CrashKill/2013-04-15

General

Socorro

• Deployment of new processor on Thu Apr 11 - bug 792484
  • Seems to be twice as fast as old processor, see http://www.twobraids.com/2013/04/new-socorro-processor-out-performs-old.html
  • Caused some regressions:
    • bug 861500 - Symbols for some libraries not used (mainly Flash and FennecAndroid)
    • bug 861524 - some recent skiplist additions missing
  • Will probably reprocess all crashes since deployment when those bugs are fixed

Flash

• Looks like Flash crashes and hangs are down by roughly a third in 21 beta due to bsmedberg's crash.

Desktop

Release

• No new crashes or surprising developments.
• bug 851934 (Outlook.com) - no reply from dvander yet to an April 10 report by the user we are working with.
• bug 836263 (Crossrider) - Even if they are working on an alternative to storing huge prefs, we should not let such a thing crash us - we could e.g. block huge prefs being set.
• bug 812683 and bug 819028 (D3D11) are down somewhat, not sure why, but also not down extremely significant.

Beta

• bug 837370 (js::types::TypeCompartment::addPendingRecompile) - waiting for approval request and uplift.
• bug 852915 (mozilla::MediaDecoderReader::DecodeToFirstAudioData) - still waiting for investigation and patch.

Aurora

• bug 858648 (js::UnwrapObject) is practically fixed.
• bug 856670 (js::RemapWrapper) seems to be practically fixed on trunk, need to find the fix to uplift to Aurora.

Trunk

• bug 859718 (mozilla::image::RasterImage::IsDecodeFinished) - patch seems to be in the bug
• Graph of nightly update curves and ADU variability: http://people.mozilla.com/~bsmedberg/bsmedberg-graphing-playground/nightly-adu.html

Mobile

• top crashing bug in nightly :
  • bug 861769 which is a dup of bug 861796 has been fixed as of today and has a patch in bound.
  • bug 860466 has been fixed and is in moz-central; there are other libmozglue.so bugs that may be related/duplicate
    • bug 861777 and several other libxul.so crashes may be related/duplicate to this bug.
  • Meta bug bug 858032 is still needs some more work.
• Aurora :
  • bug 834243 - crash in mozilla::layers::LayerManagerOGL::Initialize with abort message: "We need a context on Android"
    • looks like we need to keep track of bug 852467 - nsDisableOldMaxSmartSizePrefEvent runs on the gecko main thread, blocks for long periods of time
  • bug 752828 - android.database.sqlite.SQLiteDatabaseLockedException: database is locked at android.database.sqlite.SQLiteStatement.native_executeSql(Native Method) on ICS
    • still occurs
    • Is bug 790922 related?
• Release
  • bug 845734 - Dev can't find issue. looking for other ways to fix the problem
  • bug 824118 <- not sure what's going on with the blocklist.

B2G

• ril crashes : bug 851812
• bluetooth crash :
  • bug 857896 - [Leo] [Bluetooth] Convert device property "Connected" into a boolean if bluez returns a two-byte array
  • bug 860075 - [Inari] crash in mozilla::dom::bluetooth::BluetoothService::DistributeSignal
• libstagefright crashes are due to nexus S and keon w/ video app.
• Otoro is showing : [@ android::sp<android::GraphicBuffer>::operator= ] at various websites:
  • URL : http://bit.ly/css3d-mol , http://mrdoob.github.io/three.js/examples/css3d_molecules.html , http://www.mrdoob.com/lab/javascript/threejs/css3d/periodictable/

Notes