Record/replay support to test.py

Record/replay support to test.py

R=ricow@google.com

Committed: https://code.google.com/p/dart/source/detail?r=23216

[kustermann, 2013-05-24 16:11:25]

Request for initial feedback.

[kustermann, 2013-05-27 07:45:54]

Current usage would be like this:

user@host ~/dart $./tools/test.py -mrelease --record_to_file=recording.json
standalone/io/msk --timeout=2
user@target ~/dart $./tools/execute_recorded_testcases.py recording.json
recording_output.json
user@host ~/dart $ ./tools/test.py -mrelease
--replay_from_file=recording_output.json standalone/io/msk

[ricow1, 2013-05-27 09:05:17]

LGTM

https://codereview.chromium.org/15944005/diff/1/tools/testing/dart/test_runne...
File tools/testing/dart/test_runner.dart (right):

https://codereview.chromium.org/15944005/diff/1/tools/testing/dart/test_runne...
tools/testing/dart/test_runner.dart:1430: // FIXME: do we want to call
[addTestCase]???
FIXME -> TODO
I would probably organize this a little differently to make it easier to follow:
var newTest;

if (_testCaseRecorder != null) {
  TODO..
  newTest = _testCaseRecorder.nextTestCase;
} else {
  _tests.add(testCase);
  eventTestAdded(testCase);
  if (_testCaseOutputArchive != null) {
    newTest = (testCase) {
      Timer...
    }
      
  } else {
    newTest = _runTest;
  }
}

https://codereview.chromium.org/15944005/diff/1/tools/testing/dart/test_runne...
tools/testing/dart/test_runner.dart:1435: Timer.run(() {
add comment why we do this

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
File tools/execute_recorded_testcases.py (right):

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:9: did_timeout = False
Write the usage pattern in a comment here.

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:13: global did_timeout
why don't you send this back in the tuple like the rest of the "result"

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:21: global did_timeout
this is already global

[kustermann, 2013-05-27 11:11:37]

PTAL

https://codereview.chromium.org/15944005/diff/1/tools/testing/dart/test_runne...
File tools/testing/dart/test_runner.dart (right):

https://codereview.chromium.org/15944005/diff/1/tools/testing/dart/test_runne...
tools/testing/dart/test_runner.dart:1430: // FIXME: do we want to call
[addTestCase]???
On 2013/05/27 09:05:17, ricow1 wrote:
> FIXME -> TODO
> I would probably organize this a little differently to make it easier to
follow:
> var newTest;
> 
> if (_testCaseRecorder != null) {
>   TODO..
>   newTest = _testCaseRecorder.nextTestCase;
> } else {
>   _tests.add(testCase);
>   eventTestAdded(testCase);
>   if (_testCaseOutputArchive != null) {
>     newTest = (testCase) {
>       Timer...
>     }
>       
>   } else {
>     newTest = _runTest;
>   }
> }
>    

Done.

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
File tools/execute_recorded_testcases.py (right):

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:9: did_timeout = False
On 2013/05/27 09:05:17, ricow1 wrote:
> Write the usage pattern in a comment here.

Done.

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:13: global did_timeout
On 2013/05/27 09:05:17, ricow1 wrote:
> why don't you send this back in the tuple like the rest of the "result"

I think we need to make it global. Otherwise 'timeout_handler()' has no way of
communicating this back to 'run_command()'.

EDIT: I found now a way to do it :-)

https://codereview.chromium.org/15944005/diff/7001/tools/execute_recorded_tes...
tools/execute_recorded_testcases.py:21: global did_timeout
On 2013/05/27 09:05:17, ricow1 wrote:
> this is already global

I think global has to be set on every new scope. But we don't need it any
longer.

[ricow1, 2013-05-27 12:18:28]

Still LGTM

[kustermann, 2013-05-27 12:23:31]

Thank you.

[kustermann, 2013-05-27 12:23:55]

Committed patchset #3 manually as r23216 (presubmit successful).