Segfault when using panel function of PyNGL

[weech]

Not sure what could be causing it. Here's a sample script to produce it (on Linux, Julia 1.1.0):

using PyCall
const ngl = pyimport("Ngl")

function main()

    arr1 = ngl.generate_2d_array([180, 360], 5, 4, -5, 6)
    arr2 = ngl.generate_2d_array([180, 360], 4, 3, -4, 5)

    lons = range(0, 360, length=360)
    lats = range(-90, 90, length=180)

    wks = ngl.open_wks("png", "pyngltest")
    res = ngl.Resources()
    res.sfXArray = lons
    res.sfYArray = lats
    res.nglDraw = false
    res.nglFrame = false

    plot1 = ngl.contour_map(wks, arr1, res)
    plot2 = ngl.contour_map(wks, arr2, res)
    ngl.panel(wks, [plot1, plot2], [2, 1])

    ngl.end()

end
main()

And the stacktrace:

signal (11): Segmentation fault
in expression starting at /home/alex/Documents/julia/pyngltest.jl:28
_wrap_panel_wrap at /home/alex/miniconda3/lib/python3.7/site-packages/ngl/_hlu.cpython-37m-x86_64-linux-gnu.so (unknown line)
_PyMethodDef_RawFastCallKeywords at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
_PyCFunction_FastCallKeywords at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
call_function.lto_priv.1536 at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
_PyEval_EvalFrameDefault at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
_PyEval_EvalCodeWithName at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
_PyFunction_FastCallDict at /home/alex/miniconda3/lib/libpython3.7m.so.1.0 (unknown line)
macro expansion at /home/alex/.julia/packages/PyCall/1815y/src/exception.jl:81 [inlined]
__pycall! at /home/alex/.julia/packages/PyCall/1815y/src/pyfncall.jl:44
_pycall! at /home/alex/.julia/packages/PyCall/1815y/src/pyfncall.jl:29
_pycall! at /home/alex/.julia/packages/PyCall/1815y/src/pyfncall.jl:11
jl_fptr_trampoline at /buildworker/worker/package_linux64/build/src/gf.c:1864
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
#call#111 at /home/alex/.julia/packages/PyCall/1815y/src/pyfncall.jl:89
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
jl_apply at /buildworker/worker/package_linux64/build/src/julia.h:1571 [inlined]
jl_f__apply at /buildworker/worker/package_linux64/build/src/builtins.c:556
PyObject at /home/alex/.julia/packages/PyCall/1815y/src/pyfncall.jl:89
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
main at /home/alex/Documents/julia/pyngltest.jl:23
jl_fptr_trampoline at /buildworker/worker/package_linux64/build/src/gf.c:1864
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
do_call at /buildworker/worker/package_linux64/build/src/interpreter.c:323
eval_value at /buildworker/worker/package_linux64/build/src/interpreter.c:411
eval_stmt_value at /buildworker/worker/package_linux64/build/src/interpreter.c:362 [inlined]
eval_body at /buildworker/worker/package_linux64/build/src/interpreter.c:773
jl_interpret_toplevel_thunk_callback at /buildworker/worker/package_linux64/build/src/interpreter.c:885
unknown function (ip: 0xfffffffffffffffe)
unknown function (ip: 0x7f03977efe3f)
unknown function (ip: 0xffffffffffffffff)
jl_interpret_toplevel_thunk at /buildworker/worker/package_linux64/build/src/interpreter.c:894
jl_toplevel_eval_flex at /buildworker/worker/package_linux64/build/src/toplevel.c:764
jl_parse_eval_all at /buildworker/worker/package_linux64/build/src/ast.c:883
jl_load at /buildworker/worker/package_linux64/build/src/toplevel.c:826
include at ./boot.jl:326 [inlined]
include_relative at ./loading.jl:1038
include at ./sysimg.jl:29
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
exec_options at ./client.jl:267
_start at ./client.jl:436
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:2219
unknown function (ip: 0x40191d)
unknown function (ip: 0x401523)
__libc_start_main at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
unknown function (ip: 0x4015c4)
Allocations: 9077386 (Pool: 9076307; Big: 1079); GC: 20

The same script in Python has no issues. I haven't had issues using any other PyNGL functions, but Ngl.panel causes Julia to crash every time.

[stevengj]

Two common causes:

• MKL conflict (if you are using the MKL Julia): Segfault with Julia MKL build (library conflict) #443
• LLVM conflict (if this library uses LLVM): Add a build option to rename all LLVM symbols JuliaLang/julia#12644

Could one of these two conditions apply?

[weech]

No, I'm using the Julia from the downloads page on the website, and neither PyNGL nor any of the underlying NCL libraries make use of LLVM.

[stevengj]

Try

pycall(ngl.generate_2d_array, PyObject, [180, 360], 5, 4, -5, 6)
pycall(ngl.generate_2d_array, PyObject, [180, 360], 4, 3, -4, 5)

in case it is getting confused by Julia arrays (e.g. if it falsely assumes all NumPy arrays are row-major it will get confused by Julia column-major NumPy arrays).

[weech]

I tried that, and that doesn't work either. I went ahead and compiled PyNGL with debugging options, and it is segfaulting on the dims argument, the one that is [2, 1] in my example. I got it to break with Python by putting a np.array([2, 1]) in there, so basically this program does not accept NumPy arrays. I got it to work in Julia by passing dims as a tuple. Perhaps I should open an issue with PyNGL.