Bump to GMT 6.2.0rc1

[seisman]

The first release candidate of GMT 6.2.0 has been released (https://github.com/GenericMappingTools/gmt/releases/tag/6.2.0rc1) today.

Now GMT 6.2.0rc1 is available in the devel channel on conda-forge (https://anaconda.org/conda-forge/gmt/) and can be installed using:

conda install -c conda-forge/label/dev gmt=6.2.0rc1

We need to:

• Bump the GMT version to 6.2.0rc1 used in the CI tests (Bump the GMT version to 6.2.0rc1 in CI #1218)
• Some tests fail and we need to fix them
• Add some new tests for upstream bugs that were fixed in 6.2.0rc1

[weiji14]

Bumping to GMT 6.2.0rc1 will also affect several existing PRs. I think we should go ahead with bumping the version as soon as possible though (so that we can test and fix any new bugs). Is there any PR we would like to merge/finish before bumping PyGMT to use GMT 6.2.0rc1? Otherwise it will get lost under a torrent of PRs relating to fixing broken unit tests 😆

I was actually thinking of updating https://github.com/GenericMappingTools/pygmt/blob/e4cb775346814dce46555a27c0c94f8e0aed6f04/.github/workflows/dvc-diff.yml to have a side-by-side before/after comparison of the images (instead of just having a single new image shown). Unsure if I have time this week to update that workflow (have a presentation and my PhD defence to prep for) but will see.

[seisman]

I was actually thinking of updating e4cb775/.github/workflows/dvc-diff.yml to have a side-by-side before/after comparison of the images (instead of just having a single new image shown).

I can work on this.

[seisman]

@GenericMappingTools/pygmt-maintainers please give #1219 and #1218 a fast review and approval, otherwise, we won't finish the migration before the next 6.2.0rc2 or 6.2.0 release.

[maxrjones]

@GenericMappingTools/pygmt-maintainers please give #1219 and #1218 a fast review and approval, otherwise, we won't finish the migration before the next 6.2.0rc2 or 6.2.0 release.

I know we haven't updated the documentation for the maintainer review process yet, but I think these are good examples of when to use the 'request a review' option for pinging pygmt-maintainers. At least for me, I usually periodically skim the PRs for any that have been marked 'ready for review' or 'final review call' but will notice specific review requests immediately.

[seisman]

I've merged #1218, so now the CI tests are running with GMT 6.2.0rc1, and ~100 tests fail, due to the modern theme changes introduced in GMT 6.2.

What we need to do to make the tests pass again:

1. Install (or upgrade) 6.2.0rc1 in your local environment using conda install -c conda-forge/label/dev gmt=6.2.0rc1
2. Run tests in a specific test file, e.g., pytest pygmt/tests/test_basemap.py
3. If there are any failing tests, you should be able to see the baseline.png and result.png images in the results directory.
4. Compare the baseline.png and result.png images. Do the differences make sense to you?
5. Generate new baseline images using pytest --mpl-generate-path=baseline pygmt/tests/test_basemap.py
6. Copy new baseline images of the failing tests to the pygmt/tests/baseline directory
7. Add the baseline images to DVC and push the images following the contributing guides.
8. Open PRs

[seisman]

Here is a list of test files that contain failing tests. Anyone is welcome to work on them.

• pygmt/helpers/testing.py (Fix pygmt/helpers/testing.py doctest for GMT 6.2.0rc1 #1243 @weiji14)
• pygmt/tests/test_basemap.py (Update baseline images for basemap #1221 @seisman)
• pygmt/tests/test_coast.py (Update fig.coast baseline images for GMT 6.2.0rc1 #1222 @seisman)
• pygmt/tests/test_colorbar.py (Update fig.colorbar baseline images for GMT 6.2.0rc1 #1229 @seisman)
• pygmt/tests/test_config.py (Update pygmt.config baseline images for GMT 6.2.0rc1 #1235 @weiji14)
• pygmt/tests/test_contour.py (Update fig.contour baseline images for GMT 6.2.0rc1 #1234 @seisman)
• pygmt/tests/test_figure.py (Update fig.shift_origin baseline images for GMT 6.2.0rc1 #1237 @seisman)
• pygmt/tests/test_grd2cpt.py (Update fig.grd baseline images for GMT 6.2.0rc1 #1224 @weiji14)
• pygmt/tests/test_grdcontour.py (Update fig.grd baseline images for GMT 6.2.0rc1 #1224 @weiji14)
• pygmt/tests/test_grdfilter.py (Update fig.grd baseline images for GMT 6.2.0rc1 #1224 @weiji14)
• pygmt/tests/test_inset.py (Update fig.inset baseline images for GMT 6.2.0rc1 #1230 @seisman)
• pygmt/tests/test_legend.py (Update fig.legend baseline images for GMT 6.2.0rc1 #1233 @seisman)
• pygmt/tests/test_logo.py (Update fig.logo baseline images for GMT 6.2.0rc1 #1223 @seisman)
• pygmt/tests/test_makecpt.py (Update makecpt baseline images for GMT 6.2.0rc1 #1231 @meghanrjones)
• pygmt/tests/test_meca.py (Update fig.meca baseline images for GMT 6.2.0rc1 #1225 @seisman)
• pygmt/tests/test_plot.py (Update fig.plot baseline images for GMT 6.2.0rc1 #1227 @seisman)
• pygmt/tests/test_plot3d.py (Update fig.plot3d baseline images for GMT 6.2.0rc1 #1228 @seisman)
• pygmt/tests/test_rose.py (Update fig.rose baseline images for GMT 6.2.0rc1 #1226 @michaelgrund)
• pygmt/tests/test_solar.py (Update fig.solar baseline images for GMT 6.2.0rc1 #1232 @seisman)

Not failing, but wrong according to #1217 (comment)

• pygmt/tests/test_subplot.py
• pygmt/tests/test_text.py
• pygmt/tests/test_wiggle.py

[weiji14]

I'll handle all the test_grd*.py tests in a single PR if that's ok.

[seisman]

I'll handle all the test_grd*.py tests in a single PR if that's ok.

OK to me.