3fe72 > > # Skipping CQ checks because original CL landed less than 1 days ago. > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > BUG=708081 > > Review-Url: https://codereview.chromium.org/2809043004 > Cr-Commit-Position: refs/heads/master@{#463710} > Committed: https://chromium.googlesource.com/chromium/src/+/be26518ed4e8a2966c919a93b488689374b17d42 [email protected] BUG=708081,714386 Review-Url: https://codereview.chromium.org/2829403002 Cr-Commit-Position: refs/heads/master@{#466550} (cherry picked from commit 18633fc) Review-Url: https://codereview.chromium.org/2837873004 . Cr-Commit-Position: refs/branch-heads/3071@{#194} Cr-Branched-From: a106f0a-refs/heads/master@{#464641}

…set #2 id:20001 of https://codereview.chromium.org/2671433002/ ) Reason for revert: This breaks Chrome iOS dashboards because cpu_architecture is the only field where we have the device model info (e.g. iPhone7,2). Original issue's description: > Don't set cpu architecture field on iOS in UMA logs. > > The value we were setting it to before was not correct, so this > change instead just makes the value not get set on iOS. > > BUG=370104 > > Review-Url: https://codereview.chromium.org/2671433002 > Cr-Commit-Position: refs/heads/master@{#447578} > Committed: https://chromium.googlesource.com/chromium/src/+/a86e8e115054ad167458d0379980db5473cb20de [email protected],[email protected] BUG=370104, 718080 Review-Url: https://codereview.chromium.org/2855943005 Cr-Commit-Position: refs/heads/master@{#469062} (cherry picked from commit 9f89092) Review-Url: https://codereview.chromium.org/2863143002 . Cr-Commit-Position: refs/branch-heads/3029@{#803} Cr-Branched-From: 939b32e-refs/heads/master@{#454471}

…set #2 id:20001 of https://codereview.chromium.org/2671433002/ ) Reason for revert: This breaks Chrome iOS dashboards because cpu_architecture is the only field where we have the device model info (e.g. iPhone7,2). Original issue's description: > Don't set cpu architecture field on iOS in UMA logs. > > The value we were setting it to before was not correct, so this > change instead just makes the value not get set on iOS. > > BUG=370104 > > Review-Url: https://codereview.chromium.org/2671433002 > Cr-Commit-Position: refs/heads/master@{#447578} > Committed: https://chromium.googlesource.com/chromium/src/+/a86e8e115054ad167458d0379980db5473cb20de [email protected],[email protected] BUG=370104, 718080 Review-Url: https://codereview.chromium.org/2855943005 Cr-Commit-Position: refs/heads/master@{#469062} (cherry picked from commit 9f89092) Review-Url: https://codereview.chromium.org/2860093004 . Cr-Commit-Position: refs/branch-heads/3071@{#416} Cr-Branched-From: a106f0a-refs/heads/master@{#464641}

…irst use. (patchset #2 id:40001 of https://codereview.chromium.org/2871843003/ ) Reason for revert: The perf regression was not caused by this piece of code. Reverting to the original behavior. See https://docs.google.com/document/d/1eeTSJg18rJ4v_oxxSJU8MK20UywnEhocWk_WrWv8kQE/edit for details of the regression investigation. Original issue's description: > [translate] Defer start of translate ranker model load to first use. > > This CL is in response to perf regression (see referenced bugs). > > BUG=697947, 697665 > > Review-Url: https://codereview.chromium.org/2871843003 > Cr-Commit-Position: refs/heads/master@{#471086} > Committed: https://chromium.googlesource.com/chromium/src/+/d2876f4ee1c1a01ee262ca55752503ba3a30a6d5 [email protected],[email protected],[email protected],[email protected],[email protected],[email protected] BUG=697947, 697665 Review-Url: https://codereview.chromium.org/2918173002 Cr-Original-Commit-Position: refs/heads/master@{#478379} Review-Url: https://codereview.chromium.org/2956053003 . Cr-Commit-Position: refs/branch-heads/3112@{#481} Cr-Branched-From: b6460e2-refs/heads/master@{#474897}

This reverts commit 1b60006. Reason for revert: Causes a startup crash with chrome --mash, possibly because mash does not have a DeviceDataManager. You might want to check with sadrul@ or rjkroege@ about whether there's a good way to do this under mash. Failing build (chromeos-side waterfall): https://luci-milo.appspot.com/buildbot/chromeos.chrome/tricky-tot-chrome-pfq-informational/5205 Stack (from running manually with chrome --mash on Linux): [60089:60089:0717/114019.863351:FATAL:device_data_manager.cc(84)] Check failed: instance_. DeviceDataManager was not created. #0 0x7f06a03468bc base::debug::StackTrace::StackTrace() #1 0x7f06a036a121 logging::LogMessage::~LogMessage() #2 0x7f069aeb08ff ui::DeviceDataManager::GetInstance() #3 0x55f183402a41 chromeos::LoginDisplayHostImpl::LoginDisplayHostImpl() #4 0x55f1834066ad chromeos::ShowLoginWizard() #5 0x55f1833e4bf7 chromeos::ChromeSessionManager::Initialize() #6 0x55f1832f9cb4 chromeos::ChromeBrowserMainPartsChromeos::PostProfileInit() #7 0x55f1837e7aa2 ChromeBrowserMainParts::PreMainMessageLoopRunImpl() #8 0x55f1837e709d ChromeBrowserMainParts::PreMainMessageLoopRun() #9 0x55f1832f8f4a chromeos::ChromeBrowserMainPartsChromeos::PreMainMessageLoopRun() #10 0x7f069d9e5081 content::BrowserMainLoop::PreMainMessageLoopRun() #11 0x7f069dde6407 content::StartupTaskRunner::RunAllTasksNow() #12 0x7f069d9e353b content::BrowserMainLoop::CreateStartupTasks() #13 0x7f069d9e7a72 content::BrowserMainRunnerImpl::Initialize() #14 0x7f069d9e0b57 content::BrowserMain() #15 0x7f069e1e8462 content::ContentMainRunnerImpl::Run() #16 0x7f06a0890d39 service_manager::Main() #17 0x7f069e1e7384 content::ContentMain() #18 0x55f182e8132f ChromeMain #19 0x7f069454af45 __libc_start_main #20 0x55f182e81194 <unknown> Original change's description: > Listen to changes to touch input devices > > In https://codereview.chromium.org/2964823002 the OobeDisplayChooser > started using the DeviceDataManager to look for touchscreen devices when > searching for a good primary display to use during OOBE. > > On device cold boot the DeviceDataManager has not yet found any > touchscreen devices at the time OobeUi::ShowOobeUI() is called (likely > due to lower level systems not being fully initialized). > > This CL make LoginDisplayHostImpl an observer of changes to connected > touchscreen devices, re-triggering the OobeDisplayChooser when the > DeviceDataManager is notified of the connected touchscreens. This > overcomes the timing issues on cold boot. > > Bug: 738885 > Change-Id: Iae488ddc9428b7c5e74d36cf18e35ba3d1235bbd > Reviewed-on: https://chromium-review.googlesource.com/569958 > Reviewed-by: Jacob Dufault <[email protected]> > Commit-Queue: Felix Ekblom <[email protected]> > Cr-Commit-Position: refs/heads/master@{#487007} [email protected],[email protected],[email protected] # Not skipping CQ checks because original CL landed > 1 day ago. (cherry picked from commit 0cd134a) Bug: 738885 Change-Id: If31322734e679bbb1f4eef0a9aa802d34263cba4 Reviewed-on: https://chromium-review.googlesource.com/574731 Reviewed-by: James Cook <[email protected]> Commit-Queue: James Cook <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#487191} Signed-off-by: Bernie Thompson <[email protected]> Reviewed-on: https://chromium-review.googlesource.com/575149 Reviewed-by: Bernie Thompson <[email protected]> Cr-Commit-Position: refs/branch-heads/3112@{#623} Cr-Branched-From: b6460e2-refs/heads/master@{#474897}

