Android perf runner: print the command line in the step itself.

Android perf runner: print the command line in the step itself.

Follow up from https://codereview.chromium.org/22904005/,
on top of printing in the first "sharder" step, also print the command
at the beginning of the "print" step.

BUG=268450
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=219303

[bulach, 2013-08-19 09:22:52]

ptal

[bulach, 2013-08-22 08:33:11]

ping? :)

[frankf, 2013-08-22 17:41:12]

lgtm w/ questions

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
File build/android/pylib/perf/test_runner.py (right):

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
build/android/pylib/perf/test_runner.py:75: print '*' * 80
why not logging?

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
build/android/pylib/perf/test_runner.py:116: output, exit_code = pexpect.run(
BTW, have you seen any issue with pexpect terminating? I ran into an this issue
for another CL downstream and it seems like a bug in pexpect itself.

[bulach, 2013-08-23 09:13:54]

thanks frank! comment addressed, clarification inline:

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
File build/android/pylib/perf/test_runner.py (right):

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
build/android/pylib/perf/test_runner.py:75: print '*' * 80
On 2013/08/22 17:41:12, frankf wrote:
> why not logging?

done (the last one has to be print as it was, since it's parsed the buildbot).

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
build/android/pylib/perf/test_runner.py:116: output, exit_code = pexpect.run(
On 2013/08/22 17:41:12, frankf wrote:
> BTW, have you seen any issue with pexpect terminating? I ran into an this
issue
> for another CL downstream and it seems like a bug in pexpect itself.

yeah, sometimes exit_code is None (as per the other patch)..
I hope this will help identify which step is happening...
tbh, the only reason for pexpect here is just so it can "stream" the output to
sys.stdout _and_ capture to save to the file...

if you have any other suggestions, I'd be more than happy to use it instead!! 

hmm... actually... maybe apply a file-like object and pass that to stdout param
on subprocess.Popen? let me try that, I think we used this trick somewhere
else...

[commit-bot: I haz the power, 2013-08-23 09:15:12]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bulach@chromium.org/23316004/13001

[bulach, 2013-08-23 11:25:07]

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
File build/android/pylib/perf/test_runner.py (right):

https://codereview.chromium.org/23316004/diff/1/build/android/pylib/perf/test...
build/android/pylib/perf/test_runner.py:116: output, exit_code = pexpect.run(
On 2013/08/23 09:13:55, bulach wrote:
> On 2013/08/22 17:41:12, frankf wrote:
> > BTW, have you seen any issue with pexpect terminating? I ran into an this
> issue
> > for another CL downstream and it seems like a bug in pexpect itself.
> 
> yeah, sometimes exit_code is None (as per the other patch)..
> I hope this will help identify which step is happening...
> tbh, the only reason for pexpect here is just so it can "stream" the output to
> sys.stdout _and_ capture to save to the file...
> 
> if you have any other suggestions, I'd be more than happy to use it instead!! 
> 
> hmm... actually... maybe apply a file-like object and pass that to stdout
param
> on subprocess.Popen? let me try that, I think we used this trick somewhere
> else...

for completeness: unfortunately subprocess.Popen doesn't take a "file-like"
object, it really needs a "file" that implements a fileno() with the underlying
FD so that the child process can write directly to, which makes sense...

so there's no easy way to "capture" the stdout whilst streaming it to sys.stdout
without, well, reimplementing this pexpect.run... :-/

anyways: once we know at least which step is failing, I hope it'll give more
insights.

[commit-bot: I haz the power, 2013-08-23 12:40:37]

Retried try job too often on linux_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[commit-bot: I haz the power, 2013-08-23 14:52:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bulach@chromium.org/23316004/13001

[commit-bot: I haz the power, 2013-08-23 17:55:06]

Change committed as 219303