Closed Bug 915730 Opened 12 years ago Closed 12 years ago

Add responsiveness metrics to all Android tests with actions

Categories

(Testing Graveyard :: Eideticker, defect)

x86_64
Linux
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: wlach, Assigned: wlach)

Details

Attachments

(2 files)

We can measure this for all existing Android tests, and it's just another small addition to the datapoint for each test. This is a prelude to doing the same thing for FirefoxOS.
Various cleanup to make this possible/easier.
Attachment #804097 - Flags: review?(gbrown)
Attached patch The actual patchSplinter Review
Attachment #804098 - Flags: review?(gbrown)
Attachment #804097 - Flags: review?(gbrown) → review+
Comment on attachment 804098 [details] [diff] [review] The actual patch Review of attachment 804098 [details] [diff] [review]: ----------------------------------------------------------------- Looks great. I'm eager to see some initial results. ::: src/eideticker/eideticker/metrics.py @@ +1,3 @@ > +import videocapture > + > +INPUT_RESPONSE_THRESHOLD = 10000 Why 10000? My naive instinct would be to use something smaller. Do you expect much noise (framediffs without an action, or independent of action)? @@ +18,5 @@ > + if first_non_sleep_action: > + framediffs = videocapture.get_framediff_sums( > + capture, threshold=difference_threshold) > + for (i, framediff) in enumerate(framediffs): > + t = i/60.0 Is there an existing constant that could be used instead of "60.0"?
Attachment #804098 - Flags: review?(gbrown) → review+
(In reply to Geoff Brown [:gbrown] from comment #3) > ::: src/eideticker/eideticker/metrics.py > @@ +1,3 @@ > > +import videocapture > > + > > +INPUT_RESPONSE_THRESHOLD = 10000 > > Why 10000? My naive instinct would be to use something smaller. Do you > expect much noise (framediffs without an action, or independent of action)? I noticed some small variations before the real response in taskjs, but I agree 10000 is probably too high. Set it to 4096 for now -- we can always adjust later. > @@ +18,5 @@ > > + if first_non_sleep_action: > > + framediffs = videocapture.get_framediff_sums( > > + capture, threshold=difference_threshold) > > + for (i, framediff) in enumerate(framediffs): > > + t = i/60.0 > > Is there an existing constant that could be used instead of "60.0"? We should probably make the fps a property of the capture. Changed that before committing.
Product: Testing → Testing Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: