Thursday 18th August 2011

This month's meetings

<< <  2011 / 8 >  >>
Mon Tue Wed Thu Fri Sat Sun
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        



  • Stand up call



  • Working on EGLIBC test harness
  • Handling feedback from Joseph Meyers on string routines
  • Testing a cross build, having trouble with native
  • Tried Android build on QEMU and on real BeagleBoard, fails in same way

  • Tested on a real Beagle at all?
  • Only overlap with QEMU unfortunately
  • Having trouble with building SVN, often broken
  • ACTION: Michael to send Dave points to the cbuild stuff
  • Ramana saw problems last night with dosyscall


  • Back from vacation, getting back into working mood
  • Cross-compiling EEMBC to run on the PandaBoard, helped by Ramana

  • Want to try the different ways like native vs cross, remotely
  • Three errors in automotive
  • ACTION: Michael to check his results and corralate
  • Started a wiki page to document steps for future newcomers
  • Working with Snowball in parallel
  • ACTION: Michael to send binary toolchain specs and links


  • 2011.08 release
  • CSL data centre is changing over which is taking up some time
  • Thumb-2 constants issue is ongoing


  • libunwind support for remote unwinding is going upstream
  • Changes have been pulled!
  • Fixed up some test failures
  • Android guys saw a problem due to their earlier version of binutils
  • They're updating


  • GDB release, doing the full process himself, seems to have worked
  • Announcement hasn't come through, will ping Loic if not seen soon
  • Kicked off test
  • ACTION: Michael to look and see what happened to the x86_64 build
  • Started looking into remote libraries hook system


  • Looking at libav for SMS
  • Most of the benchmarks have loads and stores unless no aliasing
  • Will see what it's like with Richard's pass
  • Some loops have a single accumulator which SMS should improve


  • Looked at a library function being pulled in when not needed
  • Problem in libgcc with symbols that should have been hidden
  • Looked at libav
  • PowerPC constant vector fix
  • Delving into the murky depths of rtx_costs()


  • Change in default vector size has been approved
  • Working on widening shifts


  • Fix a bug upstream, been committed
  • Switch-case ranges fix is committed. Will backport.
  • Looking at issues exposed via Ira's patch. Weird things in data flow.
  • Can hack around, but hides the problem
  • More trouble with PandaBoard


  • Russled up a Kanban view onto our blueprints
  • Linting blueprints and de-assigning
  • Cloud changes with changing SMTP to be via auth
  • 4.6 release branch merge
  • Harness up libav
    • Build a natty kernel for perf
  • Talked with Doug re: next steps with STM
  • Did the 2011.08 release
  • Set up a bzr seed for the Android guys

WorkingGroups/ToolChain/Meetings/Archive/2011-08-18 (last modified 2013-08-30 11:48:21)