Change Log

• Fixed Bugs
  • N/A
• Known Issues
  • N/A
• API breaking Changes
  • N/A
• Deprecated functionality
  • N/A
• New functionality
  • N/A
• Documentation
  • ColorSchemes.html now has images for all CB schemes
• Examples
  • New
    • phoenix: Classic phoenix fractal (escape time)
    • phoenixD: Classic phoenix fractal (distance estimator outside)
    • phoenixI: Classic phoenix fractal (distance estimator inside)
    • phoenixM: A movie of the classic phoenix fractal
    • phoenixM2: Another movie of the classic phoenix fractal
  • Updated
    • color_lut_docs: Now generates images for all CB schemes
• Miscellaneous
  • N/A

• Fixed Bugs
  • N/A
• Known Issues
  • N/A
• API breaking Changes
  • WARNING: Eventually the deprecated headers will be removed.
  • The "all" target has been removed. Use the "examples" target instead.
  • New release process with 4 component version numbers
• Deprecated functionality
  • Several people reported conflicts with other packages using the same header names
    • Deprecated header color.hpp – USE MRcolor.hpp instead
    • Deprecated header point2d.hpp – USE MRpoint2d.hpp instead
    • Deprecated header colorTpl.hpp – USE MRcolorTpl.hpp instead
• New functionality
  • Library exported via build/MRasterLib.cmake
  • Package exported via build/MRasterConfig.cmake & build/MRasterConfigVersion.cmake
  • Install target created that installs:
    • MRasterLib.cmake
    • MRasterConfig.cmake & MRasterConfigVersion.cmake
    • Non-deprecated headers
  • New include file: mraster_config.hpp
    • Defines for cmake configuration information
    • Functions returning version information Note this change limits future version numbers to numeric components.
• Documentation
  • The color scheme org-mode document has been cleaned up
  • Added change log (this very file!)
  • Updated README.org to reflect change log and remove comments about very old versions
• Examples
  • New
    • kalisetJ.cpp – Julia style kaliset
    • kalisetM1.cpp – Mandelbrot style kaliset
    • kalisetM2.cpp – Mandelbrot style kaliset
    • butterfly.cpp – A kaliset inspired butterfly
    • barrymartin.cpp – Barry Martin attractors
    • tinkerbell.cpp – Tinkerbell attractors
    • tinkerbell_search.cpp – Search for interesting Tinkerbell parameters
  • Updated
    • N/A
• Miscellaneous
  • N/A

• Examples
  • New
    • newton_roter
    • newton_3updown
    • newton_min_angle
    • mandelbrot_count_movie
    • laguerre_3updown.cpp
    • laguerre_simple.cpp
  • Updated
    • mandelbrot_simple: Now uses OpenMP
• API breaking Changes
  • The ramCanvas constructors now throw. The cases where they throw would have likely caused a runtime issue later anyhow, so I don't expect very many people to notice this one…
  • The configure.sh script must now be run from the build directory. It is also much smarter about finding a good compiler.
• New functionality
  • Numerous documentation updates & fixes
  • unitTooIntLinMap: template version now handles large integers
  • 2D color maps
    • csSet(): Added two argument version
    • cs2dIdxPalArg: Create an argument color scheme from an indexed color scheme
    • cs2dFltPalArg: Create an argument color scheme from a continuous color scheme
    • cs2dRichardson: Richardson's 2D complex number coloring scheme.
    • cs2dThaller_tpl: Meta class for Bernd Thaller's 2D complex number color methods
      • cs2dThallerHSL: Thaller method using HSL
      • cs2dThallerHSVm: Thaller method using HSV with maximal V value
      • cs2dThallerHSV: Thaller method using HSV with dynamic V value

