View becomes blank after click a cross-site download link when process-swap on navigation is disabled.
<rdar://problem/47099573>
Created attachment 358934 [details] Prototype Passes all ProcessSwap API tests in debug and fixes the cross-site download case.
Created attachment 358950 [details] Prototype
Attachment 358950 [details] did not pass style-queue: ERROR: /Volumes/Data/StyleQueue/WebKit/LayoutTests/TestExpectations:2912: Duplicate or ambiguous entry lines LayoutTests/TestExpectations:130 and LayoutTests/TestExpectations:2912. [test/expectations] [5] Total errors found: 1 in 22 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 358953 [details] Prototype
Created attachment 358960 [details] Prototype (with test)
Attachment 358960 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1470: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1472: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1474: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 21 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 358961 [details] Prototype (with test)
Attachment 358961 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1470: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1472: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1474: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 21 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359047 [details] Patch
Attachment 359047 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359071 [details] Patch
Attachment 359071 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359086 [details] Patch
Attachment 359086 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359088 [details] Patch
Attachment 359088 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359092 [details] Patch
Attachment 359092 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359119 [details] Patch
Attachment 359119 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 25 files If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 359119 [details] Patch Nevermind, I found breakage.
Comment on attachment 359119 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359119&action=review > Source/WebKit/ChangeLog:77 > + Add new ProvisionalPageProxy class to wrap the provisional load in the new process > + after a swap. The provisional page is owned by the WebPageProxy and we only commit > + the provisional page when the load is committed. Until then, the WebPageProxy keeps > + using the old process and displaying the current content. Adding another proxy with message handling etc is bit unfortunate. I suppose this ends up cleaner than just expanding/renaming SuspendedPageProxy? > Source/WebKit/WebProcess/WebPage/WebPage.cpp:475 > - if (parameters.isSwapFromSuspended) > + if (parameters.isProcessSwap) > freezeLayerTree(LayerTreeFreezeReason::SwapFromSuspended); Could rename the LayerTreeFreezeReason as well.
Comment on attachment 359119 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359119&action=review >> Source/WebKit/ChangeLog:77 >> + using the old process and displaying the current content. > > Adding another proxy with message handling etc is bit unfortunate. I suppose this ends up cleaner than just expanding/renaming SuspendedPageProxy? There really isn't that much duplicated logic between SuspendedPageProxy and ProvisionalPageProxy though. The 2 classes almost have opposite purposes. Maybe I could introduce a base class and see what could be shared? Maybe some of the IPC related logic?
Comment on attachment 359119 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359119&action=review >> Source/WebKit/WebProcess/WebPage/WebPage.cpp:475 >> freezeLayerTree(LayerTreeFreezeReason::SwapFromSuspended); > > Could rename the LayerTreeFreezeReason as well. Oh, makes sense. Will do.
Created attachment 359163 [details] Patch
Attachment 359163 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359165 [details] Patch
Attachment 359165 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] Total errors found: 3 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359166 [details] Patch
Attachment 359166 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359168 [details] Patch
Attachment 359168 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359169 [details] Patch
Attachment 359169 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 359169 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359169&action=review r=me Nice and risky, just like a good old JSC patch! > Source/WebKit/UIProcess/WebPageProxy.cpp:554 > +DrawingAreaProxy* WebPageProxy::drawingArea() const > +{ > + if (m_provisionalPage && m_process.ptr() == &m_provisionalPage->process() && m_provisionalPage->drawingArea()) > + return m_provisionalPage->drawingArea(); > + return m_drawingArea.get(); > +} Why do we want to return the drawing area of the provisional page? My understanding is that we prefer to "lock the layer tree" and avoid drawing a page until it becomes committed. > Source/WebKit/UIProcess/WebPageProxy.cpp:3886 > + // The load did not fail, it merely happening in a new provisional process. it is merely happening
Comment on attachment 359169 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=359169&action=review >> Source/WebKit/UIProcess/WebPageProxy.cpp:554 >> +} > > Why do we want to return the drawing area of the provisional page? My understanding is that we prefer to "lock the layer tree" and avoid drawing a page until it becomes committed. I had to make this change for the swipe gesture snapshot logic. A drawing area is associated to a particular process and it is important for WebPageProxy::drawingArea() and WebPageProxy::process() to be in sync to the ViewGestureController logic to work.
