e10s Update: May 20
Executive Summary
- The biggest news is that the e10s Experiment for Firefox 47 Beta has been approved to run for the entire Beta cycle (vs. being disabled at Beta 6). This means that stability is trending in the right direction and that RelMan has been provided the Telemetry-based stability dashboard they need to easily decipher e10s vs. non-e10s crashes. The experiment will end when Firefox 47 goes to GA on June 7 and we will repeat a similar experiment for Firefox Beta 48 with an eye towards shipping to 1% of our Firefox 48 GA population so long as we meet our Release Criteria.
- As of this week, 800,000 users have been exposed to e10s in Beta. Our two biggest risk areas remain stability and slow script count. Stability has benefited from continued focus on fixing top crashes and uplifting them. Through continued analysis, we've found landed on some interesting conclusions which lead us to be optimistic about progress, see information, here.
Next Steps
- 05/20: Firefox 47 Beta 7 Ships
- 05/25: Fourth derived data set available (4 weeks of data)
- 05/24: Release plan review
Release Criteria
- M9 Count
- We are down to 3 M9's.
- Stability
Analyzing stability data is an ongoing effort, and we have found OS versions make a difference:
- XP Crash Rates are bringing our numbers up; excluding NT data indicates that e10s is more stable than non-e10s (see this query).
- Additionally, we have found that stability is worse for Mac OS X 10.6-10.8. These OS versions will be no longer supported as of Firefox 49.
- Next Steps include excluding experiment participants who are using these OS versions and see how it impacts our crash stats as well as figuring out if graphics acceleration is part of this (also for non-e10s).
- Slow script count increased up to 0.11 from week 2's 0.09
- We will be doing an analysis similar to the once cited above to help investigate
Add-Ons + WebExtensions
- Decision to not have an allow list has been made, we will take a more efficient approach. Only add-ons that have self identified as e10s compliant can be part of the e10s activation audience at first -in addition to all add-ons that are webextensions. See bug 1249060 for more details.
- General roll-out plan for including extensions is here.
Automated tests
- Everything we can have enabled in production is enabled, everything else is enabled on a separate branch called 'Ash'.
- We have all tests be owned
- We would like each directory signed-off by test owners week of 06/05
- We're tracking all the tests and who owns them, here
- We are tracking tests that are running and passing on some platforms, but are disabled on others.From a feature coverage point of view, it's in pretty great shape. There's very little that is disabled both on opt *and* debug on a same platform (meaning that that test is not running on this platform). Look for the "Fully disabled in a platform" entries in this spreadsheet.
A11y
- Proposed technical path to encourage a11y clients to communicate directly with the content process bug 1258839. dbolter is getting some first impressions from clients to see if this is a viable approach.
- We're also looking into Windows touchscreen/a11y compat, see more information as the plan is taking shape, here.
Milestones
We are at the point where we are setting milestones weekly; given the complexity of deliverables, we are using a Trackback Schedule to keep everything organized.
Release Schedule
| Date | Trunk | Aurora | Beta | Release |
|---|---|---|---|---|
| 3-07 | 48 default | 47 default | 46 A/B Tests | 45 off |
| 4-18 | 49 default | 48 default | 47 A/B Tests, Possible Rollout (add-ons = ally = no) | 46 off |
| 6-07 | 50 default | 49 default | 48 Possible Rollout (add-ons + ally = no) | 47 off |
| 8-02 | 51 default | 50 default | 49 Possible Rollout (add-ons + ally = no) | 48 Possible Rollout (add-ons + ally = no) |
| 9-13 | 52 default | 51 default | 50 Possible Rollout (add-ons + ally = no) | 49 Possible Rollout (add-ons + ally = no) |