Title: [244675] trunk
- Revision
- 244675
- Author
- [email protected]
- Date
- 2019-04-25 19:55:36 -0700 (Thu, 25 Apr 2019)
Log Message
REGRESSION (r234330): 3 legacy-animation-engine/compositing tests are flaky failures
https://bugs.webkit.org/show_bug.cgi?id=188357
<rdar://problem/42986633>
Reviewed by Dean Jackson.
Source/WebCore:
DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
tests would fall into low memory mode and have different behavior.
Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
to Internals, so it's shared by DRT and WTR.
We no longer need the WK2 C SPI glue.
* testing/Internals.cpp:
(WebCore::Internals::resetToConsistentState):
Source/WebKit:
DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
tests would fall into low memory mode and have different behavior.
Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
to Internals, so it's shared by DRT and WTR.
We no longer need the WK2 C SPI glue.
* WebProcess/InjectedBundle/API/c/WKBundlePage.cpp:
(WKBundlePageSetEventThrottlingBehaviorOverride):
(WKBundlePageSetCompositingPolicyOverride): Deleted.
* WebProcess/InjectedBundle/API/c/WKBundlePagePrivate.h:
Tools:
DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
tests would fall into low memory mode and have different behavior.
Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
to Internals, so it's shared by DRT and WTR.
We no longer need the WK2 C SPI glue.
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::prepare):
Modified Paths
Diff
Modified: trunk/Source/WebCore/ChangeLog (244674 => 244675)
--- trunk/Source/WebCore/ChangeLog 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Source/WebCore/ChangeLog 2019-04-26 02:55:36 UTC (rev 244675)
@@ -1,3 +1,22 @@
+2019-04-25 Simon Fraser <[email protected]>
+
+ REGRESSION (r234330): 3 legacy-animation-engine/compositing tests are flaky failures
+ https://bugs.webkit.org/show_bug.cgi?id=188357
+ <rdar://problem/42986633>
+
+ Reviewed by Dean Jackson.
+
+ DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
+ tests would fall into low memory mode and have different behavior.
+
+ Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
+ to Internals, so it's shared by DRT and WTR.
+
+ We no longer need the WK2 C SPI glue.
+
+ * testing/Internals.cpp:
+ (WebCore::Internals::resetToConsistentState):
+
2019-04-25 Sihui Liu <[email protected]>
[ iOS Sim ] REGRESSION (r242986) Layout Test storage/indexeddb/modern/idbtransaction-objectstore-failures-private.html is a flaky failure
Modified: trunk/Source/WebCore/testing/Internals.cpp (244674 => 244675)
--- trunk/Source/WebCore/testing/Internals.cpp 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Source/WebCore/testing/Internals.cpp 2019-04-26 02:55:36 UTC (rev 244675)
@@ -452,7 +452,7 @@
page.mainFrame().setTextZoomFactor(1.0f);
- page.setCompositingPolicyOverride(WTF::nullopt);
+ page.setCompositingPolicyOverride(WebCore::CompositingPolicy::Normal);
FrameView* mainFrameView = page.mainFrame().view();
if (mainFrameView) {
Modified: trunk/Source/WebKit/ChangeLog (244674 => 244675)
--- trunk/Source/WebKit/ChangeLog 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Source/WebKit/ChangeLog 2019-04-26 02:55:36 UTC (rev 244675)
@@ -1,3 +1,24 @@
+2019-04-25 Simon Fraser <[email protected]>
+
+ REGRESSION (r234330): 3 legacy-animation-engine/compositing tests are flaky failures
+ https://bugs.webkit.org/show_bug.cgi?id=188357
+ <rdar://problem/42986633>
+
+ Reviewed by Dean Jackson.
+
+ DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
+ tests would fall into low memory mode and have different behavior.
+
+ Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
+ to Internals, so it's shared by DRT and WTR.
+
+ We no longer need the WK2 C SPI glue.
+
+ * WebProcess/InjectedBundle/API/c/WKBundlePage.cpp:
+ (WKBundlePageSetEventThrottlingBehaviorOverride):
+ (WKBundlePageSetCompositingPolicyOverride): Deleted.
