Mobile/Testing/05 01 13
From MozillaWiki
Contents
Previous Action Items
- (jmaher) explain your round table items
-
(ctalbert) get kim a known good build - (kim) run tests over the weekend
- (ctalbert) to email bad news
- (kim) split out webgl from mochitest-1
- (wlach) to add stock and chome to new test
- Mostly done... see below
- (blassey) follow up with karen to get 2.2 end of life plan
Status reports
Dev team
- Trying to eliminate bug 663657 "2400 seconds without output" failures
- sutAgent updated (thanks again :kmoir!)
- bug 865944 Use foreground activity instead of process name to determine if remote browser has terminated
- bug 865006 Run more xpcshell tests on Android
Rel Eng
- (kmoir) bug 866228 - deploy sutagent 1.17 to the tegras and pandas (done), bug 865677 - rerun android panda tests on newly rewired chassis with fennec 19.0 (same results as production), working on bug 865443 create webgl mochitest run for android
- (Callek) bug 861295 Modify --testing-modules-dir for android xpcshell tests - landed
IT
- Still working on a higher density chassis. Just waiting for the prototype chassis to be fabricated.
A Team
x86 automation
- I've now run through all of the test types, tracking in bug 828571
- Crashtests and JSRefTests seem stable, so I've filed a bug to get started on running this in Cedar
- Reftests have a small number of unexpected results (20) but should also be straight forward to get running
- Mochitest and robocop have timeouts and larger number of unexpected results
- Robocop needs to specify the user to the ActivityManager since the emulator images I'm using are 4.2
- Talos tests will run.
- tsvg, ts_paint work ok
- tp4m runs for a long time, then has problems parsing the output
- haven't tried robocop based talos tests yet
- I have been running some of the tests in parallel with two emulators running on my machine
- In one case I got timeouts in both at the same time due to high CPU load
- I'm planning to focus on getting mochitests running properly next
Autophone
- [bc] Working on improving reproducibility of results in bug 866195. Will post test runs for the last several days with production phones to http://mrcote.info/phonedash_bc for comparison with previous production results at http://mrcote.info/phonedash later today.
Eideticker
- Comparative numbers for Chrome on pageload-post-startup test (forgot to generate for stock, will do later today)
- http://eideticker.wrla.ch/#/samsung-gn/nytimes-load-poststartup/timetostableframe
- Numbers indicate that pages are slow to load even after the browser has started. We are very much not competitive here.
- Working on getting HTTP request times added to dash
- Profiling support broke last week which meant no items added to galaxy nexus dash for a bit (bug 865915). Now fixed.
Round Table
Action Items
- (jmaher) figure out why our tegra orange rate went up
- one build from April 30 yielded a series of blue (retries) across most all tests
- reftest 1-4 seem to have an orange and red for each chunk, this is different than previously. If trend continues, we will need to look into this more
- (kim) split webgl tests out from mochitest-1
- (blassey) follow up with necko team on the new eideticker tests