Arch Linux User Repository

updated to latest version and patched the error

I am getting following error;

[ 58%] Building C object src/CMakeFiles/MagPlus.dir/drivers/minizip/ioapi.c.o
[ 59%] Building C object src/CMakeFiles/MagPlus.dir/drivers/minizip/zip.c.o
/var/tmp/pamac-build-bakamotokatas/magics++/src/Magics-4.15.4-Source/src/drivers/minizip/zip.c: In function ‘zipOpenNewFileInZip3’:
/var/tmp/pamac-build-bakamotokatas/magics++/src/Magics-4.15.4-Source/src/drivers/minizip/zip.c:883:28: error: assignment to ‘const long unsigned int *’ from incompatible pointer type ‘const z_crc_t *’ {aka ‘const unsigned int *’} [-Wincompatible-pointer-types]
  883 |         zi->ci.pcrc_32_tab = get_crc_table();
      |                            ^
make[2]: *** [src/CMakeFiles/MagPlus.dir/build.make:6576: src/CMakeFiles/MagPlus.dir/drivers/minizip/zip.c.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1243: src/CMakeFiles/MagPlus.dir/all] Error 2
make: *** [Makefile:166: all] Error 2 

@dquesada that looks like a memory access beyond the size of a vector, cannot assess if it is coming from cdo or magics without a stack. Have you tried to ask to the cdo developers if they have tested their software with gcc 13?

You could try to compile magics and cdo using gcc11 from the aur to pinpoint if the problem is really connected to gcc13.

@capitalaslash, thanks a lot for updating the package!

Unfortunately it did not solve my issue, which remains:

/usr/include/c++/13.1.1/bits/stl_vector.h:1125: std::vector<_Tp, _Alloc>::reference std::vector<_Tp, _Alloc>::operator[](size_type) [with _Tp = double; _Alloc = std::allocator<double>; reference = double&; size_type = long unsigned int]: Assertion '__n < this->size()' failed.

I have rebuilt and updated all the related packages in the meantime. If you have any tips on how can this be solved I would greatly appreciate it.

Cheers!

@dquesada I have updated the package to the latest version (and it does not require the previous patch). I do not use cdo so I cannot verify your problem and/or if this update fixed it, please let me know if this solves it.

Hi @capitalaslash, can you please update the package? Also, the gcc11.patch was only meant for gcc-fortran v11? Some of the plotting capabilities (e.g. shaded) of cdo with magics are not working and I am afraid it could be related to this issue, or the outdated version.

• Call:

cdo -shaded,device=pdf,lon_max=-50,lon_min=-110,lat_max=30,lat_min=-10,interval=2,colour_triad=cw,colour_min=violet,colour_max=red -seltimestep,42 gfdl-esm4_r1i1p1f1_w5e5_historical_pr_global_daily_2011_2014.nc shaded

• Output:

cdo(1) seltimestep: Process started /usr/include/c++/13.1.1/bits/stl_vector.h:1125: std::vector<_Tp, _Alloc>::reference std::vector<_Tp, _Alloc>::operator[](size_type) [with _Tp = double; _Alloc = std::allocator<double>; reference = double&; size_type = long unsigned int]: Assertion '__n < this->size()' failed. Aborted (core dumped)

Thanks!

ECMWF updated the file location. Should be working now.

I am trying to install on Manjaro with pamac. It fails to build because the validation fails.

Building magics++...
==> Making package: magics++ 4.9.4-0 (So 12 Dez 2021 11:56:43 CET)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Found Magics-4.9.4-Source.tar.gz
  -> Found gcc11.patch
==> Validating source files with sha256sums...
    Magics-4.9.4-Source.tar.gz ... FAILED
    gcc11.patch ... Passed
==> ERROR: One or more files did not pass the validity check!
Error: Failed to build magics++

Thank you, @dquesada. It works correct now.

In case somebody runs with the same issue, libgeotiff is on /usr/include on Manjaro, so, the flag -DGEOTIFF_INCLUDE_DIR=/usr/include \ is needed after -DENABLE_GEOTIFF=ON \