Arch Linux User Repository

@periode I have just checked, and the package is building perfectly fine. I cannot reproduce your issue. By your log, it looks like you are using an AUR helper, and this is not supported. Use makepkg, and make sure that you have an up-to-date Arch Linux system.

Hi, I get the following error on trying to update to 1.6.0-1:

In member function ‘deallocate’,
    inlined from ‘deallocate’ at /usr/include/c++/15.2.1/bits/alloc_traits.h:649:23,
    inlined from ‘_M_deallocate’ at /usr/include/c++/15.2.1/bits/stl_vector.h:396:19,
    inlined from ‘_M_deallocate’ at /usr/include/c++/15.2.1/bits/stl_vector.h:392:7,
    inlined from ‘__dt_base ’ at /usr/include/c++/15.2.1/bits/stl_vector.h:375:15,
    inlined from ‘__dt_base ’ at /usr/include/c++/15.2.1/bits/stl_vector.h:805:7,
    inlined from ‘Match’ at /usr/src/debug/chromaprint-fftw/chromaprint-1.6.0/src/fingerprint_matcher.cpp:178:2,
    inlined from ‘Match’ at /usr/src/debug/chromaprint-fftw/chromaprint-1.6.0/src/fingerprint_matcher.cpp:48:14,
    inlined from ‘TestBody’ at /usr/src/debug/chromaprint-fftw/chromaprint-1.6.0/tests/test_fingerprint_matcher.cpp:24:15:
/usr/include/c++/15.2.1/bits/new_allocator.h:172:66: warning: ‘operator delete’ called on pointer ‘gradient_peaks.D.6216._M_impl.D.6191._M_finish’ with nonzero offset 8 [-Wfree-nonheap-object]
  172 |         _GLIBCXX_OPERATOR_DELETE(_GLIBCXX_SIZED_DEALLOC(__p, __n));
      |                                                                  ^
/home/pierre/.cache/yay/chromaprint-fftw/src/build/tests/all_tests: symbol lookup error: /home/pierre/.cache/yay/chromaprint-fftw/src/build/tests/all_tests: undefined symbol: chromaprint_decode_fingerprint_header
CMake Error at /usr/share/cmake/Modules/GoogleTestAddTests.cmake:132 (message):
  Error running test executable.

    Path: '/home/pierre/.cache/yay/chromaprint-fftw/src/build/tests/all_tests'
    Working directory: '/home/pierre/.cache/yay/chromaprint-fftw/src/chromaprint-1.6.0'
    Result: 127
    Output:


Call Stack (most recent call first):
  /usr/share/cmake/Modules/GoogleTestAddTests.cmake:275 (gtest_discover_tests_impl)


make[2]: *** [tests/CMakeFiles/all_tests.dir/build.make:494: tests/all_tests] Error 1
make[2]: *** Deleting file 'tests/all_tests'
make[1]: *** [CMakeFiles/Makefile2:309: tests/CMakeFiles/all_tests.dir/all] Error 2
make: *** [Makefile:146: all] Error 2
==> ERROR: A failure occurred in build().
    Aborting...
 -> error making: chromaprint-fftw-exit status 4
 -> Failed to install the following packages. Manual intervention is required:
chromaprint-fftw - exit status 4

looks like chromaprint_decode_fingerprint_header is missing? Any thoughts on this?

@slapjd repository chromaprint needs ffmpeg, but here we use it in the other way around: ffmpeg-full and ffmpeg-full-git requires chromaprint. And yes, the goal is mainly to avoid circular dependency with ffmpeg, as this can be a pain, specially when ffmpeg increases the soname version (as it is happening with ffmpeg-full-git right now at the time of writing). It's a good practice to avoid circular dependencies whenever possible, specially on AUR packages. And splitting chromaprint into different packages for tools and libs does not seem a good idea to me just to separate a single binary executable that is less then 40kb, and this would also cause a circular dependency at build time.

@dbermond ah, i see. is this package trying to avoid ffmpeg dependencies entirely? thinking about it, i would assume fpcalc only uses ffmpeg directly for transcoding the input file to raw audio and then uses actual chromaprint library for any calculations.

i only ran into issues because i started using ffmpeg-full (for better AAC transcodes) which depends on this specifically, presumably to avoid circular dependency problems? for now i just copied the fpcalc binary into /usr/local/bin to make picard work (for anyone else with this issue).

it would be nice if the official package was split into chromaprint and chromaprint-tools or fpcalc or something, then this could maybe do the same and avoid circular dependency problems while still being able to depend on fpcalc (although packages only requiring the library could/should probably just depend on libchromaprint.so then this could only provide that).

regardless, thanks for clarifying!

@slapjd because 'BUILD_TOOLS' requires ffmpeg, and this package uses fftw instead of ffmpeg for the fft calculations.

Hi, is there a reason DBUILD_TOOLS is off? Official chromaprint has it on and Picard depends on it (fpcalc) for AcoustID.

@Culted package updated. Thanks for reporting.

i see what happened 2 days ago gtest from the official repository was updated and no longer ships the source files

getting an error during compilation

-- Using swresample for audio conversion -- Found GTest: /usr/src/googletest -- Configuring done (2.2s) CMake Error at tests/CMakeLists.txt:4 (add_library): Cannot find source file:

/usr/src/googletest/src/gtest-all.cc

Tried extensions .c .C .c++ .cc .cpp .cxx .cu .mpp .m .M .mm .ixx .cppm .ccm .cxxm .c++m .h .hh .h++ .hm .hpp .hxx .in .txx .f .F .for .f77 .f90 .f95 .f03 .hip .ispc

CMake Error at tests/CMakeLists.txt:4 (add_library): No SOURCES given to target: gtest

tried using googletest-git,same issue

@JeSuisNerd This is an issue with gtest 1.13.0, which does not support C++11 anymore. Fixed.