Permalink
Commits on Aug 25, 2012
  1. Bump version to 0.2.0.5

    bos committed Aug 25, 2012
  2. Kill off more upper bounds

    bos committed Aug 25, 2012
Commits on Dec 21, 2011
  1. Bump version to 0.2.0.4

    bos committed Dec 21, 2011
Commits on Dec 17, 2011
  1. Relax dependency on integer-gmp to support GHC 7.4

    hvr committed Dec 17, 2011
    GHC 7.4 ships with integer-gmp-0.4.0.0
Commits on Nov 18, 2011
  1. Quieten a compiler warning

    bos committed Nov 18, 2011
Commits on Nov 17, 2011
  1. Merge

    bos committed Nov 17, 2011
  2. Bump version to 0.2.0.3

    bos committed Nov 17, 2011
  3. Update more email addresses

    bos committed Nov 17, 2011
Commits on Nov 14, 2011
Commits on Nov 2, 2011
  1. Bump version to 0.2.0.2

    bos committed Nov 2, 2011
  2. Merge upstream changes

    bos committed Nov 2, 2011
    --HG--
    rename : src/fast-dtoa.cc => double-conversion/src/fast-dtoa.cc
Commits on Oct 29, 2011
  1. Merge in "Improve efficiency of BiggestPowerOfTen".

    floitsch committed Oct 29, 2011
    Issue 10/11.
  2. Merge pull request #4 from tibbe/ffi-fix

    bos committed Oct 29, 2011
    Fix build with GHC 7.3+
  3. Fix build with GHC 7.3+

    tibbe committed Oct 29, 2011
Commits on Oct 18, 2011
  1. merge with default.

    floitschG committed Oct 18, 2011
    --HG--
    branch : staging
  2. Improve efficiency of BiggestPowerOfTen.

    floitschG committed Oct 18, 2011
    This CL is based on a patch provided by Jeff Muizelaar (issue 10).
    
    --HG--
    branch : staging
Commits on Aug 10, 2011
Commits on Aug 2, 2011
  1. Bump version

    bos committed Aug 2, 2011
  2. Widen dependency on integer-gmp

    bos committed Aug 2, 2011
Commits on Jun 29, 2011
  1. Fix dmeo build

    bos committed Jun 29, 2011
  2. Update docs.

    bos committed Jun 29, 2011
  3. Add a ByteString module, and benchmarks.

    bos committed Jun 29, 2011
    Interestingly, the ByteString code is *slower* than Text, by a
    wide margin (2x or so). I assume this is due to the overhead of
    using malloc for ByteString, versus pointer-bump allocation for
    the MutableByteArray# used by Text.
    
    Still a big win over plain old show, and over bytestring-show (which
    is once again a surprise, being *slower* than plain old show!).
    
    --HG--
    rename : Data/Double/Conversion.hs => Data/Double/Conversion/ByteString.hs
    rename : Data/Double/Conversion.hs => Data/Double/Conversion/Text.hs
  4. Refactor in preparation for adding ByteString support.

    bos committed Jun 29, 2011
    --HG--
    rename : Data/Double/Conversion.hs => Data/Double/Conversion/FFI.hs
Commits on Jun 27, 2011
  1. Add files

    bos committed Jun 27, 2011
  2. Get more numbers.

    bos committed Jun 27, 2011
  3. Mark FFI entry points as unsafe.

    bos committed Jun 27, 2011
    They don't block or call back into Haskell. I really wish "safe" and
    "unsafe" were better named.