Changelog:

• Preliminary support for new cameras:
  • Canon: G15, S110, SX50
  • Fujifilm: F800EXR, XF1
  • Nikon: 1 J2, 1 V2, D600
  • Olympus: E-PL5, E-PM2
  • Panasonic: FZ200, GH3, LX7
  • Pentax: K-5 II, K-5 IIs, K-30, Q10
  • Sony: SLT-A99, NEX-5R, NEX-6
• LibRaw::subtract_black() return type is now integer. Function returns errorcode according to libraw error conversion.
• New processing stage LIBRAW_PROGRESS_RAW2_IMAGE (this flag is set by LibRaw::raw2image() call)
• New warning flag LIBRAW_WARN_RAWSPEED_PROCESSED
  - not really a warning, but information that RawSpeed library was used for image decoding.
• imgdata.rawdata.color_image is now split to two pointers:
  - ushort (*)[3] color3_image for 3-component RAW images (color DNGs, sRAW)
  - ushort (*)[4] color4_image for 4-component color raws (Sinar 4shot, Kodak RAWs and so on).
• imgdata.idata.sizes.raw_pitch is now in *bytes* This value should be used for access rows in imgdata.rawdata.*_image buffers.
• Enabled RawSpeed processing for non-bayer DNGs and sRAW.
  Note: there is a bug in RawStudio r499, 3-color (linear) uncompressed DNGs are read incorrectly. Filed to RawStudio bugzilla Use RawSpeed/rawspeed.uncompressed-color-dng.patch until this bug is not fixed in main RawSpeed source repository.
• Removed OpenMP support for linear interpolation because of problems on some compilers.
• Disabled wchar_t * interface if LibRaw compiled by MinGW because no wchar_t strings in MinGW STL
• ABI and API changed, all programs using 0.15-Beta1 and older should be recompiled!