Merge lp://qastaging/~kaxing/checkbox/bt4-bluez5 into lp://qastaging/checkbox

I proposed a few improvements over tiny details inline.

@kissiel Thanks for the review, will update base on your feedbacks. But for naming about unpairing() still needs a bit of advising. Please check.

One small note for the break in the for loop

Update few in-line comment replies.

Following up the discussion in the inline comments

replies updating!

Updating everything based on previous in-line comments.

For print(flush=True) I've file a new bug for further discussing.(if anyone interested)
https://bugs.launchpad.net/plainbox/+bug/1569808

Verified on a targeting system(bt4, xenial/bluez5), everything works as expected.

Only one kind of failure the script won't able to capture:
the host system/bluez thinks it's paired but the device still blinking.
but I think this is left to tester's judgment as they are user-interact-verify jobs.

I've proposed a workaround to the print(flush=True), see my inline suggestion.

idem for the mouse job

I've removed the part that requires instant flush, so the bt_connect now is not effected to the buffered issue. those print() within flush=True is working as expected.

Hi Yung,
from the revision history, it seems that the branch was not re-submitted successfully, the latest one is still rev 4276, Apr. 13

Hey @cypressyew, that's about right, I didn't made any changes since I've already removed the "buffered" part that mentioned in previous comments in revno 4276.

I tried to run the new tests using:

plainbox run -i 2013.com.canonical.plainbox::collect-manifest -i .*bluetooth4.*

I didn't work for two reasons:

1. Manifest entries does not support the requires statement
2. The two new jobs needs the following line to use manifest properly:

imports: from 2013.com.canonical.plainbox import manifest

Thanks for the hint, fixed manifest and add bug number in comment for buffered print().

One minor fix and it should be good to go, see below

And I think we could remove the TODO part in the comment, as this is quite mature now. (Convert print to logging might be the last thing)

So I tried logging with level=logging.info it appears there are a lot informations going on, And I think we will need to discuss this, will keep my fake logging outputs at the moment so the cli work as expected.

Those logging info could indeed be very useful. Let's merge this version and iterate.

A big +1 to all contributors. Thanks