[Webkit-unassigned] [Bug 193320] [GTK] Several resourceLoadStatistics tests are failing since r239816 "Create a WebResourceLoadStatisticsStore attached to the NetworkSession"

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Jan 10 08:49:23 PST 2019


https://bugs.webkit.org/show_bug.cgi?id=193320

--- Comment #1 from Brent Fulgham <bfulgham at webkit.org> ---
My guess is that I broke it with this change:

Index: trunk/Source/WebKit/UIProcess/WebPageProxy.cpp
===================================================================
--- a/trunk/Source/WebKit/UIProcess/WebPageProxy.cpp
+++ b/trunk/Source/WebKit/UIProcess/WebPageProxy.cpp
@@ -4452,6 +4448,9 @@
     WebFrameProxy* originatingFrame = m_process->webFrame(originatingFrameInfoData.frameID);

+#if ENABLE(RESOURCE_LOAD_STATISTICS)
     if (auto* resourceLoadStatisticsStore = websiteDataStore().resourceLoadStatistics())
         resourceLoadStatisticsStore->logFrameNavigation(frame, URL(URL(), m_pageLoadState.url()), request, redirectResponse.url());
+    logFrameNavigation(frame, URL(URL(), m_pageLoadState.url()), request, redirectResponse.url());
+#endif

     if (m_policyClient)
@@ -4475,4 +4474,36 @@
     m_shouldSuppressAppLinksInNextNavigationPolicyDecision = false;
 }

I was a little confused with this code on GTK/WPE, since the feature defines for non-Apple builds seems to have it off:

WebKitFeatures.cmake:    WEBKIT_OPTION_DEFINE(ENABLE_RESOURCE_LOAD_STATISTICS "Toggle resource load statistics support" PRIVATE OFF)

So, I moved the websiteDataStore().resourceLoadStatistics() access inside a macro, since that stuff shouldn't exist if the build flag is turned off. GTK (and probably others) seem to be building a weird half on/half off version of the feature.

I think GTK either needs to turn the feature flag on, or disable the tests.

As a quick fix, you could just remove the "#if ENABLE(RESOURCE_LOAD_STATISTICS)" I added in WebPageProxy.cpp.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190110/b102137a/attachment.html>


More information about the webkit-unassigned mailing list