changelog.txt

Version 1.3.5 - 11 Nov 2022
 - Added feature "-opt jpeg:rstm" / "-opt jpeg:rstr".
 - Added feature "-opt jpeg:optcoding".
 - Added feature "-opt bmp:version=auto".

Version 1.3.4 - 9 Sep 2021
 - (This release is mainly to refresh the Windows binaries.)
 - Minor improvements to code quality and stability.
 - Changed libtool versioning scheme.

Version 1.3.3 - 19 Oct 2018
 - Security and stability fixes

Version 1.3.2 - 25 May 2017
 - Security and stability fixes

Version 1.3.1 - 24 Apr 2017
 - Security and stability fixes

Version 1.3.0 - 4 Nov 2015
 - Partial support for OS/2 Bitmap Array BMP files.
 - Added -negate option.
 - Added -opt option, for format-specific options.
 - Discontinued 64-bit floating point precision support.

Version 1.2.0 - 16 Feb 2014
 - Switched to MIT-style license.
 - Support for reading and writing Netpbm PAM format.
 - Partial support for reading files from stdin.
 - Allow a filename of "-" to mean "stdin:" or "stdout:".
 - Partial support for writing files to stdout.
 - Print text output to stderr instead of stdout.
 - Added -depthcc option.
 - Support for reading binary PPM, PGM, and PBM (Netpbm) formats.
 - Fixed -condgrayscale to work when reading MIFF format.
 - Support for writing the PPM, PGM, and PBM (Netpbm) formats.
 - Added "rec709" as a colorspace option.
 - Added -colortype option.
 - Removed support for non-Unicode Windows builds.

Version 1.1.0 - 2 Feb 2013
 - Support for background colors with transparency.
 - Support for background colors in MIFF files.
 - Added -intent option, to set the rendering intent without setting the
   colorspace to sRGB.
 - By default, copy the "rendering intent" setting from the source to the
   target image.
 - Support for writing and maintaining background color metadata. Added
   -bkgdlabel option.
 - Fixed some issues with clamping interacting badly with transparency.
   Monochrome images with transparency could have other colors introduced
   unnecessarily.
 - Added iw_set_output_image_size() function and -imagesize option.

Version 1.0.0 - 4 Nov 2012
 - If -blur is used without -filter, default to "-filter gaussian".
 - Added -s and -S options.
 - Removed the --with-sample-depth configuration option.
 - Removed CMake support.
 - Renamed "boxmv" filter to "boxavg".
 - Changed how the -encoding option works. On Windows, write UTF-8 text by
   default if not writing to a console. Removed the "utf8raw" and "utf16raw"
   options.

Version 0.9.10 - 2 Jun 2012
 - Support reading and writing BMP files with more than 8 bits per sample.
 - Support writing v5 BMP files (use -bmpversion 5).
 - Support for reading some v4 and v5 BMP files.
 - Support for writing OS/2-style BMP files.
 - Added -reorient option.
 - Support for reading Orientation tags from JPEG/Exif files.
 - Ability to rotate and mirror images.
 - Added "-precision 32" option.
 - Added an alternate box filter, "boxmv".
 - Better support for -grayscale when writing BMP images.
 - Support "-cc <r>,<g>,<b>" syntax.
 - [Windows only] Support for reading images from, and writing images to, the
   clipboard.
 - Changes to allow data streams other than regular files. Filenames that may
   contain a colon now need to be prefixed with "file:".
 - Support for custom bit depths (sBIT chunks) when writing PNG images. Added
   -depthalpha and -depthgray options.
 - Support for custom bit depths when writing 16-bit BMP images.

Version 0.9.9 - 29 Mar 2012
 - Various changes to how non-square pixels are handled. Added -noresize
   option.
 - Support for writing 16-bit BMP images.
 - Support for reading 16- and 32-bit BMP images.
 - Support for RLE-compressed BMP images with transparency.
 - Support for reading BMP files.
 - Support for writing BMP files with RLE4 compression.
 - Support for writing BMP files with RLE8 compression.