[email protected] Change-Id: Ifbffc39df9024c37350df83abc7b715e52d277f9 Reviewed-on: https://chromium-review.googlesource.com/580748 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3163@{#2} Cr-Branched-From: ff259ba-refs/heads/master@{#488528}

changes: new specs require removing suggested apps indicator, specs: peeking: https://screenshot.googleplex.com/QNw4VjLHD5e fullscreen page #1: https://screenshot.googleplex.com/OT0NmfnByWg fullscreen page #2: https://screenshot.googleplex.com/7PC0iytVXHq fullscreen page#01: https://screenshot.googleplex.com/NJiyTFFzqKW fullscreen page#02: https://screenshot.googleplex.com/8Dh1Z1YC3aA [email protected] (cherry picked from commit dce53c5) screenshot: peeking: https://screenshot.googleplex.com/tMxgChqhrSm Test: test with fullscreen app list flag enabled Bug: 750292 Change-Id: I841342e95101e54ae7e022383ecbfbe56c503e2d Reviewed-on: https://chromium-review.googlesource.com/606997 Reviewed-by: Xiyuan Xia <[email protected]> Commit-Queue: Qiang(Joe) Xu <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#492775} Reviewed-on: https://chromium-review.googlesource.com/611300 Reviewed-by: Vadim Tryshev <[email protected]> Cr-Commit-Position: refs/branch-heads/3163@{#468} Cr-Branched-From: ff259ba-refs/heads/master@{#488528}

It's possible for BrowserAssociatedInterface's InternalState to have Initialize invoked after ClearBindings, which can in turn lead to a violation of the assumption that ClearBindings ensures no future message dispatches to the bound implementation. This can occur in the following rare but plausible scenario for a type X which inherits BrowserAssociatedInterface<T>: 1. Post task to IO thread which may destroy some yet-uncreated instance x of type X (e.g. maybe it posts a render process ID and we'll later have an X associated with that RPH) 2. Create x as an instance of type X (posts an IO thread task to Initialize x's Internal State) 3. Interface request is received, posting an IO thread task to bind a handle to x. 3. Task from #1 executes, locating and deleting x; resetting InternalState bindings (not yet initialized anyway). 4. Task from #2 executes, initializing the InternalState bindings. 5. Task from #3 executes, bindings a handle to the InternalState 6. Message is received on the bound handle, dispatched to deleted x 7. UAF! As it turns out, AssociatedBindingSet does not own any thread-affine state upon default construction; therefore it can be created on any thread as long as it's subsequently used and destroyed from a single thread. Since InternalState already guarantees those conditions, this CL simply removes the async Initialize step. BUG=753672 [email protected] [email protected] (cherry picked from commit 201a233) Change-Id: Id603fb57d412daf3741b61f7857a29edeaac5443 Reviewed-on: https://chromium-review.googlesource.com/610924 Reviewed-by: John Abd-El-Malek <[email protected]> Commit-Queue: Ken Rockot <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#493797} Reviewed-on: https://chromium-review.googlesource.com/617516 Reviewed-by: Ken Rockot <[email protected]> Cr-Commit-Position: refs/branch-heads/3163@{#606} Cr-Branched-From: ff259ba-refs/heads/master@{#488528}

[email protected] Change-Id: Ia12160b836ddc9bfbcf7cfe2ccc709ef0002779a Reviewed-on: https://chromium-review.googlesource.com/647208 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3202@{#2} Cr-Branched-From: fa6a5d8-refs/heads/master@{#499098}

