Arch Linux User Repository

Until the Octave binding is fixed, version 2.1.0-4 of this package, which only differs in not including it, builds perfectly well.

After the recent upgrade of Octave from 5.2.0 to 6.1.0, the "few deprecation warnings" in the Octave bindings turned into build breakage. The first error I’m hitting seems to be due to use of the deprecated method octave_value::is_map(), which was removed in 2018 in changeset f23f27e78aa2.

/build/pfstools/src/pfstools-2.1.0/src/octave/pfsget.cpp: In function ‘octave_value_list Fpfsget(const octave_value_list&, int)’:
/build/pfstools/src/pfstools-2.1.0/src/octave/pfsget.cpp:49:31: error: ‘const class octave_value’ has no member named ‘is_map’
   49 |   if( nargin != 1 || !args(0).is_map() )
      |                               ^~~~~~

This should be patchable (the deprecation notices did mention a replacement), but I’m not up to it today, sorry. Upstream has bug #49 for this, but they haven’t yet fixed this, either.

disowned

Need a fix for opencv4, here's a patch I've used:

git am < <(curl -s https://pastebin.com/raw/QtcZwRed)

Run in your local clone.

Fails to build for me:

Scanning dependencies of target pfsalign
[ 95%] Building CXX object src/camera/CMakeFiles/pfsalign.dir/pfsalign.cpp.o
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp: In function ‘bool alignImagePair(const cv::Mat&, const cv::Mat&, cv::Mat&, int, bool)’:
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp:233:44: error: ‘CV_RANSAC’ was not declared in this scope
  233 |     homography = findHomography( pp2, pp1, CV_RANSAC, 1., inliners );
      |                                            ^~~~~~~~~
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp: In function ‘void alignFrames(int, char**)’:
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp:650:62: warning: narrowing conversion of ‘frames.std::vector<FrameInfo>::operator[](((std::vector<FrameInfo>::size_type)kk)).FrameInfo::size.cv::Size_<int>::width’ from ‘int’ to ‘double’ [-Wnarrowing]
  650 |         double data[4][3] = { { 0, 0, 1 }, { frames[kk].size.width, 0, 1 }, { frames[kk].size.width, frames[kk].size.height, 1 }, { 0, frames[kk].size.height, 1 } };
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp:650:95: warning: narrowing conversion of ‘frames.std::vector<FrameInfo>::operator[](((std::vector<FrameInfo>::size_type)kk)).FrameInfo::size.cv::Size_<int>::width’ from ‘int’ to ‘double’ [-Wnarrowing]
  650 |         double data[4][3] = { { 0, 0, 1 }, { frames[kk].size.width, 0, 1 }, { frames[kk].size.width, frames[kk].size.height, 1 }, { 0, frames[kk].size.height, 1 } };
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp:650:118: warning: narrowing conversion of ‘frames.std::vector<FrameInfo>::operator[](((std::vector<FrameInfo>::size_type)kk)).FrameInfo::size.cv::Size_<int>::height’ from ‘int’ to ‘double’ [-Wnarrowing]
  650 |         double data[4][3] = { { 0, 0, 1 }, { frames[kk].size.width, 0, 1 }, { frames[kk].size.width, frames[kk].size.height, 1 }, { 0, frames[kk].size.height, 1 } };
/tmp/yaourt-tmp-felics/aur-pfstools/src/pfstools-2.1.0/src/camera/pfsalign.cpp:650:152: warning: narrowing conversion of ‘frames.std::vector<FrameInfo>::operator[](((std::vector<FrameInfo>::size_type)kk)).FrameInfo::size.cv::Size_<int>::height’ from ‘int’ to ‘double’ [-Wnarrowing]
  650 |         double data[4][3] = { { 0, 0, 1 }, { frames[kk].size.width, 0, 1 }, { frames[kk].size.width, frames[kk].size.height, 1 }, { 0, frames[kk].size.height, 1 } };
make[2]: *** [src/camera/CMakeFiles/pfsalign.dir/build.make:63: src/camera/CMakeFiles/pfsalign.dir/pfsalign.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1879: src/camera/CMakeFiles/pfsalign.dir/all] Error 2
make: *** [Makefile:152: all] Error 2
==> ERROR: A failure occurred in build().
    Aborting...

The force magick6 patch results in

-- Could NOT find ImageMagick: Found unsuitable version "7.0.8-68", but required is exact version "6" (found /usr/lib/libMagick++-6.Q16HDRI.so)

-- ImageMagick not found. The following commands will not be compiled: pfsinimgmagick pfsoutimgmagick pfsouthdrhtml.

ImageMagick CFLAGS: /usr/include/ImageMagick-7

related bug: https://sourceforge.net/p/pfstools/bugs/46/ . Probably better to disable ImageMagick than include this path which makes it fail.

Working PKGBUILD for opencv4 : https://github.com/bartoszek/AUR-pfstools

opencv3.patch need update for opencv4 https://pastebin.com/ddbEwh6W

For a build in a minimal environment package gtk3 was a missing (make) dependency.

I get the following build error when extra/imagemagick6 is installed:

/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp: In function ‘void readFrames(int, char)’:
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:117:29: error: ‘class Magick::Image’ has no member named ‘matte’
     bool hasAlpha = imImage.matte();
                             ^~~~~
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:132:39: error: ‘QuantumDepth’ was not declared in this scope
     const float maxValue = (float)(1<<QuantumDepth) - 1;
                                       ^~~~~~~~~~~~
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:132:39: note: suggested alternative: ‘QuantumRange’
     const float maxValue = (float)(1<<QuantumDepth) - 1;
                                       ^~~~~~~~~~~~
                                       QuantumRange
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:134:21: error: ‘PixelPacket’ in namespace ‘Magick’ does not name a type
       const Magick::PixelPacket pixels =
                     ^~~~~~~~~~~
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:138:31: error: ‘pixels’ was not declared in this scope
         (X)(pixInd) = (float)pixels[c].red / maxValue;
                               ^~~~~~
/path/to/build/src/pfstools-2.1.0/src/fileformat/pfsinimgmagick.cpp:138:31: note: suggested alternative: ‘pixInd’
         (*X)(pixInd) = (float)pixels[c].red / maxValue;
                               ^~~~~~
                               pixInd
make[2]:  [src/fileformat/CMakeFiles/pfsinimgmagick.dir/build.make:63: src/fileformat/CMakeFiles/pfsinimgmagick.dir/pfsinimgmagick.cpp.o] Error 1
make[1]:  [CMakeFiles/Makefile2:804: src/fileformat/CMakeFiles/pfsinimgmagick.dir/all] Error 2