Merge lp://qastaging/~zyga/lava-android-test/blackbox into lp://qastaging/lava-android-test

from this file itself, it's no problem.
but I still feel that make it a new sub command(like run-blackbox) in the commands.py
will make the source more readable.
although that needs to modify more files (including lava-dispatcher) instead of only this one file.

And also suggest adding a provider in the provider.py to support such tests that only need to run and get the test result.
no need to use the parser to parse the output.

and maybe this is just a temporary place to integrate the blackbox test,
since I can't see how the blackbox will be in the future, so please also ask Andy to review this MP.

W dniu 29.08.2012 08:34, Yongqin Liu pisze:
> from this file itself, it's no problem.
> but I still feel that make it a new sub command(like run-blackbox) in the commands.py
> will make the source more readable.
> although that needs to modify more files (including lava-dispatcher) instead of only this one file.
>
> And also suggest adding a provider in the provider.py to support such tests that only need to run and get the test result.
> no need to use the parser to parse the output.

The only long term solution is to fix the API mistakes and remove all
other subcommands. There is no justification for having all those
subcommands. All of that are tests, they should have single install/run
interface.

> and maybe this is just a temporary place to integrate the blackbox test,
> since I can't see how the blackbox will be in the future, so please also ask Andy to review this MP.

OK

Thanks
ZK

Seems like some of the code isn't that necessary:

debuggable_* - cool, but now that the feature works is it needed?

SuperAdb/AdbMixin - seems like you could pretty much just add a new listdir helper

However, that's nothing major and its isolated. Also, as Yongqin said, the blackbox stuff will eventually move us to something else, so lets get this in and see it work!

W dniu 29.08.2012 17:52, Andy Doan pisze:
> Review: Approve
>
> Seems like some of the code isn't that necessary:
>
> debuggable_* - cool, but now that the feature works is it needed?

No, it's not needed but it's still a cool thing (since l-a-t swallows
most exceptions) for analyzing problems. Perhaps I could just move it to
some utils.py module?

> SuperAdb/AdbMixin - seems like you could pretty much just add a new listdir helper

Yes, that behaves much like os.listdir() and has more predictable
behavior. It's just a helper that does parsing essentially :-)

> However, that's nothing major and its isolated. Also, as Yongqin said, the blackbox stuff will eventually move us to something else, so lets get this in and see it work!

I don't know how it will evolve over time. My current intent is to
ensure we can continue to harvest data from l-b with l-a-t, and in near
future with l-t. Once the rest of the framework solidifies and matures
we could simplify the interaction but I don't suspect we'll go away form
the simple usefulness of adb-managed communication. It simply works very
well and has little complexity.

Thanks
ZK