Releases/Post-mortems/Firefox 19.0

From MozillaWiki
Jump to: navigation, search

l== Schedule / Location / Call Information ==

  • Tuesday, 3/5/2013 @ 9:00 am PT
  • Physical spaces: MTV-3G, SFO-7N, TOR-5G
  • Vidyo Room: Release Coordination

Other communication channels

  • irc.mozilla.org #planning for back channel
  • This etherpad for real-time notes (we'll copy the wiki page in at the beginning of the meeting, and out at the end)

Previous Actions

  • joduinn - to follow up with IT about CDN pull experiment, so that we're no longer afraid to push to mirrors (we should also document the process)
  • ashughes - make sure that new locales for FF19 won't break testing (more of a reminder)
  • QA - Bug 831686 - Create Mozmill test for proxy.pac file served via a file:// protocol
  • [Relman] https://bugzilla.mozilla.org/show_bug.cgi?id=811054 discuss with Axel and get back to releng

Talking Points

Timeline

19.0

  • No changes were taken from FF19.0b6 (built 2/12), and 19.0b6 only included some minor changes (keyword reset feature disable, padding for bug 839270, last minute )
  • Final sign-off provided on 2/13, https://wiki.mozilla.org/Releases/Firefox_19/Final_Signoffs
  • Go to build sent 2/15, 11:53 AM PT
  • Push to mirrors sent 2/18, 4:40 PM PT
  • Snippets pushed out on 2/19, 5:56 AM PT (unthrottled 50%)
  • Win8 updates were disabled across all channels 2/20 AM PT due to early data from bug 830531 – [Win8] crash in XPC_WN_Helper_NewResolve mainly with AMD Radeon HD 6290/6310/6320/7290/7310/7340 (Wrestler Asic)
  • Throttled before EOD 2/20
  • FF20.0b1 was delayed till Monday to add additional logging for the 19.0beta->20.0beta hump
  • Unthrottled for non-Win8 on 2/26 PM PT

19.0.1 (Win8 only)

  • Go to build sent 2/26, 4:18 PM PT - no changes taken from 19.0, only a respin for bug 830531
  • Positive testing from external reporters and QA allowed for push to mirrors to be sent 2/27, 2 PM PT
  • Snippets were pushed out unthrottled on 2/28 at 9:45AM PT alongside https://www.mozilla.org/en-US/firefox/19.0.1/releasenotes/
    • Also updated download buttons, bug #

Release Issues

  • Softvision delays caused mobile to ship late by >2hr
    • "The cause of the delay was an external automation tool. We are considering making this a non-blocking process." -kbrosnan
    • Release has now been clarified as a priority over Aurora/Beta to SoftVision
    • This test is non-blocking
  • Uptake was still too fast for FF19.0
    • Based upon the last two releases, Annie is suggesting 50% for 24 hours, and 25% for 12 hours after that. We may just go out 25% from the start next time.
      • Clients will check for updates up to 3 times in 36 hours. If we throttle at 25%, then a client will have up to
  • The most concerning issues remain on the tracking list at bit.ly/XKQq2z
  • 791958 – android.database.sqlite.SQLiteDatabaseLockedException: database is locked (code 5): , while compiling: PRAGMA journal_mode at android.database.sqlite.SQLiteConnection.nativePrepareStatement(Native Method) on JB
    • Dropped in volume, so was deemed a wontfix for FF19
  • 830531 – [Win8] crash in XPC_WN_Helper_NewResolve mainly with AMD Radeon HD 6290/6310/6320/7290/7310/7340 (Wrestler Asic)
    • 19.0.1 verifiably resolved this crash
  • A couple of other 3rd-party specific topcrashes (FoxyProxy, Ubuntu Global Menu)

PDF.js

  • 843217 – Intuit Sites incompatible with PDF.js
    • Intuit tech support was happy with the resolution of removing headers for FF20
  • 843342 – PDF doesn't print correctly on USPS.com with pdfjs (pdf.js does not support automatic print using Adobe JavaScript)
    • Still performing outreach

IE10 Release Issues (D3D11)

  • 812695 – [D2D] Text Rendering Issues due to Windows 7 Platform Update KB2670838 (MSIE 10 Prerequisite)
  • 812683 – [Win7] startup crash in _d2d_clear_surface @ CContext::RestorePipelineStateImpl (d3d11.dll, after installing KB2670838)

QA

Bug 839270 crash in nsStyleContext::nsStyleContext mainly with AMD Radeon HD 6250/6310/6320

  • Manual testing not 100% confident, not enough crashstats data to make a fair judgement
  • What can be done to:
    • A) Have better confidence in the protections we put in place
    • B) Work with GPU vendors to improve the current situation

Bug 830531 - [Win8] crash in XPC_WN_Helper_NewResolve

Mobile QA

This is Fx 20 - 832942 – Searches to Google.com over SSL cause OOM error page on ARMv6 devices

RelEng

  • [akeybl] Exploring re-packing as opposed to re-spinning, so that 19.0rc1 could be nearly identical to 19.0b6

Security

Marketing & Press

Stability

  • Can we start emailing users who provided their email when we know 3rd-party software correlations or that we've resolved the instability issue?

Support

Other

  • How can we prevent going out with something like:
    • bug 830531 – [Win8] crash in XPC_WN_Helper_NewResolve mainly with AMD Radeon HD 6290/6310/6320/7290/7310/7340 (Wrestler Asic)
    • Bug 839270 crash in nsStyleContext::nsStyleContext mainly with AMD Radeon HD 6250/6310/6320

17.0.3esr

  • 2/25 unthrottled esr10 -> esr17 updates

Actions

  • bhearsum/joduinn - to follow up with IT about CDN pull experiment, so that we're no longer afraid to push to mirrors (we should also document the process)
  • akeybl - to follow up with axel about adding an item to the release calendar for l10n changeset pulling
  • juanb to follow up with ashughes about Bug 831686
  • STR for bug 830531 and bug 839270 will be added to testplan by ashughes/juanb
  • bhearsum to follow up about "re-packing as opposed to re-spinning, so that 19.0rc1 could be nearly identical to 19.0b6"
  • bajaj to follow up with crash-kill about "Can we start emailing users who provided their email"