EngineeringProductivity/Projects/Stockwell/Meetings/2016-10-07: Difference between revisions
Jump to navigation
Jump to search
m (→New Ideas to investigate: - note) |
m (→New Ideas to investigate: -typo) |
||
| (8 intermediate revisions by 3 users not shown) | |||
| Line 3: | Line 3: | ||
= Status = | = Status = | ||
* autostar has simple matching working and is live | |||
= Discussion Topics = | = Discussion Topics = | ||
* [jmaher] Defining Intermittent Oranges | * [jmaher] Defining Intermittent Oranges | ||
** [[https://elvis314.wordpress.com/2016/10/10/working-towards-a-productive-definition-of-intermittent-orange/ initial definition]] and thoughts | |||
* [gbrown] Disabling Intermittent Tests | * [gbrown] Disabling Intermittent Tests | ||
** [[https://gbrownmozilla.wordpress.com/2016/10/07/skipping-persistent-intermittent-failures/ blog post]] and thoughts | |||
* [ted] Attempting to reproduce failures with rr? | |||
** [https://public.etherpad-mozilla.org/p/reproducing-intermittent-failures-rr notes from a previous meeting on this topic] | |||
** [https://gist.github.com/jmaher/c3f89ec173dfe7b35e81 hacky script to scrape orangefactor for a list of possibly reproducible bugs] | |||
= New Ideas to investigate = | = New Ideas to investigate = | ||
* | * consider all mochitest- tests in orangefactor to be run in parallel in a quarantine job? ** this would allow us to see the runtime and what tests fail | ||
** we would also understand more issues with creating this | |||
** easy to do on linux64/android in taskcluster, not easy for osx/windows | |||
** QUESTION: if a test is flaky on one platform, should we consider it flaky on all platforms? | |||
* | |||
= Action Items = | = Action Items = | ||
Latest revision as of 16:21, 11 October 2016
Previous Action Items
- N/A
Status
- autostar has simple matching working and is live
Discussion Topics
- [jmaher] Defining Intermittent Oranges
- [initial definition] and thoughts
- [gbrown] Disabling Intermittent Tests
- [blog post] and thoughts
- [ted] Attempting to reproduce failures with rr?
New Ideas to investigate
- consider all mochitest- tests in orangefactor to be run in parallel in a quarantine job? ** this would allow us to see the runtime and what tests fail
- we would also understand more issues with creating this
- easy to do on linux64/android in taskcluster, not easy for osx/windows
- QUESTION: if a test is flaky on one platform, should we consider it flaky on all platforms?
Action Items
- Discussion topics for next time