Created attachment 359176 [details] Patch
Attachment 359176 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 26 files If any of these errors are false positives, please file a bug against check-webkit-style.
The commit-queue encountered the following flaky tests while processing attachment 359176 [details]: fetch/fetch-worker-crash.html bug 187257 (author: youennf@gmail.com) The commit-queue is continuing to process your patch.
Comment on attachment 359176 [details] Patch Clearing flags on attachment: 359176 Committed r239993: <https://trac.webkit.org/changeset/239993>
All reviewed patches have been landed. Closing bug.
Follow-up iOS build fix: <https://trac.webkit.org/changeset/239995>
Another follow-up build fix for iOS: http://trac.webkit.org/r239997
Reverted r239993, r239995, r239997, and r239999 for reason: Caused assertions under ViewGestureController::disconnectFromProcess() Committed r240015: <https://trac.webkit.org/changeset/240015>
Created attachment 359256 [details] WIP Patch Working on a more robust iteration which does not require temporarily overriding WebPageProxy::m_process in various places. Still needs more testing though.
Attachment 359256 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 47 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359269 [details] WIP Patch
Thanks for the patch. If this patch contains new public API please make sure it follows the guidelines for new WebKit2 GTK+ API. See http://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API
Attachment 359269 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 52 files If any of these errors are false positives, please file a bug against check-webkit-style.
Created attachment 359274 [details] Patch
Attachment 359274 [details] did not pass style-queue: ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1476: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1478: Weird number of spaces at line-start. Are you using a 4-space indent? [whitespace/indent] [3] ERROR: Tools/TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:1480: Multi-line string ("...") found. This lint script doesn't do well with such strings, and may give bogus warnings. They're ugly and unnecessary, and you should use concatenation instead". [readability/multiline_string] [5] ERROR: Source/WebKit/UIProcess/WebProcessProxy.cpp:707: More than one command on the same line [whitespace/newline] [4] Total errors found: 4 in 57 files If any of these errors are false positives, please file a bug against check-webkit-style.
Comment on attachment 359274 [details] Patch r=me
Comment on attachment 359274 [details] Patch Clearing flags on attachment: 359274 Committed r240046: <https://trac.webkit.org/changeset/240046>
Caused some non-PSON API tests failures: Test suite failed Crashed TestWebKitAPI.WebKit.ResizeWindowAfterCrash _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. libc++abi.dylib: Pure virtual function called! TestWebKitAPI.WebKit.LoadPageAfterCrash _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. libc++abi.dylib: Pure virtual function called! TestWebKitAPI.WebKit.EnumerateDevices _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. libc++abi.dylib: Pure virtual function called! TestWebKitAPI.WKNavigation.FailureToStartWebProcessAfterCrashRecovery _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. libc++abi.dylib: Pure virtual function called! TestWebKitAPI.WebKit.ReloadPageAfterCrash _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. libc++abi.dylib: Pure virtual function called! Timeout TestWebKitAPI.WKNavigation.FailureToStartWebProcessRecovery _RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL. Investigating to see if I can quickly fix it or if I need to roll out.
(In reply to Chris Dumez from comment #57) > Caused some non-PSON API tests failures: > Test suite failed > > Crashed > > TestWebKitAPI.WebKit.ResizeWindowAfterCrash > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > libc++abi.dylib: Pure virtual function called! > > TestWebKitAPI.WebKit.LoadPageAfterCrash > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > libc++abi.dylib: Pure virtual function called! > > TestWebKitAPI.WebKit.EnumerateDevices > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > libc++abi.dylib: Pure virtual function called! > > TestWebKitAPI.WKNavigation.FailureToStartWebProcessAfterCrashRecovery > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > libc++abi.dylib: Pure virtual function called! > > TestWebKitAPI.WebKit.ReloadPageAfterCrash > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > libc++abi.dylib: Pure virtual function called! > > Timeout > > TestWebKitAPI.WKNavigation.FailureToStartWebProcessRecovery > _RegisterApplication(), FAILED TO establish the default connection > to the WindowServer, _CGSDefaultConnection() is NULL. > > Investigating to see if I can quickly fix it or if I need to roll out. Fixing those crashes / timeout via https://bugs.webkit.org/show_bug.cgi?id=193509.