• Examples
  • Many examples have been improved
  • New:
    • newton_max_back.cpp'
    • newton_min_root.cpp
    • mandelbrot_ltrap.cpp
    • mandelbrot_pickover.cpp
    • biomorphMorph.cpp
    • mandelbrot_precomp.cpp
    • julia.cpp
    • juliaM2.cpp
    • juliaM.cpp
    • newton_modified.cpp
  • Renamed:
    • mandelbrot_cycles.cpp -> mandelbrot_1n2cycles.cpp
  • New/updated web pages:
    • https://www.mitchr.me/SS/biomorphMorph/index.html
    • https://www.mitchr.me/SS/newton/index.html
  • Removed:
    • newton_bm_*.cpp
    • mandelbrot_distance_old.cpp
    • mandelbrot_bm_real_eqop.cpp
    • mandelbrot_1n2cycles.cpp
• Bugs
  • mjr::ramCanvasTpl::expandCanvas. Fixed a bug with monochrome images. Changed the behavior when x1 or y1 is off the canvas. Fixed a bug with non-zero x1 argument.
  • Swapped mjr::ramCanvasTpl::rotate90CW & mjr::ramCanvasTpl::rotate90CW
  • mjr::ramCanvasTpl::autoMaxHistStrech(). Fixed bug when RGB were not on default channels.
• New functionality
  • ramCanvasTpl::getCanvasWidD()
  • New canvas converter: rcConverterMonoIntensity
  • ramCanvasTpl::adjoinCanvasRight()
  • ramCanvasTpl::adjoinCanvasLeft()
  • ramCanvasTpl::adjoinCanvasBottom()
  • ramCanvasTpl::adjoinCanvasTop()
  • ramCanvasTpl::insertCanvas() – insert a canvas into another canvas
  • mjr::sgn() – the sign function
  • mjr::point2d now has a conversion operator for complex numbers
  • ramCanvasTpl now has coordinate conversion methods that work with point2d allowing code like this: std::complex<double> c = theRamCanvas.int2real(x, y);
    • mjr::ramCanvasTpl::int2real
    • mjr::ramCanvasTpl::real2int
  • mjr::fmtInt(int inInt, int width, char fill) – Strings with formatted integers in them. Someday when compilers support the C++20 formatting features, we won't need this…
  • colorTpl::distDeltaE1976(colorArgType aColor) – The LAB Delta E*(1976) distance
  • colorTpl::distDeltaE1994(colorArgType aColor) – The LAB Delta E*(1994) distance
  • Real coordinate overloads for:
    • ramCanvasTpl::drawHersheyGlyph
    • ramCanvasTpl::drawString
    • ramCanvasTpl::drawStringBox
  • Polynomial routines
    • mjr::evalBiPoly – bivariate polynomial evaluation
    • mjr::maxExpBiPoly – figure out maximum theoretical exponent of a bivariate polynomial
    • mjr::evalUniPoly – univariate polynomial evaluation
  • New log transformation function colorTpl::tfrmLn(double scale) – note new name for the old version of this function: tfrmLn1()
• API Change
  • colorTpl::tfrmLn() renamed to tfrmLn1()
  • ramCanvasTpl::autoMaxHistStrech renamed to autoMaxHistStrechRGB
• Updated/New experimental geometric transformations functionality (geomTfrmRev* methods) These feel like they are approaching a final state, but I'm going to keep them classified as "experimental" for now.
  • Fixed a clipping bug for existing geomTfrmRef functions:
    • ramCanvasTpl::geomTfrmRevRPoly
    • ramCanvasTpl::geomTfrmRevAff
    • ramCanvasTpl::geomTfrmRevArb
  • Transform methods now take a uniform set of shift & scale arguments:
    • Xo – X coordinate for the offset from image center.
    • Yo – Y coordinate for the offset from image center.
    • oScale – Scale to apply after transform but before reverse translation
• Started making some unit tests for mjrmath.hpp. Coverage right now is just for the new polynomial routines.

• Removed recently introduced GCC'isms. The code now compiles cleanly with clang and msvc.
• New examples:
  • pickoverPopcorn.cpp
  • mandelbrot_biomorph.cpp
  • mandelbrot_simple.cpp
  • invertedMandelbrot.cpp
  • tippets.cpp
  • multibrotSnaps.cpp
  • multibrotMovie.cpp
  • tricornMovie.cpp
  • tricorn.cpp
  • collatz.cpp
• New web pages for examples:
  • https://www.mitchr.me/SS/BurningShip/index.html
  • https://www.mitchr.me/SS/multibrot/index.html
  • https://www.mitchr.me/SS/tricorn/index.html
  • https://www.mitchr.me/SS/collatz/index.html
  • https://www.mitchr.me/SS/tippets/index.html
  • https://www.mitchr.me/SS/PickoverPopcorn/index.html
  • https://www.mitchr.me/SS/mandelbrotBiomorph/index.html
  • https://www.mitchr.me/SS/mandelbrotInv/index.html
• Updated examples:
  • sic.cpp – Updated the comment on the colorizer object to reference another way to do the same thing illustrated in the pickoverPopcorn.cpp example.
  • biomorph2.cpp & biomorph1.cpp – Cleaned up the way colors were set.
  • mandelbrot_bm_real.cpp & mandelbrot_bm_real_eqop.cpp – Fixed typos in header.
• New feature:
  • ramCanvasTpl::rcConverterColorScheme – Colorize a ramCanvasTpl with integer channels using a color scheme. This gives one the ability to produce a ramCanvasTpl with some data in it, and then colorize it when the image files are written out. For example, we could store the L function for a Mandelbrot render in a grayscale image, and then write out any number of RGB images with different color schemes.
  • New class of color schemes: common fractal schemes
    • colorTpl::csCCfractal0RYBCW
    • colorTpl::csCCfractalYR
    • colorTpl::csCCfractalYB
  • New class of color schemes: Ramp from black to corner
    • colorTpl::csCCu0W – Ramp from black to white
    • colorTpl::csCCu0R – Ramp from black to red
    • colorTpl::csCCu0B – Ramp from black to blue
    • colorTpl::csCCu0G – Ramp from black to green
    • colorTpl::csCCu0C – Ramp from black to cyan
    • colorTpl::csCCu0M – Ramp from black to magenta
    • colorTpl::csCCu0Y – Ramp from black to yellow

I'm done with major changes, and the code feels solid.

