1. 26 Apr, 2017 3 commits
  2. 24 Apr, 2017 2 commits
  3. 20 Jan, 2017 1 commit
  4. 17 Jan, 2017 1 commit
    • Brendan Jackman's avatar
      base: Don't error on event field values containing '=' · 214940f3
      Brendan Jackman authored
      If a trace field value contains a '=', we currently get a ValueError('Too many
      values to unpack'). Instead, let's only split on the first '='.
      
      In practice if a field value contains a '=' it's probably because a kernel
      developer typo'd a custom event like:
      
          /* Note missing space between "%d" and "bar" */
          trace_printk("my_broken_event: foo=%dbar=%d", foo, bar)
      
      So I did consider raising an explicit 'malformed event field' error. But this
      approach is more flexible in case someone really wanted to trace fields
      containing strings with '=' in them.
      214940f3
  5. 10 Jan, 2017 3 commits
  6. 05 Jan, 2017 2 commits
  7. 04 Jan, 2017 2 commits
  8. 06 Dec, 2016 2 commits
  9. 05 Nov, 2016 2 commits
  10. 04 Nov, 2016 2 commits
  11. 02 Nov, 2016 2 commits
  12. 25 Oct, 2016 1 commit
  13. 08 Oct, 2016 2 commits
  14. 07 Oct, 2016 1 commit
  15. 04 Oct, 2016 1 commit
  16. 28 Sep, 2016 2 commits
    • Javi Merino's avatar
      Merge pull request #224 from bjackman/memory-warning · e3e9371f
      Javi Merino authored
      base: Add warning when we're at risk of running out of memory
      e3e9371f
    • Brendan Jackman's avatar
      base: Add warning when we're at risk of running out of memory · e7f31f87
      Brendan Jackman authored
      Pandas doesn't seem to properly handle running out of memory, and
      segfaults; this gives you no useful error message.
      
      Add code to check how much memory we appear to be using during each
      iteration of the generate_parsed_data generator. This uses
      resource.getrusage, which is a wrapper around a system call and
      therefore shouldn't add too much overhead. Rudimentary benchmarking with
      time.clock showed that introducing `getrusage` to the generator did not
      measurably slow down the DataFrame constructor call.
      
      _get_free_memory_kb should return None on non-Linux systems, but that
      hasn't been tested.
      e7f31f87
  17. 13 Sep, 2016 4 commits
  18. 12 Sep, 2016 1 commit
  19. 08 Sep, 2016 1 commit
  20. 05 Sep, 2016 3 commits
    • Michele Di Giorgio's avatar
      Merge pull request #221 from JaviMerino/fix_travis_build · 55552f05
      Michele Di Giorgio authored
      travis-ci: Don't fail if there isn't an X server running
      55552f05
    • Javi Merino's avatar
      travis-ci: Don't fail if there isn't an X server running · b19c84b3
      Javi Merino authored
      Travis is failing to run the tests because it is trying to use TK for
      plotting, which is not present in the environment:
      
      ======================================================================
      ERROR: Test that plotter handles triplicates fine
      ----------------------------------------------------------------------
      
      Traceback (most recent call last):
        File "/home/travis/build/ARM-software/trappy/tests/test_duplicates.py", line 95, in test_plotter_triplicates
          per_line=3)
        File "/home/travis/build/ARM-software/trappy/trappy/__init__.py", line 28, in __init__
          raise exc
      
      ImportError: No module named _tkinter, please install the python-tk package
      
      See https://travis-ci.org/ARM-software/trappy/builds/157597148 for the
      full log.
      
      Installing the python-tk package doesn't solve the issue because you
      have no X server running.  Change the matplotlib backend to "agg" (the
      Anti-Grain Geometry C++ library) that doesn't need an X server.
      b19c84b3
    • Javi Merino's avatar
      version: release trappy 5.8.0 · e189dd94
      Javi Merino authored
      e189dd94
  21. 23 Aug, 2016 1 commit
  22. 22 Aug, 2016 1 commit