[fuchsia] Disable flaky AccessibilityBridgeTests.

Several of the AccessibilityBridgeTest suite's tests expect exact
numbers of commit calls to the SemanticTree, which will not be met in
general due to the asynchronous nature of e.g. page loads.

Also rename the TestNavigation to give it a less generic sounding name.

Bug: 1168126, 1167266
Change-Id: I51d557face6c31ee3b6649cc5a28f784311d64b3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2637500
Commit-Queue: Wez <wez@chromium.org>
Commit-Queue: Marina Ciocea <marinaciocea@chromium.org>
Reviewed-by: Marina Ciocea <marinaciocea@chromium.org>
Auto-Submit: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#844791}

fuchsia/engine/browser/accessibility_bridge_browsertest.cc

@@ -182,7 +182,9 @@ IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, CorrectDataSent) {
 // Batching is performed when the number of nodes to send or delete exceeds the
 // maximum, as set on the Fuchsia side. Check that all nodes are received by the
 // Semantic Tree when batching is performed.
-IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, DataSentWithBatching) {
+// TODO(1168126): Number of commit calls is not deterministic in general,
+// leading to flakiness.
+IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, DISABLED_DataSentWithBatching) {
   LoadPage(kPage2Path, kPage2Title);
 
   // Run until we expect more than a batch's worth of nodes to be present.
@@ -197,7 +199,10 @@ IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, DataSentWithBatching) {
 
 // Check that semantics information is correctly sent when navigating from page
 // to page.
-IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, TestNavigation) {
+// TODO(1168126): Number of commit calls is not deterministic in general,
+// leading to flakiness.
+IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest,
+                       DISABLED_NavigateFromPageToPage) {
   LoadPage(kPage1Path, kPage1Title);
 
   semantics_manager_.semantic_tree()->RunUntilNodeCountAtLeast(kPage1NodeCount);
@@ -406,7 +411,10 @@ IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, Slider) {
 
 // This test makes sure that when semantic updates toggle on / off / on, the
 // full semantic tree is sent in the first update when back on.
-IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, TogglesSemanticsUpdates) {
+// TODO(1168126): Number of commit calls is not deterministic in general,
+// leading to flakiness.
+IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest,
+                       DISABLED_TogglesSemanticsUpdates) {
   LoadPage(kPage1Path, kPage1Title);
   semantics_manager_.semantic_tree()->RunUntilCommitCountIs(1);
 
@@ -435,7 +443,10 @@ IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, TogglesSemanticsUpdates) {
 // also forward the nodes in a way that leaves the tree in the Fuchsia side in a
 // valid state. Note that every time that a new tree is sent to Fuchsia, the
 // FakeSemantiTree checks if the tree is valid.
-IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, TreeModificationsAreForwarded) {
+// TODO(1168126): Number of commit calls is not deterministic in general,
+// leading to flakiness.
+IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest,
+                       DISABLED_TreeModificationsAreForwarded) {
   // Loads a page, so a real frame is created for this test. Then, several tree
   // operations are applied on top of it, using the AXTreeID that corresponds to
   // that frame.
@@ -616,7 +627,9 @@ IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, TreeModificationsAreForwarded) {
   }
 }
 
-// TODO(crbug.com/1167266): Flaky test.
+// TODO(crbug.com/1167266): Number of commit calls is not deterministic in
+// general, leading to flakiness. Hard-wired wait for 200ms may also lead to
+// expectations failures if the system running tests is overloaded.
 IN_PROC_BROWSER_TEST_F(AccessibilityBridgeTest, DISABLED_OutOfProcessIframe) {
   constexpr int64_t kBindingsId = 1234;