Release Candidate #3

• Last time I removed the filter option from the file write APIs, and that annoyed some people. ;) So I have added filters back, but the new filters are more capable in that they can be used for not just homogeneous transforms but geometric ones too.
  • Removed: pixelFormatEnum
  • New converter classes inside ramCanvasTpl:
    • ramCanvasTpl::rcConverterIdentity
    • ramCanvasTpl::rcConverterRGBbyte
    • ramCanvasTpl::rcConverterRGBAbyte
    • ramCanvasTpl::rcConverterRGBdbl
    • ramCanvasTpl::rcConverterRGBAdbl
  • convertRawToTIFF.cpp illustrates the new API
• New: "Canonical Color Types" – color types with unsigned byte and floating point channels in RGB, RGBA, and numChan. These provide a handy interface for conversions with the two most common channel types people use, and are useful for working with external interfaces.
  • Set color from a canonical color type
    • colorTpl& colorTpl::setChans_dbl(colConALLdbl dblColor)
    • colorTpl& colorTpl::setChans_byte(colConALLbyte byteColor)
    • colorTpl& colorTpl::setChansRGBA_dbl(colConRGBAdbl dblColor)
    • colorTpl& colorTpl::setChansRGB_dbl(colConRGBdbl dblColor)
    • colorTpl& colorTpl::setChansRGBA_byte(colConRGBAbyte byteColor)
    • colorTpl& colorTpl::setChansRGB_byte(colConRGBbyte byteColor)
  • convert color to a canonical color type
    • colConALLdbl colorTpl::getColCon_dbl()
    • colConALLbyte colorTpl::getColCon_byte()
    • colConRGBAdbl colorTpl::getColConRGBA_dbl()
    • colConRGBdbl colorTpl::getColConRGB_dbl()
    • colConRGBAbyte colorTpl::getColConRGBA_byte()
    • colConRGBbyte colorTpl::getColConRGB_byte()
  • typedefs for canonical color types
    • colorTpl::colConRGBdbl
    • colorTpl::colConRGBAdbl
    • colorTpl::colConRGBbyte
    • colorTpl::colConRGBAbyte
    • colorTpl::colConALLdbl
    • colorTpl::colConALLbyte
• New integer axis orientation predicates
  • ramCanvasTpl::isIntAxOrientationNaturalX()
  • ramCanvasTpl::isIntAxOrientationNaturalY()
• New helpers to guess named channel. These are helpful when we have a color with unknown structure, and want to get the best option for a named color channel.
  • colorTpl::bestRedChan()
  • colorTpl::bestGreenChan()
  • colorTpl::bestBlueChan()
  • colorTpl::bestAlphaChan()
• Unit tests
  • colorTpl::colCon*
  • colorTpl::best*Chan
• sic.cpp now uses the filter instead of making a new ramCanvasTpl object – like it did in the past.
• Added DO_CIRCLE to bmark.cpp
• Doxygen tag files & links into doxygen docs.
• Unit tests for ramCanvasTpl simple geom transforms

Release Candidate #2

• Bugs:
  • ramCanvasTpl::writeFile
    • Failed to pass return code from stream write helper function.
    • Potential beyond buffer access
• New features:
  • ramCanvasTpl::readRAWfile
• Breaking Changes
  • Replaced the toTRU & filter arguments with an enum to convert to 8-bit RGB/RGBA. In addition, writeTIFF* no longer has an alpha argument – the enum takes care of that too.
• Unit tests for ramCanvasTpl now use isEqual and readRAWfile instead of comparing file content.
• Platform Support
  • Tested with several OS/compiler combinations
    • Windows with MSVC, MSYS2 MSVC, MSYS2 clang, and MSYS2 GCC
    • Debian 11 with GCC-10, and GCC-11
    • MacOS with Apple clang and homebrew GCC
  • Made several little code changes for better platform support.
  • The big change is a couple of macros controlling conditional compilation of code. This allows the code to be compiled, with some reduced functionality, using some older compilers.
    • Added a couple of conditional compiles:
      • MISSING_P0476R2 – Omits code using bit_cast
      • MISSING_P1907R1 – Omits code using floating point template parameters
    • Added cmake rules for conditional code:
      • Detect GCC version and set MISSING_P0476R2 & MISSING_P1907R1 as required.
      • Set MISSING_P1907R1 for clang (both regular and Apple)
      • Set MISSING_P0476R2 for Apple clang

I'm still calling it a beta, but it's pretty solid right now.