[email protected] Change-Id: Ibf4495b3dddcfadd1e7a79ecd906d9ccaad72759 Reviewed-on: https://chromium-review.googlesource.com/718136 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3239@{#2} Cr-Branched-From: adb61db-refs/heads/master@{#508578}

When using --isolate-origins=... or --site-per-process, some or all sites will be given dedicated processes. Documents (HTML, XML, and some text files) and other opaque formats (JSON) from such sites should not be delivered to cross-site pages in the renderer process unless they are made available via CORS or another exception. This CL adds a ResourceHandler to enforce this restriction in the browser process. The handler inspects the response headers and decides whether to block it from reaching the renderer process. It also sniffs the content to confirm the response is labeled correctly, to avoid blocking things like JavaScript mislabeled as HTML or JSON. Blocked responses are sent back as empty response bodies rather than as a network error, to ensure existing behavior does not change. Other cross-site responses (e.g., images, scripts, etc) are still allowed. Note that the current approach assumes the renderer process is not compromised and will not lie about the initiating origin, etc. The need for assumption will be removed in future work, for bug 268640. For more details, see: http://www.chromium.org/developers/design-documents/blocking-cross-site-documents BUG=786505 [email protected] (cherry picked from commit 358baf4) (cherry picked from commit eff7652786019f8a24edd269867223ee12df7389) Cq-Include-Trybots: master.tryserver.chromium.linux:linux_site_isolation Reviewed-on: https://chromium-review.googlesource.com/783826 Commit-Queue: Charlie Reis <[email protected]> Reviewed-by: Steven Holte <[email protected]> Reviewed-by: Ilya Sherman <[email protected]> Reviewed-by: Daniel Cheng <[email protected]> Reviewed-by: Nick Carter <[email protected]> Reviewed-by: Charlie Harrison <[email protected]> Reviewed-by: Devlin <[email protected]> Reviewed-by: Mike West <[email protected]> Reviewed-by: Tsuyoshi Horo <[email protected]> Cr-Original-Original-Commit-Position: refs/heads/master@{#522016} Reviewed-on: https://chromium-review.googlesource.com/818491 Reviewed-by: Charlie Reis <[email protected]> Cr-Original-Commit-Position: refs/branch-heads/3239_84@{#2} Cr-Original-Branched-From: 8f51ed0-refs/branch-heads/3239@{#643} Cr-Original-Branched-From: adb61db-refs/heads/master@{#508578} Change-Id: I7f7b78714dd36727ccf41be79f666a61f174995c Reviewed-on: https://chromium-review.googlesource.com/823283 Cr-Commit-Position: refs/branch-heads/3239@{#668} Cr-Branched-From: adb61db-refs/heads/master@{#508578}

[email protected] Change-Id: I2b2fcc1d7b8f6dd22fdc7d807428f109d30adbf1 Reviewed-on: https://chromium-review.googlesource.com/874902 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3325@{#2} Cr-Branched-From: bc084a8-refs/heads/master@{#530369}

[email protected] Change-Id: Ib15c0a5aeb49f1134f0984dbecbaf4f9185e04ac Reviewed-on: https://chromium-review.googlesource.com/945449 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3359@{#2} Cr-Branched-From: 66afc5e-refs/heads/master@{#540276}

Bug:831762,783383 [email protected] (cherry picked from commit 787bde4) Change-Id: I3aea3afb2669f83f0d40674218c82a4f399b4e6f Reviewed-on: https://chromium-review.googlesource.com/1008270 Commit-Queue: Chris Harrelson <[email protected]> Reviewed-by: Stefan Zager <[email protected]> Reviewed-by: Alexei Svitkine <[email protected]> Reviewed-by: vmpstr <[email protected]> Reviewed-by: Mark Pearson <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#550518} Reviewed-on: https://chromium-review.googlesource.com/1012570 Reviewed-by: Chris Harrelson <[email protected]> Cr-Commit-Position: refs/branch-heads/3396@{#2} Cr-Branched-From: 9ef2aa8-refs/heads/master@{#550428}

