New failures on Windows CI

[seisman]

All the tests passed on Windows two weeks ago (commit 6bb6378), but some failed recently, likely due to some updates of the Windows agent used in CI.

	162 - doc/scripts/GMT_splines.sh (Failed)
	1179 - test/sample1d/smooth.sh (Failed)
	1180 - test/sample1d/smooth_deriv.sh (Failed)

Here is the diff image for test/sample1d/smooth.sh. There are some differences in the red box:

[joa-quim]

Those tests do not fail for me locally.

[seisman]

There are also some new failures on macOS because GitHub Actions upgraded macos-latest from macos-12 to macos-14:

	 51 - doc/scripts/GMT_Defaults_1c.sh (Failed)
	240 - doc/examples/ex38/ex38.sh (Failed)
	356 - test/gmtbinstats/nonweighted.sh (Failed)
	357 - test/gmtbinstats/nonweightedrect.sh (Failed)
	358 - test/gmtbinstats/weighted.sh (Failed)
	359 - test/gmtbinstats/weightedrect.sh (Failed)
	386 - test/gmtmath/lsfit.sh (Failed)
	412 - test/gmtspatial/duplicate.sh (Failed)
	503 - test/grdfft/power2D.sh (Failed)
	520 - test/grdgradient/aspect.sh (Failed)
	526 - test/grdgradient/illum_var.sh (Failed)
	538 - test/grdimage/globalgrid.sh (Failed)
	545 - test/grdimage/grdread.sh (Failed)
	551 - test/grdimage/lonshift.sh (Failed)
	640 - test/grdvector/met_uv.sh (Failed)
	644 - test/grdvector/shrink.sh (Failed)
	681 - test/kml/gmt2kml.sh (Failed)
	692 - test/mapproject/oblmerc_up.sh (Failed)
	724 - test/movie/movie_indicator_f_ver.sh (Failed)
	750 - test/potential/flexure_e.sh (Failed)
	751 - test/potential/flexure_en.sh (Failed)
	843 - test/pscoast/oblsuite_N.sh (Failed)
	882 - test/pshistogram/histlegend.sh (Failed)
	887 - test/pshistogram/wrappedhisto.sh (Failed)
	1026 - test/psxy/clip_ellipse.sh (Failed)
	1110 - test/psxy/repeat1.sh (Failed)
	1179 - test/sample1d/smooth.sh (Failed)
	1182 - test/segy/segy_01.sh (Failed)
	1183 - test/segy/segy_02.sh (Failed)
	1204 - test/spectrum1d/power1D.sh (Failed)
	1257 - test/surface/periodic_pix.sh (Failed)

[joa-quim]

Can we do something here? I'm afraid not. The CI failures are full of mysteries, some non reproducible like this one on Windows. I think we should close this one.