Please test and report back!

• New predefined ramCanvas types:
  • ramCanvas4c64b
  • ramCanvas3c64b
  • ramCanvasRGBA64b
  • ramCanvasRGB64b
  • ramCanvas1c32F
  • ramCanvas1c64F
• Added const:
  • colorTpl::rgb2GreyDotProd
  • colorTpl::luminanceRGB
  • colorTpl::intensityRGB
  • colorTpl::intensity
  • colorTpl::intensityScaledRGB
  • colorTpl::intensityScaled
  • colorTpl::getMaxC
  • colorTpl::getMinC
  • colorTpl::getMaxRGB
  • colorTpl::getMinRGB
  • colorTpl::dotProd
  • colorTpl::distHypot
  • colorTpl::distSumAbs
  • colorTpl::distMaxAbs
  • colorTpl::isClose
  • colorTpl::isCloseRGB
  • colorTpl::isEqual
  • colorTpl::isEqualRGB
  • colorTpl::isNotEqual
  • colorTpl::isBlack
  • colorTpl::isBlackRGB
• ramCanvasTpl::drawFillTriangleUtl
  • Better comment for cases & code
  • Much better organization of code for each case
  • Much faster for degenerate triangles using barycentric coloring
  • Much more accurate barycentric coloring, but a slight performance hit.
  • More accurate edges.
• ramCanvasTpl::triangleEdger
  • Added code to find min/max for each y value…
  • I plan on updating this to support cropped triangles at some point.

This is release candidate #1! Please test and report back!

Still beta, but approaching production.

Please test and report back!

• ramCanvasTpl refactor
  • Data member naming consistency
  • Fixed get/set methods to use uniform naming
  • Added requires
  • Comment cleanup
• Clang support
  • Duplicated requires clauses as necessary
  • Added conditional compile around floating point template parameters
  • Added defines to CMakeLists.txt
  • Added warnings about missing examples to CMakeLists.txt
  • Propagated requires outside of class definition
• Bugs
  • colorTpl::distHypot
  • colorTpl::rgb2GreyDotProd
• Upgrades
  • colorTpl::cmpRGBcornerDGradiant – Faster & safer.
• Removed FLTK from CMakeLists.txt – it was not used.

Still beta, but approaching production.

Please test and report back!

• Many more unit tests
• Bug Fixes (mostly related to getC# refactor)
  • setRGBfromUnitHSL
  • cmpRGBcornerDGradiant
  • isBlackRGB
  • isEqualRGB
  • cmpGradiant
  • rgb2colorSpace
  • tfrmStdPowRGB
  • rgb2GreyDotProd
  • luminanceRGB
  • getMaxRGB
  • getMinRGB
• New
  • isClose
  • isCloseRGB
  • linearInterpolateRGB
  • csCCdivBWR
  • csCCdivCWM
  • setChansToMean()
• The color space colorTpl<double, 3> typedef is colSpaceDbl3
• Fixed color swatch names for web safe pallets in documentation.
• Fixed bug with csWSprotanopia class name
• Fixed class names for binary schemes.
• SDL2 support
  • Added SDL2 to CMakeLists.txt so it's easy to add SDL code.
  • Added a new example: displayImageSDL2.cpp
• Better support for msys2 on windows: Added options in CMakeLists.txt to make GLUT, FLTK, & SDL2 code be compiled as a "GUI" application instead of a "console" application as before. The console applications would fail to open a window when run within some contexts. I may need to do something for MSVC too, but I don't know as much about that environment…
• Color documentation (ColorSchemes.org)
  • Added binary pallets
  • Added three segment CC diverging color schemes
  • Added continuous swatches to fixed pallets
  • Added cross reference
• Various documentation updates.

Still beta. Still a few API changes, but things are stabilizing.

Please test and report back!

