- Go/no-go status
- Merge mechanics
- How do crashes look?
- How do our Aurora triage queries look?
- Is there anything outstanding we need to land today?
- l10n status?
- Add-on compatibility implementation status?
- Web content
At this point we are not aware of any major outstanding issues that would block the Aurora to Beta transition. Most of our focus has been on bug verifications and making sure the most prominent feature of this release, channel switcher, is in good shape. In addition we have kept monitoring crash data and we haven't seen any significant spikes in the crashes per ADU rate.
- We have two concerns: add-on compatibility and the relatively low number of users in Aurora
- We are at 29% in add-con compatibility (number from desktop report sent out by Laura)
- We only have about 30k users on Aurora, compared to about 40k in the Nightly, and 100k on 5.0beta.
- Main Feature Sign-off: Channel switcher is the main feature for this release
- We've gone through the channel switcher end to end, during the test Aurora-to-Beta merge, using the test cases in the feature page and found no major problems.
- One outstanding issue worth mentioning is the "update failed" bug 652016.
- Crash Stats: Crashes per user have been hovering below the %4 mark in the Aurora channel. There aren't any significant top movers at this moment. And the statistics are very comparable between 4.0.1, 5.0, 5.0a2, and 6.0a1: http://bit.ly/mLXiZW
- The graph of "issues" in the input site, corresponding to 5.0a2 has been relatively stable. At a glance, most of the issues have to deal with changes in the look and feel or common problems already in 4.0. There are several instances of people complaining about the "update failed" problem filed as bug 652016, and it's one of the frequent complaints for this channel.
- Crash stats
- Nothing new on Aurora we are worried about but we have less users than on Beta.
- JP - Aurora and Beta are higher than release.
- Some of the top crashers have just been fixed and not reflected in the data.
- Crash levels similar to previous "pre-beta" periods.
- 6 tracking bugs
- All have been fixed in Aurora.
- Localization - anything that would block a Beta transition at this point?
- Apparently the Aurora l10n repositories didn't have a single head - had a ghook. Once that's in we need to verify this.
- Does this block doing the merge or QA on the merge?
- It probably blocks doing the builds. It breaks a bunch of assumptions and we are not sure how bad it is. The only way to find out is to do builds and figure it out.
- We will probably just proceed.
- Worst case - it pulls from the wrong head and breaks the build at a place where we can't recover - break the l10n repack.
- Christian is going to add this on his checklist add let us know if it all worked out.
- Add on compatibility
- We were originally planning on getting the last blocker bug fixed thurs/fri.
- We are aiming to push to production Wed but it hasn't been QA'd. Might be delayed due to the blocker.
- Contingency plan - we will go anyway with a throttled update.
- We said we will go out throttled by the end of this week even if the add on stuff isn't ready.
- Updates will go to the 4.0.1 users as well as other Beta users.
- Pretty much ready to go.
- Waiting on getting the future releases blog updated.
- Not a blocker for doing the "go"
- First run page - done and can get it pushed today.
- What's new page - needs some placeholder. Can we redirect? Grace will log
- Updating the channel page - not blocking the merge or the update but need to get this done.
- l10n what's new and first run pages - working on them, not all of them are ready but not a blocker.
- This merge should be simple.
- Feedback needed on the best time to merge. Christian will probably post to dev planning. Last time we did it at 7:00am PDT.
- Using the merge checklist.
- l10n-aurora/ isn't being merged to l10n-beta/, we are copying over repos (see bug 657417).
- Christian will be sending emails to release-drivers and developers list and posting to the blog with start/finish details.
- Shadow repo - nothing related to Beta. We don't have the data we need to make sure the policy we have in place will work.
- Christian will send updates to the list about where we are in the process.
- We'll want a build after the merges are done
- We do not want the update to be available anywhere except betatest yet
- The timing isn't down-to-the-minute critical. No need for RelEng to kill themselves to save seconds on the build
- treating as non-chemspill
- RelEng needs to schedule a downtime when not disruptive to drops; firmware upgrade on stage.m.o/ftp.m.o & DNS/DHCP
- Take this issue to tomorrow's call - not really related to Beta.