Add flag --require-audio-hardware-for-testing.

Add flag --require-audio-hardware-for-testing.

When enabled, tests that require audio hardware will fail if the system
doesn't provide it. Otherwise, they will silently succeed.

TEST=media_unittests (and AudioInputTests with !InputDevicesAvailable())
BUG=432624
Committed: https://crrev.com/b0ddebe2ea0b00c272af7820b9c108c4438f5b88
Cr-Commit-Position: refs/heads/master@{#315823}

[watk, 2015-02-09 21:07:22]

watk@chromium.org changed reviewers:
+ dalecurtis@chromium.org

[watk, 2015-02-09 21:07:22]

I made InputDevicesAvailable() return false.

Output without the flag:

[ RUN      ] AudioInputTest.OpenAndClose
[28838:28838:0209/122657:1726369119143:WARNING:audio_input_unittest.cc(191)]
Skipping test: Requirement(s) not satisfied (InputDevicesAvailable())
[       OK ] AudioInputTest.OpenAndClose (4 ms)

With the flag:

[ RUN      ] AudioInputTest.OpenStopAndClose
../../media/audio/audio_input_unittest.cc:204: Failure
Failed
Requirement(s) not satisfied (InputDevicesAvailable())
[  FAILED  ] AudioInputTest.OpenStopAndClose (2 ms)

[DaleCurtis, 2015-02-09 21:21:32]

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_input_volu...
File media/audio/audio_input_volume_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_input_volu...
media/audio/audio_input_volume_unittest.cc:49: bool CanRunAudioTests() {
HasCoreAudioAndInputDevices ?

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
File media/audio/audio_unittest_utils.h (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
media/audio/audio_unittest_utils.h:10: #define
CAN_RUN_AUDIO_TEST_IF(requirements_satisfied)                 \
Hmm, naming feels odd. What about something like:

ABORT_AUDIO_TEST_IF_NOT(...)

Both paths are abort, just some are hard vs soft failures, right?

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
media/audio/audio_unittest_utils.h:12: bool satisfied = requirements_satisfied; 
                        \
Most of this should be moved into a static function and only the FAIL() or
return part kept in the macro.

https://codereview.chromium.org/914483002/diff/1/media/audio/win/audio_output...
File media/audio/win/audio_output_win_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/win/audio_output...
media/audio/win/audio_output_win_unittest.cc:601:
REQUIRE_FOR_AUDIO_TEST(audio_man->HasAudioOutputDevices());
Different name?

[watk, 2015-02-09 22:03:58]

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_input_volu...
File media/audio/audio_input_volume_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_input_volu...
media/audio/audio_input_volume_unittest.cc:49: bool CanRunAudioTests() {
On 2015/02/09 21:21:31, DaleCurtis wrote:
> HasCoreAudioAndInputDevices ?

Done.

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
File media/audio/audio_unittest_utils.h (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
media/audio/audio_unittest_utils.h:10: #define
CAN_RUN_AUDIO_TEST_IF(requirements_satisfied)                 \
On 2015/02/09 21:21:31, DaleCurtis wrote:
> Hmm, naming feels odd. What about something like:
> 
> ABORT_AUDIO_TEST_IF_NOT(...)
> 
> Both paths are abort, just some are hard vs soft failures, right?

Much better, thanks

https://codereview.chromium.org/914483002/diff/1/media/audio/audio_unittest_u...
media/audio/audio_unittest_utils.h:12: bool satisfied = requirements_satisfied; 
                        \
On 2015/02/09 21:21:31, DaleCurtis wrote:
> Most of this should be moved into a static function and only the FAIL() or
> return part kept in the macro.

Done.

https://codereview.chromium.org/914483002/diff/1/media/audio/win/audio_output...
File media/audio/win/audio_output_win_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/1/media/audio/win/audio_output...
media/audio/win/audio_output_win_unittest.cc:601:
REQUIRE_FOR_AUDIO_TEST(audio_man->HasAudioOutputDevices());
On 2015/02/09 21:21:32, DaleCurtis wrote:
> Different name?

Done.

[DaleCurtis, 2015-02-09 22:10:50]

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_low_la...
File media/audio/audio_low_latency_input_output_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_low_la...
media/audio/audio_low_latency_input_output_unittest.cc:378:
ABORT_AUDIO_TEST_IF_NOT(audio_manager()->HasAudioInputDevices() &&
Fix alignment.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
File media/audio/audio_unittest_utils.cc (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.cc:1: #include "base/command_line.h"
Needs license header.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.cc:14: *should_fail = fail_if_unsatisfied;
Should always be set?

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
File media/audio/audio_unittest_utils.h (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:1: #include
"testing/gtest/include/gtest/gtest.h"
Needs license header. #ifdef protections for multiple include.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:4: // required audio device or library. If
the |kRequireAudioHardwareForTesting|
I'd use the actual --flag-name.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:19: bool should_abort_audio_test(bool
requirements_satisfied,
Should be CamelCase method name.

[watk, 2015-02-09 22:43:07]

Patchset #3 (id:40001) has been deleted

[watk, 2015-02-10 01:38:33]

Had a bunch of back and forth with silly compiler errors on windows. Should be
good now.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_low_la...
File media/audio/audio_low_latency_input_output_unittest.cc (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_low_la...
media/audio/audio_low_latency_input_output_unittest.cc:378:
ABORT_AUDIO_TEST_IF_NOT(audio_manager()->HasAudioInputDevices() &&
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Fix alignment.

Done.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
File media/audio/audio_unittest_utils.cc (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.cc:1: #include "base/command_line.h"
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Needs license header.

Done.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.cc:14: *should_fail = fail_if_unsatisfied;
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Should always be set?

May as well

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
File media/audio/audio_unittest_utils.h (right):

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:1: #include
"testing/gtest/include/gtest/gtest.h"
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Needs license header. #ifdef protections for multiple include.

Done.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:1: #include
"testing/gtest/include/gtest/gtest.h"
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Needs license header. #ifdef protections for multiple include.

Oh jeez, I forgot a lot of stuff. I put it in the media ns, and removed the
plural from 'utils' to be consistent with the rest of media/.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:4: // required audio device or library. If
the |kRequireAudioHardwareForTesting|
On 2015/02/09 22:10:50, DaleCurtis wrote:
> I'd use the actual --flag-name.

Done.

https://codereview.chromium.org/914483002/diff/20001/media/audio/audio_unitte...
media/audio/audio_unittest_utils.h:19: bool should_abort_audio_test(bool
requirements_satisfied,
On 2015/02/09 22:10:49, DaleCurtis wrote:
> Should be CamelCase method name.

Done.

[DaleCurtis, 2015-02-11 19:29:35]

lgtm

[watk, 2015-02-11 19:35:21]

The CQ bit was checked by watk@chromium.org

[commit-bot: I haz the power, 2015-02-11 19:35:58]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/914483002/180001

[commit-bot: I haz the power, 2015-02-11 20:24:52]

Committed patchset #9 (id:180001)

[commit-bot: I haz the power, 2015-02-11 20:25:49]

Patchset 9 (id:??) landed as
https://crrev.com/b0ddebe2ea0b00c272af7820b9c108c4438f5b88
Cr-Commit-Position: refs/heads/master@{#315823}