Permalink
Switch branches/tags
Commits on Sep 21, 2012
  1. Merge 'rurban/bytebuffer-resize-gh835'

    Reini Urban committed Sep 21, 2012
    Fixed conflict in ChangeLog
  2. Merge branch 'gh837_concat'

    Reini Urban committed Sep 21, 2012
    Fixes [GH #837], a imcc optimizer problem with SREG's and const
    not holding STRING*, thus no good encoding information.
  3. [GH #837] New comparison benchmark against stress_stringsu.pir

    Reini Urban committed Sep 20, 2012
    Bad news:
    There is almost no measurable speed advantage from ascii strings to encoded strings.
    With the fix and without the fix, as sprintf handling slows down the test 4x times.
    
    So converting a SREG and const to encoded STRING* will not gain much performance.
  4. [GH #837] Add benchmark to test encoding slowdown

    Reini Urban committed Sep 20, 2012
    stress_stringsu.pir tests encoding via sprintf and concat.
    The fixed version is only a bit slower then the old broken version.
    13.666s (broken) vs 13.506s (fixed) vs 3.022s (without encoding).
    
    So fixing the encoding representation as STRING* in SREG and consts 'U'
    is a worthwile goal.
  5. [GH #837] ChangeLog line for imcc optimizer bug

    Reini Urban committed Sep 20, 2012
  6. [GH #835] Fix set ByteBuffer from string

    Reini Urban committed Sep 21, 2012
    set_string_native() did not reset char* content, only the STRING *source.
  7. Add bytebuffer resize tests for [GH #835]

    Reini Urban committed Sep 18, 2012
    As it turns out adding the failing test to t/pmc/bytebuffer.t
    does not reveil any problem, but if called seperately as main it does.
    Add t/pmc/bytebuffer2.t to MANIFEST
  8. [cage] Remove windowism from pprof2cg.pl

    Reini Urban committed Sep 21, 2012
    use /usr/bin/perl code
Commits on Sep 20, 2012
  1. [GH #837] Honor encoding in imcc optimizer

    Reini Urban committed Sep 20, 2012
    Context registers unfortunately do not store type information,
    just the values. We would need at least type 'U" here.
    So recreate an proper const'ed encoding string from the register
    value.
  2. Revert "[t] Add some failing tests for #837 reported by pmichaud++"

    leto committed Sep 20, 2012
    This reverts commit 5d7aa2d.
Commits on Sep 19, 2012
  1. Minor doc corrections.

    ayardley committed Sep 19, 2012
Commits on Sep 18, 2012
  1. The 4.8.0 Release

    ayardley committed Sep 18, 2012
  2. ChangeLog: Add int3 trap fix on Solaris with gcc [GH #583]

    Reini Urban committed Sep 18, 2012
  3. ChangeLog: note improved VALGRIND env

    Reini Urban committed Sep 18, 2012
Commits on Sep 17, 2012
  1. Fix solaris gcc build: int3 trap not supported

    Reini Urban committed Sep 17, 2012
    Apparently the solaris 10 gcc assembler does not define the
    int3 (0xCC) opcode. Disable it for solaris, i.e defined(sun).
    gcc (GCC) 4.5.2 SunOS solaris 5.11 11.0 i86pc i386 i86pc
    
            "<stdin>", line 127437 : Illegal mnemonic
            Near line: "    int3"
            "<stdin>", line 127437 : Syntax error
            Near line: "    int3"
  2. .travis.yml: smoke more branches with less configurations

    Reini Urban committed Sep 17, 2012
    Add smoke-me branches and threads, but reduce the configurations a bit. Shuffle optimize and g++ a bit.
    1-2 * 9 vs. now 3-4 * 5
  3. .travis.yml: smoke threads branch also

    Reini Urban committed Sep 17, 2012
  4. [codingstd] t/perl/Parrot_IO.t copyright year

    Reini Urban committed Sep 17, 2012
  5. [GH #832] TODO t/op/sprintf.t 217 on mingw (%+u prints +)

    Reini Urban committed Sep 17, 2012
Commits on Sep 16, 2012
  1. Fixed dynpmc/os windows reveiled duplicate names

    Reini Urban committed Sep 16, 2012
    Fixed several bugs in the dummy dynpmc os:
    The dynpmc/os was a static pmc. msvc could not link duplicate names.
    I had to change the pmc to dynpmc, and had to rename the os to osdummy,
    which is installed as os.so/bundle/dll. Tested ok on msvc.
  2. dynpmc/os: nmake fix, $< invalid for os.str

    Reini Urban committed Sep 16, 2012
    $< is only valid for suffix rules, and also then not for nmake
  3. os.pmc actually needs os.str

    Benabik committed Sep 16, 2012
    I find this somewhat bizzare, given that it's basically empty, but
    don't know how to fix that.  Instead of using the old nmake-breaking
    rule, let the suffix rule in the main Makefile handle it.
  4. dynpmc/Rules: Revert OS build rules to pre 97aa16a

    Benabik committed Sep 16, 2012
    The rule to build os.str seems to be unnecessary and breaks nmake.
Commits on Sep 15, 2012
  1. Enable VALGRIND tests with more test functions

    Reini Urban committed Sep 15, 2012
    Now also pir. E.g. VALGRIND=tsan prove `find t/src -name \*.t`
  2. Enable VALGRIND tests with c_output test functions

    Reini Urban committed Sep 15, 2012
    E.g. POSTMORTEM=1 VALGRIND=tsan perl t/src/embed.t
    or VALGRIND=valgrind perl t/src/extend.t
  3. fix bogus t/src/checkdepend.t warnings

    Reini Urban committed Sep 15, 2012
    PARROT_HAS_(CORE|EXTRA)_NCI_THUNKS is in %PConfig, not %PConfig_Temp.
    
    Skip warning for src/gc/malloc.c /usr/include/malloc.h when
    HAVE_USR_INCLUDE_MALLOC_H is undefined.
    
    Skip warning for src/glut_nci_thunks.c glut_nci_thunks.str at all.
  4. [cage] Makefile cover-clean. Make less noise and use echo

    Reini Urban committed Sep 15, 2012
    make -s clean listed all dirs to be cleaned up for gcov-generated files.
    Print only one line
  5. skip outdated extend tests with old embed API [GH #829]

    Reini Urban committed Sep 15, 2012
    I also tried to convert extend_15 to the new API but failed to understand Parrot_runcode()
Commits on Sep 13, 2012
  1. [CAGE] remove duplicate os.bundle in MANIFEST.generated

    Reini Urban committed Sep 13, 2012
  2. dynpmc/os: and now actually install it

    Reini Urban committed Sep 13, 2012
    Sigh