• Tons of unit tests
• Rearranged some code in colorTpl for better Doxygen results.
• colorTpl::csFP_tpl now provides both discreet and continuous color schemes via overloads to the c member. Much like csCB_tpl.
• New:
  • colorTpl::csWS_tpl – A color scheme template for web safe colors
  • colorTpl::setRGBAfromLogPackIntBGRA
  • colorTpl::setRGBfromLogPackIntBGRA
• Fixed bugs
  • colorTpl::tfrmDiracTot
  • colorTpl::tfrmStdPowRGB
  • colorTpl::tfrmLn
  • colorTpl::tfrmLinearGreyLevelScaleRGB
  • colorTpl::tfrmWebSafeRGB
    • When clrChanT wasn't uint8_t, this function would scale the components incorrectly. Not sure this is really a bug as "web safe" doesn't really apply to such images. Still, it's fixed.
    • Now only operates on the red, blue, and green channels.
  • colorTpl::intensityScaledRGB Scaling was incorrect for images with more than 3 channels.
    • RGBA Bugs Fixed (using getC instead of getRed/Blue/Green/Alpha)
      • colorTpl::tfrmGreyScaleRGB
      • colorTpl::tfrmWebSafePro216
      • colorTpl::tfrmWebSafeDeu216
      • colorTpl::tfrmWebSafeTri216
• Renamed
  • tfrmWebSafe216 -> tfrmWebSafeRGB
  • interplColors(dbl, c1, c2) -> linearInterpolate
  • interplColors(dbl, col) -> tfrmMix
• Removed:
  • colorTpl::tfrmGmeanClamp
  • colorTpl::colorComp2CloseColorComp
  • colorTpl::tfrmWebSafePro216
  • colorTpl::tfrmWebSafeDeu216
  • colorTpl::tfrmWebSafeTri216
  • colorTpl::tfrmLinearGreyLevelScale taking colorTpl
• Web safe color schemes
  • Now use csWS_tpl instead of csFP_tpl.
  • Provide a setter that takes a color object
  • Renamed:
    • csFPwebSafeNormalVision -> csWSnormalVision
    • csFPwXebSafeProtanopia -> csWSnrotanopia
    • csFPwXebSafeDeutanopia -> csWSdeutanopia
    • csFPwXebSafeTritanoptia -> csWStritanoptia
    • csFPwXebSafeProtanopiaAlt -> csWSprotanopiaAlt
    • csFPwXebSafeDeutanopiaAlt -> csWSdeutanopiaAlt
    • csFPwXebSafeTritanoptiaAlt -> csWStritanoptiaAlt

Moving it from alpha to beta. I hope this is the last of the API changes for a while.

Please test and report back!

The API changes are mostly member name changes in colorTpl. These will impact some users, but the refactor should be a simple search-n-replace. Unit tests are fleshing out, and I've fixed a few bugs.

• Fixed a bugs:
  • colorTpl::isBlack()
  • colorTpl::rgbSumIntensity()
• Removed loop from cmpRGBcornerCGradiant()
• Tons of unit tests – up to about 40% coverage for colorTpl
• Removed methods
  • colorTpl::distP2sq – this one wasn't very useful because the sum of the squares was frequently too large to fit into an Sums/Products integer type.
• Renamed methods
  • colorTpl::distAbs -> distSumAbs
  • colorTpl::rgbSumIntensity -> intensitySum
  • colorTpl::rgbLuminance -> luminanceRGB
  • colorTpl::sumIntensity -> intensity
  • colorTpl::rgbScaledIntensity -> IntensityScaledRGB
  • colorTpl::sumScaledIntensity -> intensityScaled
• Changed methods
  • colorTpl::rgb2GreyDotProd & colorTpl::dotProd now return a channelArithFltType and use floating point arithmetic for all intermediate computations.
• New methods
  • colorTpl::distMaxAbs
  • colorTpl::distHypot – Like distP2sq, but the square root so it is much smaller. All arithmetic is done with doubles.
  • colorTpl::cmpGradiant(csFltType csX, csIntType numColors, const packed4Cint* colors).
• colorTpl::cmpGradiant() now uses csFltType for inputs and anchors. Not really an API change as csFltType is double. Just makes things more consistent and readable.
• Added some typedefs for clarity
  • Packed integers
  • color type for color space computations
• Continuous gradient colorbrewer2
  • Added continuous gradients to colorbrewer2 color schemes
  • Added swatches to documentation

Still alpha. Still changing the API. :)

Please test and report back!

The API changes are mostly member name changes in colorTpl. These will impact some users, but the refactor should be a simple search-n-replace. Unit tests are fleshing out, and I've cleaned up the doxygen output.

• Renamed "clp" tfrm methods to "clamp" to better align with C++ use.
  • colorTpl::tfrmMultClp -> tfrmMultClamp
  • colorTpl::tfrmDiffClp -> tfrmDiffClamp
  • colorTpl::tfrmNegDiffClp -> tfrmNegDiffClamp
  • colorTpl::tfrmAddClp -> tfrmAddClamp
  • colorTpl::tfrmAddDivClp -> tfrmAddDivClamp
  • colorTpl::tfrmGmeanClp -> tfrmGmeanClamp
• Renamed "clip" methods to "clamp" to better align with C++ use.
  • colorTpl::clipTop -> clampTop
  • colorTpl::clipAll -> clampAll
  • colorTpl::clipBot -> clampBot
• Renamed "CLIP" drawModeType values to "CLAMP" to better align with C++ use.
  • ramCanvasTpl::drawModeType::ADDCLAMP
  • ramCanvasTpl::drawModeType::AND
  • ramCanvasTpl::drawModeType::OR
  • ramCanvasTpl::drawModeType::DIFFCLAMP
  • ramCanvasTpl::drawModeType::MULTCLAMP
• Many many many more unit tests. Coverage for colorTpl is now at about 25%.
• Published doxygen documentation on web
  • https://www.mitchr.me/SS/mraster/doc-examples/autodocs/html/index.html
  • https://www.mitchr.me/SS/mraster/doc-lib/autodocs/html/index.html
• Added links to published doxygen documentation to github documentation
• Made some updates to doxygen config changes
  • Including the comments in the published source
  • Excluded some problematic bits of code that was confusing Doxygen.

