[test] Add random seed stress mode to test runner.

[test] Add random seed stress mode to test runner.

NOTRY=true

Committed: https://crrev.com/f3e8c11afdee02da98150fb9634b3b1e49e83f9e
Cr-Commit-Position: refs/heads/master@{#28936}

[Michael Achenbach, 2015-06-10 15:40:30]

machenbach@chromium.org changed reviewers:
+ jkummerow@chromium.org

[Michael Achenbach, 2015-06-10 15:40:30]

PTAL

[Jakob Kummerow, 2015-06-11 08:50:35]

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py
File tools/run-tests.py (right):

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode386
tools/run-tests.py:386: if options.random_seed_stress_count <= 1:
Shouldn't this condition include a check that options.random_seed == 0?
Otherwise it's a bit surprising that any explicitly specified random seed will
silently be overwritten unless the random seed stress count is > 1.

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode565
tools/run-tests.py:565: options.random_seed,
In case of --random-seed=1 --random-seed-stress-count=2, this will lead to
duplicate --random-seed flags on the final test command line, right? One coming
from the test case (cf. CopyAddingFlags below), and one coming from the context.

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode620
tools/run-tests.py:620: yield ["--random-seed=%s" % str(seed)]
nit: explicit str() is not necessary when using %s (arguably %d would be more
appropriate, and  still wouldn't need manual conversion)

[Michael Achenbach, 2015-06-11 09:45:05]

Done. PTAL

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py
File tools/run-tests.py (right):

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode386
tools/run-tests.py:386: if options.random_seed_stress_count <= 1:
On 2015/06/11 08:50:35, Jakob wrote:
> Shouldn't this condition include a check that options.random_seed == 0?
> Otherwise it's a bit surprising that any explicitly specified random seed will
> silently be overwritten unless the random seed stress count is > 1.

Good catch!

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode565
tools/run-tests.py:565: options.random_seed,
On 2015/06/11 08:50:35, Jakob wrote:
> In case of --random-seed=1 --random-seed-stress-count=2, this will lead to
> duplicate --random-seed flags on the final test command line, right? One
coming
> from the test case (cf. CopyAddingFlags below), and one coming from the
context.

Done.

https://codereview.chromium.org/1180473003/diff/1/tools/run-tests.py#newcode620
tools/run-tests.py:620: yield ["--random-seed=%s" % str(seed)]
On 2015/06/11 08:50:35, Jakob wrote:
> nit: explicit str() is not necessary when using %s (arguably %d would be more
> appropriate, and  still wouldn't need manual conversion)

Done.

[Jakob Kummerow, 2015-06-11 10:53:43]

lgtm

[Michael Achenbach, 2015-06-11 11:40:39]

The CQ bit was checked by machenbach@chromium.org

[commit-bot: I haz the power, 2015-06-11 11:40:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1180473003/20001

[commit-bot: I haz the power, 2015-06-11 11:42:37]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-06-11 11:42:46]

Patchset 2 (id:??) landed as
https://crrev.com/f3e8c11afdee02da98150fb9634b3b1e49e83f9e
Cr-Commit-Position: refs/heads/master@{#28936}