CrashKill/2013-07-08
From MozillaWiki
Contents
General
Socorro
- Didn't ship a release last week since Wednesday (the usual ship date) was effectively Friday.
- 52 is a big release, will ship this week instead.
- HBase upgrade last week went well
- API should be staged this week
Flash
Desktop
Beta
- A lot of the explosiveness report is once again colored by GC signatures shifting around, probably without overall increase or decline.
- bug 860641 (RelevantKnowledge) might be slowly decreasing, but the in-product blocklist will only ship with 23.0b4 (and the majority of those users seem to be stuck with b1, probably not getting any update to either builds or blocklists because this is a startup crash.
Aurora
- bug 885668 is unowned and around the #10 position in both 24 and 25, should be looked at.
Trunk
- bug 885668 is a new bug rising on trunk.
- trying to figure out where [@ CRpcChannelBuffer::SendReceive(tagRPCOLEMESSAGE*, unsigned long*) ] belongs
Mobile
- bug 889433 (libcutils/libstagfright, Sony Xperia Z, JB 4.2) has been rising over the weekend on 22 release. Probably needs a stagefright block.
- bug 871290 (js::ion::Assembler::TraceJumpRelocations) is #2 and the main concern on 23 beta.
- bug 890272 (SkScalerContext::MakeRec), bug 890515 (org.mozilla.gecko.gfx.BitmapUtils.decodeByteArray), bug 890590 (_ZN17AutoDecoderCancel13RequestCancelEP8_jobject) are new topcrashes that have popped up over the weekend on trunk, responsible for bad crash rate there.
- bug 845867 and bug 8863313 don't seem much better in b2, where bug 887097 should be included already.
STR Wanted
- bug 881018 - crash in mozilla::layers::floor_div
- bug 839854 - java.lang.NullPointerException: at org.mozilla.gecko.AboutHomeContent/TopSitesView$<n>.onActivityResult
- bug 782223 - crash in NvGl214081946277 on Asus Nexus 7 (NVIDIA Tegra 3) running JB
- bug 882697 - Android crash in ExecuteRegExpImpl
- bug 848810 - crash in base::CreatePlatformFile on ICS and above
- bug 889433 - crash in mozilla::ReentrantMonitorAutoEnter::~ReentrantMonitorAutoEnter @ libcutils.so@0x4f54 on Sony Xperia Z running JB 4.2
- bug 808378 - crash in nsMediaPluginHost::CreateDecoder @ libOMX_Core.so@0x1... on devices with rk<n>board HW and Mali-400 MP GPU running JB
- bug 849589 - Intermittent robocop PROCESS-CRASH | java-exception | java.lang.RuntimeException: Only ByteBufferInputStream is supported at org.mozilla.gecko.mozglue.NativeZip.<init>(NativeZip.java:32)
Need Device
- bug 752828 - android.database.sqlite.SQLiteDatabaseLockedException: database is locked at android.database.sqlite.SQLiteStatement.native_executeSql(Native Method) on ICS
B2G
- Launch devices status:
- Symbol pushing accounts for ZTE and TCL exist, need to make Socorro use them - bug 889978
- We're starting to get info about which build IDs to expect in the wild, channel names are very probably "nightly", so not too easy to distinguish with current tools.
- We haven't seen a single crash report from TCL devices yet, apparently (but no launch with them was done yet). Their build IDs should be 20130606070202 for Poland and 20130602070207 for the South America launch.
- The crash reports we're seeing from ZTE Open seem to mostly be from a different build ID than what we are told has been released to the public in Spain. Apparently, a "20130621xxxxxx" Build ID ("xxxxxx" probably being "152332") should have been shipped, the crash reports we see are for 20130531232151.
- Nobody did get around yet to do explicit custom reports on them yet, Socorro will support appropriate searches in the future, but not yet.
- KaiRo's daily crash lists - e.g. https://crash-analysis.mozilla.com/rkaiser/2013-07-07/2013-07-07.b2g.crashes.html - now list the ZTE Open as "roamer2".
Top crashes
- bug 881723 - crash in mozilla::dom::PContentChild::Write with abort message: "actor has been |delete|d"
- malloc_mutex_unlock%7CSFTK_SlotReInit concerns me as it's a startup crash (probably from the same user), we don't have the symbols so we can't do anything about it.
Notes
- Symbols are showing unexpectedly for certain partners.