Fix handling of unexpected cases in bit-flip test.

There can be potentially unexpected message names in the bit-flip test.

This removes some NOTREACHED()s in the base class which may be
reached when bad data comes along.  Merely tally such so that it
can be checked in the cases where we know it's "impossible".

We also need to be more lenient about the results of reading
from pipes.

BUG=376173
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272385

[Tom Sepez, 2014-05-22 18:15:12]

Trung, please review.  Thanks.

[viettrungluu, 2014-05-22 19:24:17]

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
File mojo/apps/js/test/js_to_cpp.mojom (right):

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
mojo/apps/js/test/js_to_cpp.mojom:31: // Note: For messages which control test
flow, pick numbers that are unlikely
Would it be better to use a completely separate interface for control purposes?
Doing so would ensure that no matter how hard you're fuzzing, you won't end up
accidentally starting/stopping the test.

(Also, at some point, we may probably enforce that the ordinals be consecutive
or something like that.)

[Tom Sepez, 2014-05-22 19:42:03]

On 2014/05/22 19:24:17, viettrungluu wrote:
>
https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
> File mojo/apps/js/test/js_to_cpp.mojom (right):
> 
>
https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
> mojo/apps/js/test/js_to_cpp.mojom:31: // Note: For messages which control test
> flow, pick numbers that are unlikely
> Would it be better to use a completely separate interface for control
purposes?
> Doing so would ensure that no matter how hard you're fuzzing, you won't end up
> accidentally starting/stopping the test.
>
Probably.  Separate interface implies separate pipe, right?
I'd like to do this in a follow-up so as to unblock yzshen's testing.

[Tom Sepez, 2014-05-22 21:00:45]

Ah. In other words, if they're on separate OS pipes, then we'd hit the sync
issue (e.g. no reason to believe all the responses have been read when the
finished message is seen). If they're on the same OS pipe, then hard fuzzing
corrupts the value.  So I'm not sure there's a way around this. Am I being
dense?

[viettrungluu, 2014-05-22 21:18:55]

On 2014/05/22 21:00:45, Tom Sepez wrote:
> Ah. In other words, if they're on separate OS pipes, then we'd hit the sync
> issue (e.g. no reason to believe all the responses have been read when the
> finished message is seen). If they're on the same OS pipe, then hard fuzzing
> corrupts the value.  So I'm not sure there's a way around this. Am I being
> dense?

I guess we'll fix that with "child/associated" message pipes. :)

So LGTM w/nits for now.

[viettrungluu, 2014-05-22 21:19:11]

FFFFFFFFFFFFFUUUUUUUUUUUUUUUUUUUUUUUUU.

Replying doesn't send comments.

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
File mojo/apps/js/test/js_to_cpp.mojom (right):

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
mojo/apps/js/test/js_to_cpp.mojom:35: StartTest @88888888 ();
Style nit: Probably format this as:

  StartTest@88888888();

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
mojo/apps/js/test/js_to_cpp.mojom:38: TestFinished @99999999 ();
"

[Tom Sepez, 2014-05-22 21:22:21]

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
File mojo/apps/js/test/js_to_cpp.mojom (right):

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
mojo/apps/js/test/js_to_cpp.mojom:35: StartTest @88888888 ();
On 2014/05/22 21:19:12, viettrungluu wrote:
> Style nit: Probably format this as:
> 
>   StartTest@88888888();

Done.

https://codereview.chromium.org/297833007/diff/40001/mojo/apps/js/test/js_to_...
mojo/apps/js/test/js_to_cpp.mojom:38: TestFinished @99999999 ();
On 2014/05/22 21:19:12, viettrungluu wrote:
> "

Done.

[Tom Sepez, 2014-05-22 21:23:00]

The CQ bit was checked by tsepez@chromium.org

[commit-bot: I haz the power, 2014-05-22 21:23:50]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/tsepez@chromium.org/297833007/50001

[commit-bot: I haz the power, 2014-05-23 00:49:24]

Change committed as 272385