Still alpha. Still changing the API. :)

The big API change is the addition of an argument to ramCanvasTpl, and the removal of ramConfig.hpp. That will all be transparent to people using the color.hpp & ramCanvas.hpp typedefs…

• Added color types for common 8-bit channel RGBA orderings:
  • colorBGR8b -> ramCanvasBGR8b
  • colorARGB8b -> ramCanvasARGB8b
  • colorBGRA8b -> ramCanvasBGRA8b
  • colorABGR8b -> ramCanvasABGR8b
• ramCanvasTpl::writeTGAstream now uses the R, G, & B channels when writing a file if they are available. If not it reverts to the old behavior of using the first three channels.
• ramCanvasTpl::colorizeFltCanvas & ramCanvasTpl::colorizeIntCanvas.
  • color an entire canvas based upon a function that takes pixel coordinates and returns a color.
  • mandelbrot_bm_real_cfun.cpp – Demonstrates colorizeFltCanvas.
• ramCanvasTpl
  • Cleaned up some casts making things a bit safer, and eliminating compiler warnings for edge cases.
  • Moved a few small functions into the class. I hope to move more eventually so as to avoid all the extra requires clauses for externally defined methods.
  • Cleaned up the doxygen overload tags – they just make the document harder to read.
  • New: isOnCanvas – logical inverse of isCliped.
• Removed ramConfig.hpp include
  • The SUPPORT_DRAWING_MODE macro has been replaced with a template parameter (enableDrawModes) to ramCanvasTpl.
  • The REAL_CORD & INT_CORD macros have been replaced with typedefs rcSimpleFltCrdT & rcSimpleIntCrdT in ramCanvas.hpp.
  • The ramConfig class is simply gone.
  • All the 128 bit stuff is now at the top of colorTpl.hpp:
    • macros: MJR_LOOK_FOR_128_BIT_TYPES & MJR_HAVE_128_BIT_TYPES
    • typedefs: mjr_uint128_t & mjr_int128_t
    • Note this stuff is not inside the mjr name space – it's outside just like before.
• Fixed a bug in bmark.cpp – used wrong set_chans function.

Several serious bugs, and two API changing refactors in this update.

This release is BETA quality!!! Please test it, and report bugs.