[email protected] Change-Id: Ic0dfa4e19ec9328a162a535ad1990cab3ce69996 Reviewed-on: https://chromium-review.googlesource.com/1072325 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3440@{#2} Cr-Branched-From: 010ddcf-refs/heads/master@{#561733}

[email protected] Change-Id: I9321fc52b8b4731488b4a3ce2dc7214e6448f784 Reviewed-on: https://chromium-review.googlesource.com/1144494 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3497@{#2} Cr-Branched-From: 271eaf5-refs/heads/master@{#576753}

[email protected] Change-Id: I5488772ea9b1e107a39dd776258413833ad56ac8 Reviewed-on: https://chromium-review.googlesource.com/1198315 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3538@{#2} Cr-Branched-From: 79f7c91-refs/heads/master@{#587811}

This CL is a collection of cherry-picks related to crbug.com/868592 fix. Specifically, this is: - The original mega-patch crrev.com/222c9ba7c6 - creis@ follow-up fix crrev.com/27986c7c955 - kouhei@ follow-up fix crrev.com/6be8b5a07bdf The original change descriptions are captured below % Change-Id lines --- Speculative crash fix for navigator.serviceworker access during unload This should fix crash/caab6eb137e58385 This CL addresses the unhandled case in crrev.com/582126 [email protected] Bug: 881126, 868592 Reviewed-on: https://chromium-review.googlesource.com/1207781 Commit-Queue: Kouhei Ueno <[email protected]> Reviewed-by: Kinuko Yasuda <[email protected]> Reviewed-by: Nasko Oskov <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#589419}(cherry picked from commit 6be8b5a) Reviewed-on: https://chromium-review.googlesource.com/1212368 Reviewed-by: Kouhei Ueno <[email protected]> Cr-Commit-Position: refs/branch-heads/3545@{#2} Cr-Branched-From: a2bbe9d-refs/heads/master@{#589377} Speculative fix for additional History DocumentLoader crashes. There is no guarantee that the DocumentLoader is always attached [1], so let's introduce null checks in StateInternal and setScrollRestoration. [1] The DocumentLoader may be detached while FrameLoader::PrepareForCommit. BUG=879477, 872672 Reviewed-on: https://chromium-review.googlesource.com/1200075 Commit-Queue: Charlie Reis <[email protected]> Reviewed-by: Nate Chapin <[email protected]> Cr-Commit-Position: refs/heads/master@{#588227} Speculative fix for History::ScrollRestorationInternal null deref This is a speculative fix for crash reported on crbug.com/872672 . There is no guarantee that the DocumentLoader is always attached [1], so let's introduce a null check. [1] The DocumentLoader may be detached while FrameLoader::PrepareForCommit. Bug: 872672 Reviewed-on: https://chromium-review.googlesource.com/1171972 Reviewed-by: Kinuko Yasuda <[email protected]> Commit-Queue: Kouhei Ueno <[email protected]> Cr-Commit-Position: refs/heads/master@{#582509} (cherry picked from commit a2a107d) NavigatorServiceWorker: Avoid instantiating if being navigated away. This CL fixes a clusterfuzz crash which fails to minimize. Bug: 872320 Reviewed-on: https://chromium-review.googlesource.com/1170160 Reviewed-by: Hiroki Nakagawa <[email protected]> Reviewed-by: Matt Falkenhagen <[email protected]> Reviewed-by: Kinuko Yasuda <[email protected]> Commit-Queue: Kouhei Ueno <[email protected]> Cr-Commit-Position: refs/heads/master@{#582126} (cherry picked from commit d995adf) Flush microtask queue before commit Bug: 868592 Reviewed-on: https://chromium-review.googlesource.com/1164148 Reviewed-by: Yutaka Hirano <[email protected]> Commit-Queue: Kouhei Ueno <[email protected]> Cr-Commit-Position: refs/heads/master@{#581124} (cherry picked from commit 38894f6) Prevent promise reject to be sync scheduled during DocumentLoader detach (% mod: revert fetch_manager.cc change) (cherry picked from commit c415acc) Bug: 868592 Change-Id: I50029416f0441a9f09c538716684a01cb8af93e1 Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_slimming_paint_v2;master.tryserver.blink:linux_trusty_blink_rel Reviewed-on: https://chromium-review.googlesource.com/1163235 Reviewed-by: Yuki Shiino <[email protected]> Reviewed-by: Yutaka Hirano <[email protected]> Commit-Queue: Kouhei Ueno <[email protected]> Cr-Original-Original-Commit-Position: refs/heads/master@{#580814} Reviewed-on: https://chromium-review.googlesource.com/1184122 Cr-Original-Commit-Position: refs/branch-heads/3497@{#760} Cr-Original-Branched-From: 271eaf5-refs/heads/master@{#576753} Reviewed-on: https://chromium-review.googlesource.com/1218183 Reviewed-by: Kouhei Ueno <[email protected]> Reviewed-by: Kentaro Hara <[email protected]> Reviewed-by: Charlie Reis <[email protected]> Cr-Commit-Position: refs/branch-heads/3497@{#938} Cr-Branched-From: 271eaf5-refs/heads/master@{#576753}

