1. 18 Aug, 2014 1 commit
  2. 16 Apr, 2013 1 commit
  3. 20 Mar, 2013 1 commit
  4. 27 Feb, 2013 1 commit
  5. 31 Dec, 2012 1 commit
    • Alexander Hesse's avatar
      Initial package rename: · 31c597f6
      Alexander Hesse authored
      pypy.rpython -> rpython.rtyper
      pypy.translator -> rpython.translator
      pypy.rlib -> rpython.rlib
      pypy.jit -> rpython.jit
      pypy.objspace.flow -> rpython.flowspace
      pypy.annotation -> rpython.annotator
      31c597f6
  6. 26 Nov, 2012 1 commit
  7. 14 Apr, 2012 1 commit
  8. 04 Mar, 2012 1 commit
  9. 27 Dec, 2011 1 commit
  10. 20 Nov, 2011 1 commit
  11. 15 Jun, 2011 1 commit
  12. 14 Jun, 2011 1 commit
  13. 09 Jun, 2011 1 commit
  14. 07 Jun, 2011 1 commit
  15. 06 Jun, 2011 1 commit
  16. 14 May, 2011 1 commit
  17. 12 May, 2011 2 commits
  18. 24 Mar, 2011 1 commit
  19. 27 Jan, 2011 1 commit
  20. 12 Dec, 2010 1 commit
  21. 03 Dec, 2010 1 commit
  22. 28 Nov, 2010 1 commit
  23. 27 Nov, 2010 1 commit
  24. 25 Nov, 2010 1 commit
  25. 24 Nov, 2010 1 commit
    • Armin Rigo's avatar
      (antocuni, arigo) · d8d19663
      Armin Rigo authored
      Merge branch/jit-free.
      
      The major point of the branch is to reduce memory usage of long-running
      pypy's with the JIT (such as "pypy translate.py").  Done by playing with
      weakrefs and having a __del__ that kills the ResumeGuardDescrs of old
      loops.  This is controlled by "--jit loop_longevity=N".  (Actually
      freeing the assembler code would be the next step, but is less important,
      because the ResumeDescrs take more memory.)
      
      In addition, this branch contains a number of semi-related things:
      
          * Generators: set their 'frame' to None when they finish.  This
            helps a lot the GC.  (The issue is that generators need an
            interp-level __del__; this let the GC free the frame earlier.)
      
          * rgc.dump_rpy_heap() now works with almost no need for
            additional memory.
      
          * pypy.rlib.debug.debug_print(r_longlong(...)).
      
          * add debug_{start,stop,print} to the __pypy__ module.  This lets
            you write to the PYPYLOG file from app-level.
      
          * use this new interface in translator/driver, to record the begin
            and end of tasks (via pypy.tool.debug_print so that it does not
            crash on CPython).
      
          * improve a bit reftracker.py.
      
      The net result is that you can now translate pypy with a pypy-jit on
      32-bits, using roughly 2.5 GB of RAM.  It is twice as big (for now) as
      CPython, but twice as fast :-)
      d8d19663
  26. 18 Nov, 2010 1 commit
  27. 09 Nov, 2010 1 commit
  28. 08 Nov, 2010 1 commit
  29. 31 Oct, 2010 1 commit
  30. 23 Oct, 2010 2 commits
  31. 23 Nov, 2010 1 commit
  32. 30 Sep, 2010 1 commit
  33. 22 Sep, 2010 1 commit
    • Antonio Cuni's avatar
      Merge the resoperation-refactoring branch. · 956fb5de
      Antonio Cuni authored
      The idea is to reduce the memory needed by ResOperations, by putting certain
      fields (e.g. a variable-sized list of arguments, or descr, or fail_args) only
      on the operations that actually need them.
      
      The benchmarks don't show any speedup or slowdown, but it's still worth
      merging because it saves a bit of memory: list of operations are thrown away
      immediately right now, but we might want to keep them in the future.
      956fb5de
  34. 25 Nov, 2009 1 commit
    • Armin Rigo's avatar
      Merge the branch shorter-guard-path: · 105d952c
      Armin Rigo authored
       * Cancel a previous refactoring that moved
         the numbering of the guard descrs to the front-end.  Instead, move
         that counting to AbstractCPU.
      
       * Generate much shorter failure paths from the x86 backend:
         generate_quick_failure() writes just a jump to control code which
         can itself do the guard recovery by reading some byte-compressed
         data.
      
       * Guards in x86 don't need to store the attribute _x86_faillocs
         any more.
      
       * Improve the testing of "machine code block full".
         Although it is still a bit random, now it runs all
         runner tests a second time with a 1KB limit.
      105d952c
  35. 24 Nov, 2009 1 commit
    • Samuele Pedroni's avatar
      merge compress-virtuals-resumedata(2) · 4f4e3c54
      Samuele Pedroni authored
      (cfbolz, pedronis)
      reuse vinfos by making virtuals and boxes out of them (and not the environment) have more stable numberings
      by introducing holes (None) in fail_args
      4f4e3c54
  36. 12 Nov, 2009 1 commit
  37. 02 Nov, 2009 1 commit
    • Armin Rigo's avatar
      Merge the 'logging2' branch. Add a general way to record logging events · 172def9e
      Armin Rigo authored
      to pypy.rlib.debug.  In addition to debug_print() there is now
      debug_start() and debug_stop() to mark the start/stop of a category.
      
          svn merge -r68854:68901 svn+ssh://codespeak.net/svn/pypy/branch/logging2
      
      Use it in the GC (replaces the gcconfig.debugprint config option) and in
      the JIT (similar to jitprof.py, which is not dead yet but obsoleting).
      
      To enable logging run the compiled program with 'PYPYLOG=:file'.  To
      enable lightweight logging only, suitable for profiling, run it with
      'PYPYLOG=file'.  See translator/c/src/debug.h for more options.
      
      To turn the profiling logs into nice graphics, see pypy/tool/logparser.py.
      172def9e
  38. 17 Oct, 2009 1 commit