(gdt's judgement is that anyone paying attention to color management
will surely have jpeg and tiff libraries installed anyway, so this
isn't bloat. If it is thought to be bloat, a better fix is to
PKG_OPTIONS other programs to refrain from using lcms.)
* linux & windows packages are now joined into a single package.
* Improved (and now fast!) gamut checking.
* New virtual NULL profile for building gamut check transforms.
* CGATS.13/IT8 parser now supports multiple tables and other goodies like
specifying flot format.
* More python support.
* New Matlab wrapper
* Revamped icctrans
* Improved jpegicc
* jpegicc and tifficc now can save (grab) embedded profiles.
* tifficc now supports up to 15 channels.
* Improved Lab prelinearization.
* Fixed absolute intent on PostScript CRD
* Man pages for utilities.
* Access to creation & calibration date time.
* Lots of bug fixes
Changes:
Overall speedup
Full ICC v4 support
New CGATS.13/IT8 parser
8-bit transforms optimization
Improved named color -- lcms can now write such profiles as well
lcms can now write 8 bit profiles as well as 16 bit
softproof now honors proofing intent
Absolute colorimetric works fine on mixed v2/v4 transforms
Improved black point compensation
Several new utility functions
Black point compensation on PostScript CRD
Added support for HiFi (> 4 inks) on tifficc
Improved icclink
Add lcms.pc pkg-config file
(http://www.freedesktop.org/Software/pkgconfig).
USE_TETRAHEDRAL and USE_TRILINEAR no longer used -- engine now
chooses between interpolation methods
on depending on several heuristics.
All library names listed by *.la files no longer need to be listed
in the PLIST, e.g., instead of:
lib/libfoo.a
lib/libfoo.la
lib/libfoo.so
lib/libfoo.so.0
lib/libfoo.so.0.1
one simply needs:
lib/libfoo.la
and bsd.pkg.mk will automatically ensure that the additional library
names are listed in the installed package +CONTENTS file.
Also make LIBTOOLIZE_PLIST default to "yes".
Package changes:
The include files moved from $PREFIX/include/lcms to
$PREFIX/include. Support added to buildlink*.mk to provide
compatibility symlinks include/lcms/*.h in BUILDLINK_DIR.
No list of changes since the previously packaged version (1.06), I do not
even know if there was any versions in between... Anyway changes
in version 1.12 are (from the homepage):
- Brightness/Contrast/Hue/Saturation/WhitePoint modification across abstract
profile
- License changed to MIT
- pseq tag handling: cmsReadProfileSequenceDescription and cmsSEQ,
cmsPSEQDESC structures
- CRD generation now supports black point compensation, see
cmsGetPostScriptCRDEx
- cmsTakeManufacturer and cmsTakeModel for uncooked info on these tags
- Writing 8 bit profiles is now supported
- Named color profiles support. This turns lcms from a "wide subset" into
a "full implementation" of ICC 3.4, with some ICC 4.0 support.
- PostScript CSA, CRD generation
- Ink-Limiting capabilities for CMYK
- Devicelink profile generation.
- Gray scale virtual profiles
- Linearization virtual device link profiles
- New ICCLINK and ICC2PS utilities
- SWIG wrapper. This enables lcms from Python.
- Floating-point formats are now accepted as well.
- More ICC 4.0 compatibility. Some 4.0 profiles are now are fully understood
(still experimental)
- Profiles can now be saved to memory (thanks to Steven Greaves for providing
the code)
- Char Target data are now handled. Some profiles does store the data
profiler has used. This is all information needed to rebuild the profile
from scratch.
- New low-resolution flag cmsFLAGS_LOWRESPRECALC to save memory.
- User-defined encodings are now supported.
- cmsChangeBuffersFormat() to change the encoding of buffers on runtime
allows reuse of existing transforms.
- Gamma estimation routines cmsEstimateGamma() and cmsEstimateGammaEx()
- multilocalized unicode is now supported. Language and codepage is
selected via cmsSetLanguage() (ICC 4.0 only)
- LUT handling has been enhanced with enumerators. (SAMPLER_INSPECT)
- Improved TIFFICC, JPEGICC and ICCTRANS utilities.
- cmsOpenProfileFromMem() no longer creates temporary files.
- Transforms does accept now a maximum of 8 channels on input and 16 on
output. (last version did accept 6 on input)
- 8 <-> 16 bits per sample are now always computed accurately.
- Some minor bugs fixed