rogerwang · 2018-09-26T01:50:01Z

In the new site isolation architecture in upstream, webview now runs in separate process. So I'm closing this.

[email protected] Change-Id: Ic56dc53e7abf1fdb6a44a10e31ebfc53aecd86e6 Reviewed-on: https://chromium-review.googlesource.com/c/1278330 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3578@{#2} Cr-Branched-From: 4226ddf-refs/heads/master@{#599034}

Pressing the skip button leads to two LoginDisplayHost::Finalize() calls, which causes a double deletion eventually leading to a crash. The PIN setup skip button triggers two 'discover-done' events (fired in discover_ui.js::end_), and the observer in screen_discover.js sends a finished user action whenver it receives a discover-done event. Discover should not be sending discover-done twice. This CL just prevents the crash. Finalize #1 at HTMLElement.end_ at HTMLElement.fire at HTMLElement.onSkipButton_ at HTMLElement.onAuthTokenChanged_ at HTMLElement._observerEffect at HTMLElement._effectEffects at HTMLElement._propertySetter at HTMLElement.setter at HTMLElement.onSkipButton_ at HTMLElement.handler Finalize #2 at HTMLElement.end_ at HTMLElement.fire at HTMLElement.onSkipButton_ at HTMLElement.handler at HTMLElement.fire at Object.fire at Object.forward at Object.click at HTMLElement.handleNative [email protected] (cherry picked from commit 41254eb) Bug: 895250 Change-Id: I784ee5ef0ecfc5bffad515973451331bfc0aee3f Reviewed-on: https://chromium-review.googlesource.com/c/1299145 Reviewed-by: Alexander Alekseev <[email protected]> Commit-Queue: Jacob Dufault <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#602880} Reviewed-on: https://chromium-review.googlesource.com/c/1308414 Reviewed-by: Jacob Dufault <[email protected]> Cr-Commit-Position: refs/branch-heads/3578@{#409} Cr-Branched-From: 4226ddf-refs/heads/master@{#599034}

