NEW 128979
[Win] Some failures in inspector-protocol tests
https://bugs.webkit.org/show_bug.cgi?id=128979
Summary [Win] Some failures in inspector-protocol tests
Brent Fulgham
Reported 2014-02-18 09:53:25 PST
I noticed that the inspector-protocol test suite had been turned off on Windows, and never revisited. Nearly all tests pass, but a handful have slightly different output than the Mac baseline: inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local.html inspector-protocol/debugger/setBreakpoint-dfg.html inspector-protocol/heap-profiler/heap-snapshot-with-detached-dom-tree.html inspector-protocol/heap-profiler/heap-snapshot-with-event-listener.html inspector-protocol/heap-profiler/take-heap-snapshot.html Flaky tests (failed the first run and passed on retry) (2): flag all inspector-protocol/model/content-node-region-info.html inspector-protocol/debugger/pause-on-assert.html
Attachments
Radar WebKit Bug Importer
Comment 1 2014-02-18 09:53:49 PST
Joseph Pecoraro
Comment 2 2014-02-18 10:49:47 PST
Where can we find the differences? Could attach a diff, or link to a bot's results?
Brent Fulgham
Comment 3 2014-02-18 11:10:57 PST
I'll be pushing some updates to the bots today that will start showing these failures. In the meantime, I'll capture the local results and add them to this bug.
Brent Fulgham
Comment 4 2014-02-18 14:30:11 PST
1. Failure for inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local.html: It just looks like a difference in the order of operations. --- /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local-expected.txt +++ /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/debugger/setBreakpoint-dfg-and-modify-local-actual.txt @@ -7,8 +7,8 @@ Breakpoint set in breakpointBasic() Hit Breakpoint! Evaluating in DFG frame at frame[1]: 'x = 20000;' -Resumed from breakpoint inside breakpointBasic PASS: result is 20003 +Resumed from breakpoint Test complete 2. Failure for inspector-protocol/debugger/setBreakpoint-dfg.html: Again, it just looks like a difference in the order of operations. --- /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/debugger/setBreakpoint-dfg-expected.txt +++ /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/debugger/setBreakpoint-dfg-actual.txt @@ -13,6 +13,6 @@ dfgWithoutInline result: 502500 Hit Breakpoint 2! Removed Breakpoint 2! +dfgWithInline result: 504500 PASS -dfgWithInline result: 504500
Brent Fulgham
Comment 5 2014-02-18 14:31:46 PST
Heap Profiler Failures: 1. inspector-protocol/heap-profiler/heap-snapshot-with-detached-dom-tree.html: --- /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/heap-snapshot-with-detached-dom-tree-expected.txt +++ /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/heap-snapshot-with-detached-dom-tree-actual.txt @@ -1,7 +1,13 @@ +ALERT: Error in: ../../../../inspector-protocol/heap-profiler/resources/heap-snapshot-common.js +ALERT: importScript@file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:156:24 +eval code +eval@[native code] +importScript@file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:158:16 +test +eval code +eval@[native code] +file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:265:13 +CONSOLE MESSAGE: line 269: Error: Invalid script URL: ../../../../../Source/WebCore/inspector/front-end/HeapSnapshot.js Test that all nodes from the detached DOM tree will get into one group in the heap snapshot. Bug 107819. - SUCCESS: found (Detached DOM trees) -SUCCESS: found Detached DOM tree / 4 entries -SUCCESS: found 3 DIVs in Detached DOM tree / 4 entries -SUCCESS: didGetHeapSnapshot 2. inspector-protocol/heap-profiler/heap-snapshot-with-event-listener.html: --- /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/heap-snapshot-with-event-listener-expected.txt +++ /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/heap-snapshot-with-event-listener-actual.txt @@ -1,6 +1,13 @@ +ALERT: Error in: ../../../../inspector-protocol/heap-profiler/resources/heap-snapshot-common.js +ALERT: importScript@file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:156:24 +eval code +eval@[native code] +importScript@file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:158:16 +test +eval code +eval@[native code] +file:///C:/Projects/WebKit/OpenSource/LayoutTests/http/tests/inspector-protocol/resources/InspectorTest.js:265:13 +CONSOLE MESSAGE: line 269: Error: Invalid script URL: ../../../../../Source/WebCore/inspector/front-end/HeapSnapshot.js Test that all nodes from the detached DOM tree will get into one group in the heap snapshot. Bug 107819. - SUCCESS: found myEventListener -SUCCESS: found link from HTMLBodyElement to HTMLBodyElement -SUCCESS: didGetHeapSnapshot 3. inspector-protocol/heap-profiler/take-heap-snapshot.html: This makes it looks like it's not running the test with "DumpAsText" style output. --- /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/take-heap-snapshot-expected.txt +++ /cygdrive/c/Projects/WebKit/OpenSource/WebKitBuild/Release/bin32/layout-test-results/inspector-protocol/heap-profiler/take-heap-snapshot-actual.txt @@ -1,5 +1,25 @@ -Test that heap profiler doesn't crash while taking snapshot on a page where iframe was navigated to a new location after storing a hold of a function from the previous page. Bug 103076. - - -SUCCESS: didTakeHeapSnapshot - +layer at (0,0) size 800x600 + RenderView at (0,0) size 800x600 +layer at (0,0) size 800x600 + RenderBlock {HTML} at (0,0) size 800x600 + RenderBody {BODY} at (8,8) size 784x584 + RenderBlock {P} at (0,0) size 784x36 + RenderText {#text} at (0,0) size 754x36 + text run at (0,0) width 754: "Test that heap profiler doesn't crash while taking snapshot on a page where iframe was navigated to a new location after" + text run at (0,18) width 327: "storing a hold of a function from the previous page. " + RenderInline {A} at (0,0) size 83x18 [color=#0000EE] + RenderText {#text} at (327,18) size 83x18 + text run at (327,18) width 83: "Bug 103076." + RenderBlock (anonymous) at (0,52) size 784x154 + RenderInline {A} at (0,0) size 304x18 [color=#0000EE] + RenderText {#text} at (0,0) size 0x0 + RenderIFrame {IFRAME} at (0,0) size 304x154 [border: (2px inset #0000EE)] + layer at (0,0) size 300x150 + RenderView at (0,0) size 300x150 + layer at (0,0) size 300x150 + RenderBlock {HTML} at (0,0) size 300x150 + RenderBody {BODY} at (8,8) size 284x134 + RenderBlock {DIV} at (0,0) size 284x18 + RenderText {#text} at (0,0) size 85x18 + text run at (0,0) width 85: "resource.html" + RenderText {#text} at (0,0) size 0x0
Timothy Hatcher
Comment 6 2014-02-18 15:39:05 PST
(In reply to comment #5) > Heap Profiler Failures: The heap-profiler tests could be skipped. We don't support the heap profiler right now.
Note You need to log in before you can comment on or make changes to this bug.