• Fixed bug with set/clear_byte funcs with clrChanT integer
• Added "color scheme" classes to colorTpl
  • They know how many colors they support
  • Much better encapsulation.
  • The csSet method of colorTpl can be used to set the current color object's color from a color scheme class aColor.csSet<mjr:color3c8b::webSafeNormalVision>(25);
  • Color scheme classes may be used as a factory for colorTpl objects aColor = mjr:color3c8b::webSafeNormalVision.c(25);
  • Color scheme classes may also set the color of a colorTpl object passed in mjr:color3c8b::webSafeNormalVision.c(aColor, 25);
  • Quite a few new color schemes.
  • Some of the old U of O schemes are gone.
  • Many methods are gone replaced by color scheme classes
    • setRGBtoWebSafe216 -> webSafeNormalVision
    • setRGBfromICP -> csCb* & webSafe* classes
    • setRGBcmpBinaryColorRampRG -> csBinRG
    • setRGBcmpBinaryColorRampRB -> csBinRB
    • setRGBcmpBinaryColorRampGR -> csBinRG
    • setRGBcmpBinaryColorRampGB -> csBinGB
    • setRGBcmpBinaryColorRampBR -> csBinRB
    • setRGBcmpBinaryColorRampBG -> csBinGB
    • NEW -> csBin01
    • NEW -> csBinCM
    • NEW -> csBinMY
    • NEW -> csBinYM
    • setRGBcmpUpDownRampRg -> csCCudRg
    • setRGBcmpUpDownRampRb -> csCCudRb
    • setRGBcmpUpDownRampGr -> csCCudGr
    • setRGBcmpUpDownRampGb -> csCCudGb
    • setRGBcmpUpDownRampBr -> csCCudBr
    • setRGBcmpUpDownRampBg -> csCCudBg
    • setRGBcmpGrey -> csCCdiag01
    • setRGBcmpDiagRampCR -> csCCdiagCR
    • setRGBcmpDiagRampMG -> csCCdiagMG
    • setRGBcmpDiagRampYB -> csCCdiagYB
    • setRGBcmpConstTwoRamp -> csCCconsTwo
    • setRGBcmpConstOneRamp -> csCCconsOne
    • setRGBcmpFireRamp -> csCColdeFireRamp
    • setRGBcmpColdToHot -> csCColdeColdToHot
    • setRGBcmpIceToWaterToHot -> csCColdeIceToWaterToHot
    • setRGBcmpSumRampRGB -> csCCsumRGB
    • setRGBcmpSumRampBGR -> csCCsumBGR
    • setRGBcmpSumRampGRB -> csCCsumGRB
    • setRGBcmpSumRampBRG -> csCCsumBRG
    • setRGBcmpSumRampGBR -> csCCsumGBR
    • setRGBcmpSumRampRBG -> csCCsumRBG
    • setRGBcmpGreyRGB -> csCCdiag01
    • setRGBcmpGrey3x -> csPGrey3x
    • setRGBcmpGrey4x -> csPGrey4x
    • setRGBcmpRampGrey2R -> csHSLhR
    • setRGBcmpRampGrey2G -> csHSLhG
    • setRGBcmpRampGrey2B -> csHSLhB
    • setRGBcmpRampGrey2C -> csHSLhC
    • setRGBcmpRampGrey2M -> csHSLhM
    • setRGBcmpRampGrey2Y -> csHSLhY
    • setRGBcmpClrCubeRainbow -> csCColdeRainbow
    • setRGBcmpRainbowHSV -> csCColdeRainbow
    • colorTpl& setRGBcmpRainbowLA -> csRainbowLA
    • colorTpl& setRGBcmpRainbowCM -> csRainbowCM
• Colorbrewer
  • Updated to colorbrewer2
  • Everything is now implemented as color schemes in the colorTpl class
  • Some of the old schemes are gone now.
• Removed colorData.cpp & colorData.hpp
• Added RGBA specific set & get methods
  • The R, G, B, & A channel indexes are specified as colorTpl arguments
  • Allows one to specify the RAM layout for colors. For example we can make frame buffers in GL_RGBA or GL_BGRA layout.
• Reduced some of the Doxygen comment duplication, and regrouped some members. This makes the code much easier to read.
• Added heart2022.cpp
• hershey refactor
  • The old hershey data include and source are gone.
  • The library is 100% header only now!
  • The old hershey stuff is now in a hershey class.
• Several doc & bug fixes

One might consider this a brand new library; however, I expect most users will need to make few changes – and the changes made will be pretty simple search-n-replace type changes.

• This release ie PRE-ALPHA!!! Please test it, and report bugs.
• C++20 is now the language
• Much of the krazy code required for performance – with zero performance impact with gcc 11.2.0, clang 14.0.0, and MSVC 2022. In many cases this was achievable simply due to better compilers, but much of it is a direct result of new C++20 features.
• Template parameters are simplified
  • Many former parameters are now derived from remaining parameters.
  • Concepts are used to constrain types and specialize
• Removed non-RGB references where not really related to RGB – partly related to removing the optimizing "name" struct.
• Floating point channels are back!
• Many performance improvements.

• Bugs
  • Fixed bug with << operator for colorTpl
  • Fixed a few edge case casts with very deep images
  • Tighter constraints in colorTpl static_assert block
• Performance improvements
  • ramCanvasTpl::convolution is roughly 2x faster
  • Added performance notes to ramCanvasTpl::clrCanvasToBlack()
• Compiler support
  • A few code & CMakeLists.txt updates for clang
  • More testing on MSVC++
• Improved colorspace support
  • More spaces: HSL, HSV, LAB, RGB, LCH, & XYZ
  • Added a new enum to specify color space: colorSpaceEnum
    • Removed the old enum: colorInterpMethEnum
  • Specific color space functions replaced with general ones taking a colorspace argument.
    • setColorFromColorSpace replaces setColorFromNaturalHSV & setColorFromNaturalHSL
    • rgb2colorSpace replaces rgb2hsv & rgb2hsl
    • interplColorSpace replaces interplColorsHSL
• mapclamp.hpp is now mjrmath.hpp
  • Has a constant for PI.
  • Added several new functions
    • interpolateLinear
    • interpolateLinearAnglesDeg
    • realWrap
  • Corrected/changed negative wrap behavior in intWrap
• Examples
  • Added high resolution timers to most examples
  • bmark.cpp has more tests (including one for convolution)
  • Examples use types from ramCanvasTpl not colorTpl
  • Reworked circles example
  • Uniform formatting in examples
  • Fleshed out cplxColor
    • Expanded code demoing how to make nice colors
    • Added a few Eisenstein series examples to cplxColor.cpp
  • color_interp_hsl_vs_rgb.cpp
    • Demos all the new interpolation methods
    • Labels the methods in the resulting image which is why it moved in CMakeLists.txt
    • Illustrates the color space conversion functions as well
  • color_interp_hsl_vs_rgb is improved
    • Demonstrates the various interpolation methods
    • Labels the image with which method is used for each strip
• Documentation
  • Updated to newer Doxygen
  • Changed flowerbox comments take advantage of Doxygen's new JAVADOC_BANNER option
  • Lots of little fixes (spelling, grammar, whitespace cleanup)
  • Better docs for colorChanArithT
• Code reformatting
  • Put member return type separate line
  • Made sure multi-line argument lists are lined up