…Mojave This change is required due to new security restrictions in Mojave. We can no longer inject input w/o being added as an accessibility app in the security applet. While this sounds like a usefulk speedbump, it causes remote access applications quite a bit of trouble: 1.) We don't use the restricted API until a user connects so they cannot approve remotely 2.) The dialog appears to only show up once (regardless of approve/deny status) 3.) Users connecting to a locked machine will never see the dialog This is affecting quite a few CRD users, basically everyone who upgrades to Mojave will experience this one way or another. This is the simplest fix (and easiest to merge) that I could think of to unblock users. The prompt will only be shown on 10.14+ platforms and the request is only shown if the app has not been approved. I'd like to look at the user feedback after releasing this change to see if more work is needed. One problem I anticipate is that the dialog shown doesn't have a lot of context and it refers to the wrapper script (org.chromium.chromoting.me2me.sh) instead of Chrome Remote Desktop. If this is confusing, we can wrap the prompt request in a dialog where we control the text. My concern with checking in the feature first is that the new strings won't be available for merging. Another behavior to call about this impl is that the prompt will be displayed in two instances: 1.) When the host is first started (choosing enable via app/website) 2.) When the user signs in and the host service is started Scenario #2 will have less context but that is the only way to ask for permission for users who upgraded and had CRD installed previously. The dialog is not displayed at the login screen (i.e. when no one is signed in). One last note, there is no way that I can see to specify this permission in the manifest or set up via a script / at install time. It requires a user action to complete. Bug: 901021 Change-Id: I54247dd56111feee9608e1b50387cb3501a659a4 Reviewed-on: https://chromium-review.googlesource.com/c/1315468 Reviewed-by: Gary Kacmarcik <[email protected]> Commit-Queue: Joe Downing <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#604947}(cherry picked from commit 19e52db) Reviewed-on: https://chromium-review.googlesource.com/c/1315706 Reviewed-by: Joe Downing <[email protected]> Cr-Commit-Position: refs/branch-heads/3578@{#478} Cr-Branched-From: 4226ddf-refs/heads/master@{#599034}

[email protected] Change-Id: Ic479ff470024e8f324896b71e9a89153da3a2350 Reviewed-on: https://chromium-review.googlesource.com/c/1356267 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3626@{#2} Cr-Branched-From: d897fb1-refs/heads/master@{#612437}

Background: My first attempt [1] to implement KeyboardFocusableScrollers made scrollers click-focusable. This got reported as a regression in Issue 908809. [1] https://chromium-review.googlesource.com/c/chromium/src/+/1258331. This happens when HandleMouseFocus() uses SupportsFocus(): Element::SupportsFocus() <--- Element::IsMouseFocusable() Element::IsFocusable() MouseEventManager::HandleMouseFocus() EventHandler::HandleMousePressEvent() Problem: After [1], SupportsFocus() and therefore also IsMouseFocusable() returns true for all scrollable scrollers. If a clicked element is "mouse focusable", HandleMouseFocus() calls FocusController:: SetFocusedElement(element) which makes |element| the activeElement. Changing activeElement on clicks is reported as a regression. Solution: This CL reverts [1]'s change in SupportFocus() and instead decouples IsMouseFocusable() and IsKeyboardFocusable(). Now we're back to the old mouse focus semantics: clicking a scroller does *not* make it the activeElement, does not trigger :focus CSS and does not target the scroller with a focus-event. However, tabbing to a scroller *does* all that. That's the feature of KeyboardFocusableScrollers. This CL does not add any new tests, instead, it reverts some of [1]'s test changes so tests again expect that clicking or tapping a scroller doesn't move focus. Bug: 585413, 908809 Change-Id: Ia0f41cb2e0bf2c093eb2c6d6fa15af5d00aa69c9 Reviewed-on: https://chromium-review.googlesource.com/c/1354450 Reviewed-by: Dave Tapuska <[email protected]> Commit-Queue: Hugo Holgersson <[email protected]> Cr-Original-Commit-Position: refs/heads/master@{#613920}(cherry picked from commit d01359a) Reviewed-on: https://chromium-review.googlesource.com/c/1370171 Cr-Commit-Position: refs/branch-heads/3626@{#223} Cr-Branched-From: d897fb1-refs/heads/master@{#612437}

