[ios clean] Remove stop on Browser coordinator dealloc

ios/clean/chrome/browser/browser_coordinator_unittest.mm

Index: ios/clean/chrome/browser/browser_coordinator_unittest.mm
diff --git a/ios/clean/chrome/browser/browser_coordinator_unittest.mm b/ios/clean/chrome/browser/browser_coordinator_unittest.mm
index c4fabd2f3d336d47aefc65683adc7c8824d1ab65..8f545fcf35143cd5e95ebd750f28e7d7a6587394 100644
--- a/ios/clean/chrome/browser/browser_coordinator_unittest.mm
+++ b/ios/clean/chrome/browser/browser_coordinator_unittest.mm
@@ -50,7 +50,9 @@
 
 namespace {
 
-TEST(BrowserCoordinatorTest, TestStopOnDealloc) {
+// HACK: This test will stay disabled, removed or changed depending on how
+// BrowserCoordinator behaves on dealloc.

[edchin, 2017/02/02 22:43:20]

I think it is best to remove this code entirely. You completely removed the
actual dealloc code without any comments. So this code is confusing to someone
who has no context.

[sczs, 2017/02/02 23:25:51]

I was thinking the exact same thing. But decide to disable it since we're not
sure of 
the final behavior yet, Mark mentioned that we might still call stop on dealloc.
If not this can become a Test to check that stop is NOT called on dealloc. 

The HACK: tag will hopefully give some context to people. In any case, I
wouldn't 
mind removing it entirely and re-introducing it later if you feel strongly about
it. 

[marq (ping after 24h), 2017/02/03 08:36:15]

I want to get out of the mindset of disabling tests, especially in a case like
this where the behavioral change is intentional. 

For this change, it's probably correct just to change line 65 to EXPECT_FALSE. A
comment that stop-on-dealloc used to be the behavior but isn't at the moment is
fine, but I don't think it needs to be marked as a hack or placeholder.

[sczs, 2017/02/03 18:51:23]

Done.

+TEST(DISABLED_BrowserCoordinatorTest, TestStopOnDealloc) {
   __block BOOL called = NO;
 
   {