Mozilla Status Board

mh+mozilla@glandium.org

Current Status - 2015-05-19

Done:

Next:

  • More of the same.

Older Posts

Member of Projects

Team Posts

mh+mozilla@glandium.org 3 days ago

Done:

Next:

  • More of the same.

erahm@mozilla.com 4 days ago

Done:

  • Primarily logging work, removed all instances of PR_LOGGING from the gecko codebase
  • Mentored several bugs

Next:

  • More logging cleanup
  • j2me.js warm start profiling
  • More mentoring
  • Look at coverity

Coordination:

nfroyd@mozilla.com 4 days ago

Done:

  • Worked on libstdc++/libc++ builds for Android/B2G.
  • Investigated building libc++ et al. standalone via cmake, discovered that's a very poorly supported build configuration.
  • Discovered we can use the NDK's already-compiled libstdc++/libc++ binaries if necessary, rejoiced.
  • Discovered that B2G builds don't use the NDK for building, less rejoicing. (This means we have to build something--probably libc++--ourselves.)
  • Started work on moz.build files for libc++ et al.
  • TemporaryRef-related rebasing and testing, once Ms2ger reviewed the patch in bug 1116905 instead of Waldo. Discovered the patches make Valgrind builds (and Valgrind builds only) error about uninitialized variables.
  • Reviews.
  • MemShrink meeting.
  • 1:1 with Nick.

Next:

  • njn (PLDHashTable patches) and poiru (nsTArray fallibility changes) have shoved a lot of reviews in my queue over the weekend.
  • Try to get TemporaryRef things landed.
  • ekr indicated he's willing to review reasonable patches in bug 1162026, so going to try to do that.
  • Want to write some more Rust code.

Coordination:

  • Temperamental compilers, bug 116905.
  • mozilla::Tuple is blocking work in bug 116905, at least until I can figure out how nsRunnableMethodArgs can be made to work there. botond says he's going to land mozilla::Tuple soon, though.

nnethercote@mozilla.com 5 days ago

Done:

Gecko landings:

  • Bug 1161377 - Backed four patches out due to weird Android startup perf regressions, and a plugin problems (sigh). Started re-landing in smaller chunks in bug 1165206.
  • Bug 1163909, bug 1163916, bug 1164373 - Removed five static constructors, which helps start-up time, esp. on Android.

Other

Next:

  • Continue PLDHashTable clean-ups.
  • Continue Servo memory reporting.

jseward@acm.org last week

Done:

First report in way too long:

  • Bug 1141712: Make LUL work with inplace ticking (not using the unwinder thread): multiple iterations, land. Result: major simplification of CFI based unwinding without much of a loss of unwinding performance.
  • Bug 1157194: Make LUL able to deal with Dwarf expressions in CFI unwind info: analyse unwind failures, implement Dwarf expression support. Works, with test cases currently under development.
  • A couple more V/Mochitest runs; not much by way of bugs to see there though.
  • V: bug 345125 (major overhaul of register allocator): finish off and land
  • V: NCode (inlining of Memcheck instrumentation): get initial version working for arm32. Disappointing results though -- slowdown due to increased icache misses. Unsure how to proceed.
  • V: quite a lot of review/merging: TileGX port (merge), Solaris port (review), various Helgrind speedups (discuss, review)

Next:

  • Continue with test case work for 1157194 (LUL Dwarf expressions).
  • Try to figure out if anything useful can be pulled out of the V/NCode work.