[email protected] Change-Id: Ib6b8d1056f9600b2a553341d5a95cafa8a4f2f01 Reviewed-on: https://chromium-review.googlesource.com/c/1436410 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3683@{#2} Cr-Branched-From: e510299-refs/heads/master@{#625896}

[email protected] Change-Id: I5645af3484987dd8bf93bd2efab40f9854fdc98a Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1511013 Reviewed-by: [email protected] <[email protected]> Cr-Commit-Position: refs/branch-heads/3729@{#2} Cr-Branched-From: d4a8972-refs/heads/master@{#638880}

rogerwang force-pushed the nw13 branch 2 times, most recently from d77fa71 to 4bbe944 Compare February 24, 2016 08:54

Publish DEPS for Chromium 49.0.2623.75

9909a1c

rogerwang force-pushed the nw13 branch 3 times, most recently from 9285daa to f255843 Compare March 3, 2016 00:17

nwjs master

d689b34

rogerwang force-pushed the nw13 branch from f255843 to d689b34 Compare March 3, 2016 06:51

Allow create webview in new process.

e76413d

-Attribute 'usenewprocess' added on webview. If used it ensure that webview is started in new process. If not used chrome will decide whether or not new process will be used.

rogerwang force-pushed the nw13 branch from d689b34 to 7c92241 Compare March 4, 2016 02:35

nwjs master

242db83

rogerwang force-pushed the nw13 branch from 7c92241 to 242db83 Compare March 4, 2016 03:01

Merge branch 'nw13' into cookiesProcess49

e366af2

janRucka force-pushed the cookiesProcess branch from ed52c7b to e366af2 Compare March 4, 2016 13:07

rogerwang force-pushed the nw13 branch from 242db83 to 96977b3 Compare March 5, 2016 01:06

nwjs master

3edfd64

rogerwang force-pushed the nw13 branch from 96977b3 to 3edfd64 Compare March 7, 2016 02:19

jtg-gg referenced this pull request in jtg-gg/chromium.src Mar 7, 2016

Incrementing VERSION to 48.0.2564.2

1d73ad1

Cr-Commit-Position: refs/branch-heads/2564@{#2} Cr-Branched-From: 1283eca-refs/heads/master@{#359700}

Merge branch 'nw13' into cookiesProcess49

7b7f975

rogerwang force-pushed the nw13 branch from 3edfd64 to 2c53f2b Compare March 7, 2016 12:25

Publish DEPS for Chromium 49.0.2623.87

14be212

rogerwang force-pushed the nw13 branch from 2c53f2b to 43cb767 Compare March 8, 2016 06:27

nwjs master

7fee480

rogerwang force-pushed the nw13 branch from 43cb767 to 7fee480 Compare March 8, 2016 16:20

Merge branch 'nw13' into cookiesProcess49

4c8d365

GnorTech pushed a commit that referenced this pull request Apr 26, 2017

Incrementing VERSION to 59.0.3071.2

7825827

Cr-Commit-Position: refs/branch-heads/3071@{#2} Cr-Branched-From: a106f0a-refs/heads/master@{#464641}

rogerwang closed this Sep 26, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

‘usenewprocess‘ attribute for webview allowing to create new process for every webview. #2

‘usenewprocess‘ attribute for webview allowing to create new process for every webview. #2

janRucka commented Feb 23, 2016

janRucka commented Mar 4, 2016

rogerwang commented Mar 4, 2016

janRucka commented Mar 4, 2016

rogerwang commented Sep 26, 2018

‘usenewprocess‘ attribute for webview allowing to create new process for every webview. #2

‘usenewprocess‘ attribute for webview allowing to create new process for every webview. #2

Conversation

janRucka commented Feb 23, 2016

Reasons for pull request:

Solution:

Examples:

janRucka commented Mar 4, 2016

rogerwang commented Mar 4, 2016

janRucka commented Mar 4, 2016

rogerwang commented Sep 26, 2018