Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(1583)

Unified Diff: content/browser/site_per_process_browsertest.cc

Issue 1345353003: Fix detection of RenderViewHosts pending deletion in CreateRenderViewHost. (Closed) Base URL: https://chromium.googlesource.com/chromium/src.git@master
Patch Set: Nits Created 5 years, 3 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
Index: content/browser/site_per_process_browsertest.cc
diff --git a/content/browser/site_per_process_browsertest.cc b/content/browser/site_per_process_browsertest.cc
index d500ce60e40f4ba222745b220de50392867d9365..c1d588d66e5d973a48508d8a08bea4391dd5c37d 100644
--- a/content/browser/site_per_process_browsertest.cc
+++ b/content/browser/site_per_process_browsertest.cc
@@ -275,6 +275,23 @@ bool ConsoleObserverDelegate::AddMessageToConsole(
return false;
}
+// A BrowserMessageFilter that drops SwapOut ACK messages.
+class SwapoutACKMessageFilter : public BrowserMessageFilter {
+ public:
+ SwapoutACKMessageFilter() : BrowserMessageFilter(FrameMsgStart) {}
+
+ protected:
+ ~SwapoutACKMessageFilter() override {}
+
+ private:
+ // BrowserMessageFilter:
+ bool OnMessageReceived(const IPC::Message& message) override {
+ return message.type() == FrameHostMsg_SwapOut_ACK::ID;
+ }
+
+ DISALLOW_COPY_AND_ASSIGN(SwapoutACKMessageFilter);
+};
+
} // namespace
//
@@ -3371,4 +3388,43 @@ IN_PROC_BROWSER_TEST_F(SitePerProcessBrowserTest,
EXPECT_EQ(expected_title, title_watcher.WaitAndGetTitle());
}
+// Test for https://crbug.com/515302. Perform two navigations, A->B->A, and
+// delay the SwapOut ACK from the A->B navigation, so that the second B->A
+// navigation is initiated before the first page receives the SwapOut ACK.
+// Ensure that the RVH(A) that's pending deletion is not reused in that case.
+IN_PROC_BROWSER_TEST_F(SitePerProcessBrowserTest,
+ RenderViewHostPendingDeletionIsNotReused) {
+ GURL a_url(embedded_test_server()->GetURL("a.com", "/title1.html"));
+ NavigateToURL(shell(), a_url);
+
+ FrameTreeNode* root = static_cast<WebContentsImpl*>(shell()->web_contents())
+ ->GetFrameTree()
+ ->root();
+ RenderFrameHostImpl* rfhi = root->current_frame_host();
nasko 2015/09/18 19:06:17 nit: Either rfh or rvhi, for consistency. IMHO, th
alexmos 2015/09/18 22:37:05 Done.
+ RenderViewHostImpl* rvh = rfhi->render_view_host();
+
+ // Install a BrowserMessageFilter to drop SwapOut ACK messages in A's
+ // process.
+ scoped_refptr<SwapoutACKMessageFilter> filter = new SwapoutACKMessageFilter();
+ rfhi->GetProcess()->AddFilter(filter.get());
+
+ // Navigate to B. The SwapOut completion should be delayed.
+ GURL b_url(embedded_test_server()->GetURL("b.com", "/title2.html"));
+ NavigateToURL(shell(), b_url);
+
+ // Navigate back to A. Since there was no SwapOut ACK for A->B, this should
+ // happen while first page's RenderFrameHost and RenderViewHost are pending
+ // deletion.
+ EXPECT_TRUE(root->render_manager()->IsPendingDeletion(rfhi));
nasko 2015/09/18 19:06:17 nit: I would put those expectations right after th
alexmos 2015/09/18 22:37:05 Done.
+ EXPECT_TRUE(rvh->is_pending_deletion());
+ NavigateToURL(shell(), a_url);
nasko 2015/09/18 19:06:17 nit: I wonder if we should make this navigation as
alexmos 2015/09/18 22:37:05 Done.
+
+ // Check that the RenderViewHost wasn't reused.
+ EXPECT_NE(rvh, shell()->web_contents()->GetRenderViewHost());
+
+ // Simulate that the dropped SwapOut ACK message arrives now on the original
+ // RenderFrameHost.
+ rfhi->OnSwappedOut();
nasko 2015/09/18 19:06:17 For going the extra mile, should we put a RenderFr
alexmos 2015/09/18 22:37:05 Done.
+}
+
} // namespace content

Powered by Google App Engine
This is Rietveld 408576698