Re-enable SitePerProcessBrowserTest.RecreateMainFrameAfterCancelPending

The test was flaky but the underlying flakiness was fixed in
86ae1baf.

The MainThreadEventQueue now destroys its OneShotTimer on the main
thread in ClearClient() before the main thread reference to the
MainThreadEventQueue is dropped. This prevents the timer from being
destroyed along with the MainThreadEventQueue on the compositor
thread, if that is the last reference to the MainThreadEventQueue.

R=dtapuska@chromium.org

Bug: 1104826
Change-Id: I0e46906f731ea5c628cb997eb9b9e0b57f407daf
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2429803Reviewed-by: Dave Tapuska <dtapuska@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Auto-Submit: danakj <danakj@chromium.org>
Commit-Queue: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#810456}

content/browser/site_per_process_browsertest.cc

@@ -8773,9 +8773,8 @@ IN_PROC_BROWSER_TEST_P(SitePerProcessBrowserTest,
 // same site as the canceled RFH doesn't lead to a renderer crash.  The steps
 // here are similar to ReuseNonLiveRenderViewHostAfterCancelPending, but don't
 // involve crashing the renderer. See https://crbug.com/651980.
-// Flaky on all platforms, see https://crbug.com/1104826.
 IN_PROC_BROWSER_TEST_P(SitePerProcessBrowserTest,
-                       DISABLED_RecreateMainFrameAfterCancelPending) {
+                       RecreateMainFrameAfterCancelPending) {
   GURL a_url(embedded_test_server()->GetURL("a.com", "/title1.html"));
   GURL b_url(embedded_test_server()->GetURL("b.com", "/title2.html"));
   GURL c_url(embedded_test_server()->GetURL("c.com", "/title3.html"));