plotting causes SEGFAULT

[dlfivefifty]

julia> versioninfo()
Julia Version 1.0.2
Commit d789231e99 (2018-11-08 20:11 UTC)
Platform Info:
  OS: macOS (x86_64-apple-darwin18.2.0)
  CPU: Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-6.0.0 (ORCJIT, skylake)

julia> using PyPlot

julia> plot(1:10)
Segmentation fault: 11

A window briefly appears before the SEGFAULT.

[cormullion]

A bit more information from the crash on my mac:

julia-1.1> using PyPlot

julia-1.1> x = range(0; stop=2*pi, length=1000); y = sin.(3 * x + 4 * cos.(2 * x));

julia-1.1> plot(x, y, color="red", linewidth=2.0, linestyle="--")
2019-01-24 09:11:10.174 julia[662:29121764] -[NSApplication _setup:]: unrecognized selector sent to instance 0x7fbafd8318c0
2019-01-24 09:11:10.174 julia[662:29121764] *** Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '-[NSApplication _setup:]: unrecognized selector sent to instance 0x7fbafd8318c0'
*** First throw call stack:
(
)
libc++abi.dylib: terminating with uncaught exception of type NSException

signal (6): Abort trap: 6
in expression starting at no file:0
__pthread_kill at /usr/lib/system/libsystem_kernel.dylib (unknown line)
Allocations: 12760827 (Pool: 12759145; Big: 1682); GC: 27
Abort trap: 6

using [d330b81b] PyPlot v2.7.0

Julia Version 1.1.0
Commit 80516ca202 (2019-01-21 21:24 UTC)
Platform Info:
OS: macOS (x86_64-apple-darwin14.5.0) (Mojave)
CPU: Intel(R) Core(TM) i5-7500 CPU @ 3.40GHz
WORD_SIZE: 64
LIBM: libopenlibm
LLVM: libLLVM-6.0.1 (ORCJIT, skylake)

[cormullion]

(May be the same as #253.)

[stevengj]

It works fine for me on my Mac with the Conda python. What are PyPlot.backend and PyPlot.PyCall.python? I have:

julia> PyPlot.backend
"qt5agg"

julia> PyPlot.PyCall.python
"/Users/stevenj/.julia/conda/3/bin/python"

[dlfivefifty]

Exactly the same thing:

julia> PyPlot.backend
"qt5agg"

julia> PyPlot.PyCall.python
"/Users/solver/.julia/conda/3/bin/python"

[stevengj]

Does it work in Python? Try:

using PyCall
ENV["MPLBACKEND"]="qt5agg"
run(`$(PyCall.python) -c "import matplotlib.pyplot as plt; plt.plot([1,2,3,4]); plt.show()"`)

[dlfivefifty]

Yes, it works!

[stevengj]

I still can't reproduce. I'm using Julia 1.1.0 on macos 10.14, and I ran Conda.update() to make sure I'm using the latest Conda packages; can you try that?

[dlfivefifty]

Just did and still it crashes

[stevengj]

Are you using JuliaPro, the standard julialang.org download, or a custom build?

[dlfivefifty]

Julia 1.0.2 built from source with MKL:

julia> versioninfo()
Julia Version 1.0.2
Commit d789231e99 (2018-11-08 20:11 UTC)
Platform Info:
  OS: macOS (x86_64-apple-darwin18.2.0)
  CPU: Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-6.0.0 (ORCJIT, skylake)

[stevengj]

Duplicate of JuliaPy/PyCall.jl#443.

Julia's MKL conflicts with the one used in NumPy.

[stevengj]

By the way, what does LinearAlgebra.BLAS.vendor() print for such builds? It might be nice to print a warning.

[dlfivefifty]

julia> LinearAlgebra.BLAS.vendor()
:mkl

[stevengj]

Warning added in JuliaPy/PyCall.jl#637