Global earth relief data loaded as xarray doesn't work if projection center is not 0

Description of the problem

It may work in GMT 6.0.0 + PyGMT v0.1.2, or not. Currently, PyGMT master + GMT 6.1.0 gives me a wrong result.

Full code that generated the error

import pygmt
from pygmt.datasets import load_earth_relief

grid = load_earth_relief()
fig = pygmt.Figure()
fig.grdimage(grid, cmap="geo", region='d', projection="H10c")

fig.shift_origin(xshift="12c")
fig.grdimage(grid, cmap="geo", region='g', projection="H10c")

fig.shift_origin(xshift="-12c", yshift="6c")
fig.grdimage(grid, cmap="geo", region='d', projection="H120/10c")

# this one sometimes crashes
#fig.shift_origin(xshift="12c")
#fig.grdimage(grid, cmap="geo", region='g', projection="H120/10c")
fig.savefig("map.png")

Output

[seisman]

The issue is caused by upstream GMT bugs, which was fixed in a series of upstream PRs (https://github.com/GenericMappingTools/gmt/pull/3813, https://github.com/GenericMappingTools/gmt/pull/3813, https://github.com/GenericMappingTools/gmt/pull/3829) and already merged into GMT’s master and 6.1 branches.

GMT may release v6.1.1 this month. When GMT v6.1.1 is released, I think we can bump the minimum required GMT version to v6.1.1 and release PyGMT v0.2.0.

---

HELP NEEDED

PyGMT users who build GMT from source codes may try GMT’s 6.1 branch (for the upcoming v6.1.1 release) and see if you can find any related bugs before GMT v6.1.1 is released.

[PaulWessel]

OK, so -JH120/10.c does not to any rotation and messes up at -180 + 120 = 60W it seems.