Mozilla Status Board

mh+mozilla@glandium.org

Current Status - 2015-04-14

Done:

Next:

  • More of the same.

Older Posts

Member of Projects

Team Posts

nfroyd@mozilla.com 5 days ago

Done:

  • Random acts of build system hacking (bug 1155816, bug 1155776, bug 1155760, bug 1155738, bug 870891)
  • Volunteered to review some editor/ patches for Ehsan...that was (and still is) interesting.
  • Discussed merits of glob specifications in build system, filed bug 1155377 for supporting globbing in EXTRA_JS_MODULES.
  • Stared at build system issues surrounding addons. No inspiration there yet. Current plan is to keep moving addon-related things into moz.build and hope that helps make things clearer, or at least move to context-based things for addons.
  • Did some do_CreateInstance profiling, wrote patch for bug 1153988.
  • Rebased cleanup patch for do_QueryObject, bug 1095098.
  • Back-and-forth on the merits of an nsAutoPtr patch, bug 1155500.
  • Reviews. Feels like I've been missing things on reviews lately, trying to figure out how to not do that.
  • Talked to bent about strategies for reviewing threaded code.
  • Reserved filghts for Whistler. Travel time to the airport is sad-making.
  • Transferred Q2 goals into Workday.

Next:

  • editor/ reviews.
  • More build system hacking.
  • More cleanup patches.
  • More TSan runs with a modern clang.
  • Probably some other things.

mh+mozilla@glandium.org last week

Done:

Next:

  • More of the same.

erahm@mozilla.com last week

Done:

  • Bug 1142693 - worked w/ jld to track down a bug in IPC code, wrote tests to evaluate theories and performed detailed analysis
  • Used DMD to track down large spikes during j2me.js app launch, looks like it's primarily ion compilation
  • Did some work on security bugs
  • Landed bug 1145056, trying to silence some coverity warnings
  • Minor AWSY cleanup
  • Booked SF trip on 4/20 for j2me.js, booked whistler trip
  • j2me.js memory benchmarking

Next:

  • More of the same

nnethercote@mozilla.com 4 weeks ago

Done:

Gecko:

  • Bug 1129786 - Unlazify initialization of some PLDHashTables. Simplifies some code, thanks to PLDHashTable now being automatically lazy.
  • Bug 1146216 - Remove MakeSlotArray(), because it's dead.

Servo:

Shumway:

j2me:

Next:

A month of vacation.

jseward@acm.org last month

Done:

  • Chased profiler regressions (LUL failing); turned out to be bug 1132586
  • Looked a bit at bug 1139386 (jemalloc crashes on EC2), failed to be useful.
  • More Fx/V mochi runs.
  • Started work on reinstating LUL (native unwind for SPS).
  • bug 1022583 (enable LUL by default on desktop linux). Refresh. Re-consider xpcshell test failures.
  • Conclude cannot fix it properly with the unwinder thread in place. Make plan to enable LUL without unwinder thread (bug 1141712). Start to implement. Initial run successful.
  • V: quite some patch reviewing and coordination: ubsan cleaning, icc as compiler, stack seg 344139, sigaltstack extension 344886.
  • V: bug 345125. major overhaul of register allocator and representation of registers. Reduces reg-alloc time by about 1/3 and gives up to 7% memcheck speedup in very JIT intensive situations. Like starting Fx on V. x86/amd64 only right now.

Next:

  • More LUL work -- tidy up and nuke all traces of the infamous Unwinder Thread.
  • V: fix up remaining backends to use new reg-alloc framework.