+ * WebProcess/InjectedBundle/API/c/WKBundlePagePrivate.h:
+
2019-04-25 Chris Dumez <[email protected]>
Trying to load a main resource outside the sandbox should not crash the WebContent process
Modified: trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePage.cpp (244674 => 244675)
--- trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePage.cpp 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePage.cpp 2019-04-26 02:55:36 UTC (rev 244675)
@@ -803,21 +803,3 @@
WebKit::toImpl(page)->corePage()->setEventThrottlingBehaviorOverride(behaviorValue);
}
-
-void WKBundlePageSetCompositingPolicyOverride(WKBundlePageRef page, WKCompositingPolicy* policy)
-{
- Optional<WebCore::CompositingPolicy> policyValue;
- if (policy) {
- switch (*policy) {
- case kWKCompositingPolicyNormal:
- policyValue = WebCore::CompositingPolicy::Normal;
- break;
- case kWKCompositingPolicyConservative:
- policyValue = WebCore::CompositingPolicy::Conservative;
- break;
- }
- }
-
- WebKit::toImpl(page)->corePage()->setCompositingPolicyOverride(policyValue);
-}
-
Modified: trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePagePrivate.h (244674 => 244675)
--- trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePagePrivate.h 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Source/WebKit/WebProcess/InjectedBundle/API/c/WKBundlePagePrivate.h 2019-04-26 02:55:36 UTC (rev 244675)
@@ -134,16 +134,6 @@
// Passing null in the second parameter clears the override.
WK_EXPORT void WKBundlePageSetEventThrottlingBehaviorOverride(WKBundlePageRef, WKEventThrottlingBehavior*);
-enum {
- kWKCompositingPolicyNormal = 0,
- kWKCompositingPolicyConservative
-};
-
-typedef uint32_t WKCompositingPolicy;
-
-// Passing null in the second parameter clears the override.
-WK_EXPORT void WKBundlePageSetCompositingPolicyOverride(WKBundlePageRef, WKCompositingPolicy*);
-
#if TARGET_OS_IPHONE
WK_EXPORT void WKBundlePageSetUseTestingViewportConfiguration(WKBundlePageRef, bool);
#endif
Modified: trunk/Tools/ChangeLog (244674 => 244675)
--- trunk/Tools/ChangeLog 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Tools/ChangeLog 2019-04-26 02:55:36 UTC (rev 244675)
@@ -1,3 +1,22 @@
+2019-04-25 Simon Fraser <[email protected]>
+
+ REGRESSION (r234330): 3 legacy-animation-engine/compositing tests are flaky failures
+ https://bugs.webkit.org/show_bug.cgi?id=188357
+ <rdar://problem/42986633>
+
+ Reviewed by Dean Jackson.
+
+ DumpRenderTree had no code that set page.setCompositingPolicyOverride() to Normal, so some
+ tests would fall into low memory mode and have different behavior.
+
+ Fix by moving the code that calls setCompositingPolicyOverride(Normal) from the WK2 layer
+ to Internals, so it's shared by DRT and WTR.
+
+ We no longer need the WK2 C SPI glue.
+
+ * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
+ (WTR::InjectedBundlePage::prepare):
+
2019-04-25 Jonathan Bedard <[email protected]>
webkitpy: Including skipped tests in the final count
Modified: trunk/Tools/WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp (244674 => 244675)
--- trunk/Tools/WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp 2019-04-26 01:58:33 UTC (rev 244674)
+++ trunk/Tools/WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp 2019-04-26 02:55:36 UTC (rev 244675)
@@ -420,10 +420,6 @@
// Force consistent "responsive" behavior for WebPage::eventThrottlingDelay() for testing. Tests can override via internals.
WKEventThrottlingBehavior behavior = kWKEventThrottlingBehaviorResponsive;
WKBundlePageSetEventThrottlingBehaviorOverride(m_page, &behavior);
-
- // Force consistent compositing behavior, even if the test runner is under memory pressure. Tests can override via internals.
- WKCompositingPolicy policy = kWKCompositingPolicyNormal;
- WKBundlePageSetCompositingPolicyOverride(m_page, &policy);
}
void InjectedBundlePage::resetAfterTest()
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes
