Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
branch: clang-with-o2
Commits on May 23, 2012
  1. malloc

    authored
Commits on May 22, 2012
  1. channel

    authored
  2. @dbussink

    Merge pull request #1737 from ryoqun/libffi-with-clang

    dbussink authored
    Fix libffi build issue with Clang
  3. Remove an unneeded C-style cast

    authored
  4. @dbussink
  5. @dbussink

    Remove unused LLVM file

    dbussink authored
  6. @dbussink

    Add a bench for ruby executable invocation

    dbussink authored committed
Commits on May 21, 2012
  1. @ileitch
  2. @ileitch

    Exit after -Xconfig.print

    ileitch authored
  3. Fix libffi build issue with Clang

    authored
    This cherry-picks a commit by hand to fix a libffi build failure with clang.
    In particular, I applied the following patch from the official libffi
    repository:
      https://github.com/atgreen/libffi/blob/master/patches/x86_pcrel_test
    
    References:
      http://bugs.python.org/issue12812
      https://bugzilla.mozilla.org/show_bug.cgi?id=631928
      http://sourceware.org/ml/libffi-discuss/2011/msg00024.html
  4. @ileitch
Commits on May 20, 2012
  1. @ileitch
  2. @ileitch

    Fix IO#popen tags.

    ileitch authored
  3. @dbussink
  4. @dbussink

    Update to libffi 3.0.11

    dbussink authored
  5. Don't use unions for cmsghdr

    authored
    Work around build issues on mac instroduced this commmit:
      e2e7b25 Use union instead of struct for cmsghdr
    
    Reference:
      http://alas.matf.bg.ac.rs/manuals/lspe/snode=153.html
  6. Use union instead of struct for cmsghdr

    authored
    Use of struct with cmsghdr and pad is a GNU extension. When built with
    Clang, warnings are issued as such. So take the union approach instead.
    
    To fix this, I used an example at:
      http://www.rootr.net/man/man/CMSG_SPACE
    
    This is a second try to fix clang warnings in io.cpp. The reverted first try
    was:
      a9648cf Ignore all GNU-extension warnings
    
    As an alternative approach to it, I fixed the code itself.
  7. @Gibheer
  8. @Gibheer
Commits on May 19, 2012
  1. @nazgob

    Revert "fixed Rake deprecation warning"

    nazgob authored
    This reverts commit 95bb2fe.
  2. @ileitch
  3. @dbussink

    Revert "Ignore all GNU-extension warnings"

    dbussink authored
    This reverts commit a9648cf.
  4. Regenerated Rapa

    authored
  5. Ignore all GNU-extension warnings

    authored
    This commit suppresses following warnings from Clang:
      vm/builtin/io.cpp:1101:22: error: field 'hdr' with variable sized type
        'struct cmsghdr' not at the end of a struct or class is a GNU extension
        [-Werror,-Wgnu]
            struct cmsghdr hdr;
                           ^
      vm/builtin/io.cpp:1146:22: error: field 'hdr' with variable sized type
        'struct cmsghdr' not at the end of a struct or class is a GNU extension
        [-Werror,-Wgnu]
            struct cmsghdr hdr;
                           ^
      2 errors generated.
  6. @nazgob

    fixed Rake deprecation warning

    nazgob authored
  7. Define the prototype of rb_cvar_set_internal

    authored
    This suppresses a following warning by running specs with Rubinius built by
    clang:
    
      spec/ruby/optional/capi/ext/class_spec.c:108:2: warning:
        implicit declaration of function 'rb_cvar_set_internal' is invalid in C99
          [-Wimplicit-function-declaration]
            rb_cvar_set(klass, rb_intern(StringValuePtr(name)), val);
            ^
      vm/capi/19/include/ruby/ruby.h:1096:23: note: expanded from:
        #define rb_cvar_set   rb_cvar_set_internal
                              ^
      1 warning generated.
  8. @IPGlider

    Make some Zlib::ZStream methods private

    IPGlider authored
    These methods should't be public, there are more methods that shouldn't
    be public but they need more work or is not currently possible.
  9. @IPGlider

    Remove Zlib::ZStream#detach_input & #detach_output

    IPGlider authored
    These methods where called directly by #flush_next_in and #flush_next_out
    and are not in MRI so let's remove them for the sake of consistency.
Commits on May 18, 2012
  1. @ileitch
  2. @nazgob
Something went wrong with that request. Please try again.