Mozilla Status Board

mh+mozilla@glandium.org

Current Status - 2015-07-28

Done:

Next:

  • More Gtk*3 fallout
  • Build system stuff
  • More jemalloc3 testing.

Older Posts

Member of Projects

Team Posts

mh+mozilla@glandium.org yesterday

Done:

Next:

  • More Gtk*3 fallout
  • Build system stuff
  • More jemalloc3 testing.

erahm@mozilla.com 2 days ago

Done:

  • Coverity triage
  • Memshrink triage
  • Bug 1173091- working w/ ateam to attempt to chunk xpcshell tests by runtime (debug linux tests were pushing 100 minutes)
  • WoW - Down to 78,600 warnings
    • Bug 1179048 - Carried as far as a could, found new owner in devtools
    • Bug 1179069 - Removed 8,000 warnings
    • Bug 1179568 - NS_WARNING: Re-registering a CID? Looked into warnings, tracked down to a new network service registering for the dbus services CID. Pending feedback from Daniel Stenberg
    • Bug 1180036 - Removed 7,500 warnings. Tested whitelisting, landed.
    • Bug 1180048 - ~5,000 instances of ' WARNING: zero axis length'. Poked, pending feedback from Robert Longson
    • Bug 1183786 - 4,500 instances of "Shouldn't call SchedulePaint in a detached pres context". Poked, pending feedback from :tn
    • Bug 1183828 - Removed 2,700 warnings. Detailed analysis for roc, got patch reviewed, landed.
    • Bug 1183845 - 5,200 instances of "NS_ENSURE_TRUE(domDoc && target) failed". Proposed removal, detailed analysis for bz, pending feedback from Neil Deakin
    • Bug 1183873 - 5,000 instances of "'!mContentCache.CacheEditorRect(this, &aIMENotification)'". masayuki says this is an e10s IME bug, flagged as tracking e10s
    • Bug 1184249 - Removed 1,900 warnings
    • Bug 1184260 - Removed 1,600 warnings
    • Bug 1184275 - Removed 4,300 warnings
    • Bug 1184285 - Removed 1,200 warnings. Detailed analysis for dhylands.
    • Bug 1184293 - Removed 2,200 warnings
    • Bug 1184295 - 1,100 instances of "have unconstrained inline-size ..." Analysis, got redirect to the right people, heycam landed fix
    • Bug 1184661 - 900 instances of "Silently denied access to property..." Feedback from past, need to look into whether or not it fixes the issue
    • Bug 1185793 - 5,800 instances of "No inner window available!" Filed, grumbled about backing out bug 1182316, khuey landed fix
    • Bug 1186113 - Disable mozSettings debugging by default. Filed for egregious emulator log spam, got workaround landed in bug 1119727

Next:

More of the same, probably coverity focused, look at startup spam.

Coordination:

Feedback from:

  • Daniel Stenberg
  • Robert Longson
  • Timothy Nikkel
  • Neil Deakin

nfroyd@mozilla.com 2 days ago

Done:

  • Reviews.
  • Got use counter patches up, seems bz was happy with them.
  • Worked on libc++ patches for Fennec.
  • Too much futzing with new desktop machine.
  • Talked with dmajor about clang-cl.
  • Struggled with minor cold all week.
  • 1:1 with Nick.

Next:

  • Get Fennec libc++ patches sorted and landed.
  • Need to figure out a plan for B2G libc++; doesn't look like fun.
  • Need to run some TSan tests and get back to filing bugs.
  • Use counter patches with smaug.

Coordination:

  • Waiting for new laptop to poke at clang/ASan stuff on Windows. Possible that dmajor will have all the warts figured out first, though!

nnethercote@mozilla.com 3 days ago

Done:

Gecko landings:

Servo landings:

Other:

  • Filed ~35 bugs about replacing the ~350 calls to nsBaseHashtable::EnumerateRead() and nsBaseHashtable::Enumerate() with Iterator uses.
  • Got working a user-space test program on Linux that reads energy measurements from machine-specific registers.
  • Discovered that the Gecko Profiler already has some integration with Intel Power Gadget on Windows.

Next:

  • More power profiling experimentation.

jseward@acm.org last week

Done:

V/Mochi runs:

  • Landed bug 1180195 (Uninitialised value use in Predictor::SpaceCleaner::OnMetaDataElement)
  • Landed bug 1137169 (Uninitialised value uses related to mozilla::dom::WebAudioUtils::SpeexResamplerProcess)
  • Investigated bug 1100338 (Uninitialised value use in nsUrlClassifierLookupCallback::Completion) Clearly something ungood here, but nobody knows how to fix it.

B2G on V on Z3c (Aries-KK):

  • Received Z3c, got that set up, built, about 15 V runs, liaise with Gregor and A Lissy about various bugs
  • Filed bug 1184066 (redacted) potential Z3c crasher
  • Filed bug 1184616 (Uninitialised value use in android::OMXCodec::read)
  • Filed bug 1184624 (Uninitialised value use in overlay::Overlay::configDone)
  • Started to read about MTBF testing for B2G.

Tried out Fx on Dr Memory on Win7: * Filed bug 1183093 (Uninitialised value use in Probe::Trigger)

V/mach improved automation: * bug 1185244 (Improve mach support for running mochitests on Valgrind): Lots of hackery, have a partial initial patch

Valgrind hacking

  • Bug triage: waded through 4 months of bug reports
  • 346411 (MIPS: SysRes::_valEx handling is incorrect): fixed
  • 345248 (add support for Solaris OS in valgrind): landed
  • Look at completing AVX2 support
  • More assembly helpers: LOADV64le for ARM32, commit 349879 (x86 (32 bit) versions)

Next:

  • More automation stuff -- try to get to first complete patch for mach.
  • V releng -- fix up AVX2 state save/restore insns
  • Build B2G for Aries-L (KK is outdated); run on V, look for bugs
  • Look into MTBF testing for Aries-{L,KK} (not sure which)