This should make ProcessUtilTest.KillSlowChild not-flaky.

This should make ProcessUtilTest.KillSlowChild not-flaky.

There is no need to check process count (which tends to induce flakiness).
In fact, we are already checking all important facts:

- after checking that process handle is a not-null handle we know
  that the process started successfully

- after WaitForSingleProcess returns true we know that the process
  finished successfully

BUG=8811
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=11728

[Paweł Hajdan Jr., 2009-03-16 12:24:19]

It seems that this flakiness really annoys people, see
http://codereview.chromium.org/18832 and http://codereview.chromium.org/28277 .

[Dean McNamee, 2009-03-16 12:35:54]

Looks ok to me.  I chased the history back a bit, but I'm not sure who wrote the
original test.

You should add bugs to a BUG= line in your description.

[Dean McNamee, 2009-03-16 12:42:12]

Yeah, can you make a bug quick saying that this test was flakey, and that a good
solution it to just drop these flakey checks which aren't adding anything.  Then
there is a better history / place to comment than a review.

With a BUG= line, I think LGTM, and you can commit.

On 2009/03/16 12:35:54, Dean McNamee wrote:
> Looks ok to me.  I chased the history back a bit, but I'm not sure who wrote
the
> original test.
> 
> You should add bugs to a BUG= line in your description.

[dank, 2009-03-16 12:54:53]

LGTM