Kernel crashing when setting region in grdfill

[willschlitzer]

When trying to set the region on grdfill, it routinely causes my notebook to crash (I'm using Jupyter Lab). There don't appear to be any issues when I don't set the region.

Getting the grid file

grid = load_static_earth_relief()
grid[3:6, 3:5] = np.nan
grid[6:8, 2:4] = np.inf

Code that didn't cause any error

result = pygmt.grdfill(grid=grid, mode="c20")

Full code that generated the error

result = pygmt.grdfill(grid=grid, mode="c20", region=[-55, -50, -24, -16])

Full error message

I don't get a Python error message when this crashes. This is what pops up:

System information

Please paste the output of python -c "import pygmt; pygmt.show_versions()":

PyGMT information:
  version: v0.5.1.dev74+g3b896adc.d20220207
System information:
  python: 3.9.9 | packaged by conda-forge | (main, Dec 20 2021, 02:41:03)  [GCC 9.4.0]
  executable: /home/will/miniconda3/envs/pygmt-dev/bin/python
  machine: Linux-5.13.0-28-generic-x86_64-with-glibc2.31
Dependency information:
  numpy: 1.21.4
  pandas: 1.3.3
  xarray: 0.21.1
  netCDF4: 1.5.7
  packaging: 21.3
  ghostscript: 9.54.0
  gmt: 6.3.0
GMT library information:
  binary dir: /home/will/miniconda3/envs/pygmt-dev/bin
  cores: 12
  grid layout: rows
  library path: /home/will/miniconda3/envs/pygmt-dev/lib/libgmt.so
  padding: 2
  plugin dir: /home/will/miniconda3/envs/pygmt-dev/lib/gmt/plugins
  share dir: /home/will/miniconda3/envs/pygmt-dev/share/gmt
  version: 6.3.0

[seisman]

Here is a full script to reproduce the issue:

import numpy as np
import pygmt
from pygmt.helpers.testing import load_static_earth_relief

grid = load_static_earth_relief()
grid[3:6, 3:5] = np.nan
grid[6:8, 2:4] = np.inf
result = pygmt.grdfill(grid=grid, mode="c20", region=[-54, -50, -23, -16])

Ping @meghanrjones or @PaulWessel

[PaulWessel]

Unless Conda and the python ecosystem now fully supports M1 architecture then I am unable to debug this here. @seisman, do you know if there has been progress on that front?

dlopen(/Users/pwessel/UH/RESEARCH/CVSPROJECTS/GMTdev/gmt-dev/xbuild/src/Debug/libgmt.dylib, 0x0006): tried: '/Users/pwessel/UH/RESEARCH/CVSPROJECTS/GMTdev/gmt-dev/xbuild/src/Debug/libgmt.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'x86_64')), '/usr/local/lib/libgmt.dylib' (no such file), '/usr/lib/libgmt.dylib' (no such file), '/Users/pwessel/UH/RESEARCH/CVSPROJECTS/GMTdev/gmt-dev/xbuild/src/Debug/libgmt.6.4.0.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'x86_64')), '/usr/local/lib/libgmt.6.4.0.dylib' (no such file), '/usr/lib/libgmt.6.4.0.dylib' (no such file)

[maxrjones]

I debugged this yesterday and have a fix implemented (sorry for not posting an update promptly). I am still working on a regression test. I'll open a PR on Monday.