• Removed floating point argument from color template (API breaking change)
  • I expect most people probably use the pre-defined types in color.hpp and ramCanvas.hpp, so they may not even have exposure to this change – Other than fewer compiler complaints about implicit type casts.
  • I looked through my code and didn't find a single place where I took advantage of the floating pint arithmetic type. It was useful back when the library could have floating point channels, but now it's just sort of a vestigial toe.
  • All floating point color computations are now done with doubles
  • This simplifies the code, and makes it easier to use
  • I have included some compatibility typedefs to mask the change:
    • clrChanIArthT
    • channelIntArithType
    • clrChanFArthT
    • channelFltArithType
  • In general one should avoid the compatibility typedefs and use
    • double instead of clrChanFArthT or channelFltArithType
    • channelArithType instead of clrChanIArthT or channelIntArithType
• Several typedef changes in ramCanvasTpl. I have attempted to make the typedef names more uniform, and simplify access to contained types. All the examples have been updated to use the new names.
  • New typedefs
    • pointFltType …….. Real coordinate pair type
    • pointIntType …….. Integer coordinate pair type
    • coordIntType …….. Integer type for coordinates
    • coordFltType …….. Real type for coordinates
    • colorType ……….. Color type for pixels
    • colorChanType ……. Channel type for color type for pixels
    • colorChanArithType .. Type for integer channel arithmetic
  • Deprecated typedefs
    • rcPointFlt
    • rcPointInt
    • rcCordInt
    • rcCordFlt
    • rcColor

• New/Improved interpolation methods
  • getPxColorInterpBLin
  • getPxColorInterpTrunc
  • getPxColorInterpNear
  • getPxColorInterpAvg4
  • getPxColorInterpAvg9
• Example test_gtfrm_scale.cpp illistrates all interpolation methods by scaleing an image
• Example bmark.cpp has benchmarks for getPxColorInterpAvg9

• The string drawing functions now take std::string instead of c-style strings
• Added static_cast in several places we had automatic conversions to document that we really do want that cast.
• More careful about using types from the classes instead of integers in the examples for things like image coordinates and color channels. It's still a bit uneven – even sloppy in some of the examples.
• Experimental support for MSVC++ (tested with MSVS CC 2022). I may well have introduced some bugs or different behavior – especially in the examples. So be on the lookout for oddness.
• I have recently become interested in arbitrary geometric transformations, and just added a bilinear interpolation method. I expect to adding several interpolation methods along with a generic interface to them later. I also plan on adding a canvas transformation method. Lastly, the example test_gtfrm_scale.cpp will eventually be used to demonstrate that transformation method – right now it just demonstrates the interpolation bits…
  • White space cleanup across the code base. So if you do diffs, make sure to tell them to ignore white space changes.
  • Removed all the extra template arg bits in the colorTpl template. Not sure how they got in there – probably a fit of cut-n-paste at some point? They were not errors, just unnecessary.
  • Hard coded the C++ version to 2011 just to make sure the compiler was using something at least that new.

• New version numbersk
• Several new examples:
  • examples/brownianDiffusion.cpp
  • examples/dlaDriver.sh
  • examples/lorenz_fuz.cpp
  • examples/lorenz_multi.cpp
  • examples/sprott2d.cpp
• Added some aux files:
  • examples/mandelbrot_potential_a_0.pov
  • examples/mandelbrot_potential_a_1.pov
  • examples/mandelbrot_potential_a_1t.pov
  • examples/mandelbrot_potential_a_2.pov
  • examples/mandelbrot_potential_b_0.bov
  • examples/mandelbrot_potential_b_0.pov
  • examples/mandelbrot_potential_b_1.pov
  • examples/mandelbrot_potential_b_2.pov
• Updated a few examples:
  • examples/circles.cpp
  • examples/dlaSeed.cpp
  • examples/peterdejong.cpp
  • examples/sic_search.cpp
• A few cmake & build changes
• A couple updates to the base library