Version 0.9.8 - 18 Feb 2012
 - Support for building using GNU autotools.
 - Support for two types of custom grayscale formulas.
 - Added -gsf as an alias of -grayscaleformula, and made it unnecessary to
   specify "-grayscale" if you set the grayscale formula.
 - Support reading CMYK/YCCK JPEG images.

Version 0.9.7 - 30 Dec 2011
 - Support for custom memory allocation functions.
 - Renamed -pngcmprlevel option to -zipcmprlevel, and made it apply whenever
   zip/deflate compression is used.
 - Added Zip compression support for MIFF images, and made it the default.
 - Added "-compress" option.
 - Added support for transparent virtual pixels ("-edge t").
 - Added "-translate s" option, so you can specify the translation amount in
   source pixels instead of target pixels.
 - Enabled transparency support for WebP images.
 - Made the "edge policy" setting be per-dimension.
 - Removed the -filteralpha feature.
 - Stopped disallowing the "s" edge policy when -translate or -offset is used.
 - Changed the default downscaling filter to catrom.
 - Changed the box filter to have more predictable behavior.

Version 0.9.6 - 1 Dec 2011
 - Added -translate option, to fine-tune the position of the image.
 - Added "-blur x" option, to make it easy to do simple interpolation.
 - Support -blur when using the "mix" filter.
 - Added a way to request a specific density label.
 - Added an option to write JPEG files using arithmetic coding.
 - Don't clamp alpha samples before using them to apply a background color.
 - Added -noincludescreen option, for more control over how GIF images are
   handled.
 - Added -page option, to read arbitrary images from GIF files.
 - Changed the default bit depth to always be 8, instead of having it depend
   on the depth of the input file.

Version 0.9.5 - 8 Oct 2011
 - Use a binary search to find the nearest available color. Improves speed
   significantly in many cases.
 - Random dithering now uses a private pseudorandom number generator.
 - Changed the default PNG compression level to 9.
 - Updated the format of grayscale MIFF files to be compatible with the
   current version of ImageMagick.
 - Added -encoding option, to set the character encoding for text output.
 - Allow -cs to set the output colorspace even when IW doesn't think it's a
   good idea.
 - No longer attempting to make the output colorspace be the same as the
   input colorspace.

Version 0.9.4 - 9 Aug 2011
 - Added -density option, to give some control over how the density of the
   target image is calculated.
 - Added -nowarn and -noinfo options.
 - Support for reading GIF files.

Version 0.9.3 - 30 May 2011
 - Support for rational numbers like "1/3" in the command line.
 - Basic support for Google's WebP file format.
 - Partial support for ImageMagick's MIFF file format.
 - Support for image sizes that are relative to the original image size.
 - Moved strings to string tables.

Version 0.9.2 - 7 Apr 2011
 - Made the library use only UTF-8 for character encoding.
 - Improved resampling speed.
 - Read the background color metadata from PNG files, and use it as the
   background color in certain circumstances.
 - Detect input file formats based on the contents of the file, instead of the
   file name.
 - Added "sinc" filter.
 - Support for writing uncompressed TIFF files.
 - Replaced "-nobinarytrn" option with "-noopt".
 - Added a "halftone" ordered dither.
 - Support for reduced input bitdepths; respect PNG sBIT chunks.
 - The image format modules now use abstracted I/O.
 - Support for writing BMP files.
 - Added option to set the PNG compression level.
 - Improved speed of colorspace conversion.

Version 0.9.1 - 22 Feb 2011
 - Added a crop feature.
 - Support for optimizing to an image with binary transparency.
 - Added "-interlace" option.
 - Added "-jpegsampling" option.
 - Retain pixel density information if the image isn't being resized.
 - Support for optimizing to a 4- or 16-color grayscale image.

Version 0.9.0 - 6 Feb 2011
 - Initial public release.