Add a startup benchmark and rudimentary scripts for running benchmarks.

Add a startup benchmark and rudimentary scripts for running benchmarks.

BUG=None
TEST=None
R=abarth@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/5b7319a91ea9c3613cdd6ce73dac3ea62479cf38

[yzshen1, 2014-11-14 00:52:50]

yzshen@chromium.org changed reviewers:
+ abarth@chromium.org

[yzshen1, 2014-11-14 00:52:51]

Hi, Adam.

Would you please take a look? Thanks!

https://codereview.chromium.org/724243002/diff/1/benchmarks/startup/run.py
File benchmarks/startup/run.py (right):

https://codereview.chromium.org/724243002/diff/1/benchmarks/startup/run.py#ne...
benchmarks/startup/run.py:15: __file__, os.pardir, os.pardir, os.pardir, 'mojo',
'tools')))
I am not quite sure whether this is allowed.
Please advise. Thanks!

[abarth-chromium, 2014-11-14 04:44:13]

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
File benchmarks/benchmark_runner.py (right):

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
benchmarks/benchmark_runner.py:9: import subprocess
Unused

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
benchmarks/benchmark_runner.py:24: return tests
I probably would have used a generator.

https://codereview.chromium.org/724243002/diff/20001/benchmarks/startup/run.py
File benchmarks/startup/run.py (right):

https://codereview.chromium.org/724243002/diff/20001/benchmarks/startup/run.p...
benchmarks/startup/run.py:24: paths = Paths(build_directory)
Why not just pass in |paths| instead of making the benchmark do this work.  It
seems like every benchmark will need to repeat the boiler plate.

[abarth-chromium, 2014-11-14 04:44:37]

On 2014/11/14 at 00:52:51, yzshen wrote:
> Hi, Adam.
> 
> Would you please take a look? Thanks!
> 
> https://codereview.chromium.org/724243002/diff/1/benchmarks/startup/run.py
> File benchmarks/startup/run.py (right):
> 
>
https://codereview.chromium.org/724243002/diff/1/benchmarks/startup/run.py#ne...
> benchmarks/startup/run.py:15: __file__, os.pardir, os.pardir, os.pardir,
'mojo', 'tools')))
> I am not quite sure whether this is allowed.
> Please advise. Thanks!

We should ask James what the proper way of finding mopy is.

[abarth-chromium, 2014-11-14 04:44:59]

This seems like a good start.

[yzshen1, 2014-11-14 18:00:15]

Hi, Adam.

Thanks for the review! Please take another look.

> We should ask James what the proper way of finding mopy is.

I asked James and he said he is not very familiar with the right python approach
to find modules, but he think it is reasonable to use mopy.

Besides, we have used this way elsewhere, for example, in:
mojo/public/tools/bindings/pylib/mojom/parse/parser.py

So I will leave it as it is.

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
File benchmarks/benchmark_runner.py (right):

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
benchmarks/benchmark_runner.py:9: import subprocess
On 2014/11/14 04:44:13, abarth wrote:
> Unused

Done.

https://codereview.chromium.org/724243002/diff/20001/benchmarks/benchmark_run...
benchmarks/benchmark_runner.py:24: return tests
On 2014/11/14 04:44:13, abarth wrote:
> I probably would have used a generator.

Done.

https://codereview.chromium.org/724243002/diff/20001/benchmarks/startup/run.py
File benchmarks/startup/run.py (right):

https://codereview.chromium.org/724243002/diff/20001/benchmarks/startup/run.p...
benchmarks/startup/run.py:24: paths = Paths(build_directory)
On 2014/11/14 04:44:13, abarth wrote:
> Why not just pass in |paths| instead of making the benchmark do this work.  It
> seems like every benchmark will need to repeat the boiler plate.

Done.

[abarth-chromium, 2014-11-14 22:20:38]

lgtm

[yzshen1, 2014-11-14 22:22:24]

Committed patchset #3 (id:40001) manually as
5b7319a91ea9c3613cdd6ce73dac3ea62479cf38 (presubmit successful).