CrashKill/2013-04-08

General

Socorro

• Shipped a very minor release last week due to release week (forced correlations)
• Completed PG upgrades smoothly on Thursday: 17 minutes of user facing downtime
• Trying to ship twice this week: today (new processor, ES indexing) and Thursday (b2g stuff)
• Tracker for Django migration - 11 blockers right now

Flash

Desktop

Release

• Crossrider bug 836263 is the #4 crash in 20.0 release now.
• bug 851934 (outlook.com crash) is #5.
• D3D/D2D11-6.2/Win7 crashes bug 819028 and bug 812683 are at #7 and #10.
• EMTPY crash dumps are at 20% of all crashes, bsmedberg seems to have some ideas in bug 837835 on some of the increases.

Beta

• bug 837370 (js::types::TypeCompartment::addPendingRecompile) is at #5 in 21.0b1, we should see to get the apparent fix into beta.
• bug 852915 (mozilla::MediaDecoderReader::DecodeToFirstAudioData) is #15, probably in new feature.

Aurora

• bug 856670 (js::RemapWrapper) - #2 on 22 Aurora, billm and bsmedberg are discussing in the bug, dolske could apparently hit that crash.
• bug 858648 (js::UnwrapObject) - this crash is #4 and has stopped on Nightly recently, if we find what fixed it, we should port that to 22.

Trunk

• Baseline compiler introduced the EnterBaseline signature that is basically equivalent to the EnterMethodJIT we had on JM. Due to some incompatibilities with some older processors (non-SSE2 x86 or some ARMv6), this is #1 on desktop and mobile atm. See dependencies of the meta bug 858032.
• bug 858655 (js::ScopeIter::settle) is a new crash in Nightly 23.

Mobile

• 20.0 release is under 2 crashes per 100 ADU and around 25% startup crashes, both of which is lower than any release of Firefox for Android we had before.

Addendum:

• bug 858022 - BaselineCompiler: Crashes on hardware without SSE2 support
  • Nightly shows : bug 858032 - (SadBaseline) [meta] crashes in EnterBaseline
    • as a major crash.
• NS_Debug Break affects aurora/nightly:
  • bug 834243 - crash in mozilla::layers::LayerManagerOGL::Initialize with abort message: "We need a context on Android"
  • bug 846319 - crash in LayerManagerOGL::CreateFBOWithTexture with abort message: "Framebuffer not complete -- error 0x0, mFBOTextureTarget 0xde1, aRect.width 0, aRect.height 0"

Note:

• bug 752828 - android.database.sqlite.SQLiteDatabaseLockedException: database is locked at android.database.sqlite.SQLiteStatement.native_executeSql(Native Method) on ICS
  • mfinkle made a change on friday to help :
    • https://bugzilla.mozilla.org/show_bug.cgi?id=843005
    • https://bugzilla.mozilla.org/show_bug.cgi?id=856739
  • Concern is with possible crashing due to double accessing the record as write, or corrupting the database.
• bug 845734 - crash in mozilla::MediaPluginHost::CreateDecoder @ libcutils.so@0x50.. on Sony Xperia S, T, V and Z running JB
  • Release shows this bug as the highest; doublec has the device now.

B2G

• Commercial RIL crashes : bug 859915 - crash in [@ libqc_b2g_ril.so@0x41ba4]
• libui.so@0x105fa is a Nexus S crash

Notes