CI fixes

[penelopeysm]

Description

• Updates tests to account for v0.16.5 of Latexify.jl.
• Removes Julia 1.6 tests on ARM macOS and readds one for Intel macOS, as described in option 2 of https://discourse.julialang.org/t/how-to-fix-github-actions-ci-failures-with-julia-1-6-or-1-7-on-macos-latest-and-macos-14/117019

Attribution

• I am listed in .zenodo.json (see How to get acknowledged #3503)

Things to consider

• Does it work on log scales?
• Does it work in layouts?
• Does it work in recipes?
• Does it work with multiple series in one call?
• PR includes or updates tests?
• PR includes or updates documentation?

[codecov]

Codecov Report

Attention: Patch coverage is 66.66667% with 1 line in your changes missing coverage. Please review.

Project coverage is 89.75%. Comparing base (fa65e7d) to head (3873ecc).
Report is 23 commits behind head on master.

Files	Patch %	Lines
src/backends/plotlybase.jl	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4974      +/-   ##
==========================================
- Coverage   89.77%   89.75%   -0.02%     
==========================================
  Files          40       40              
  Lines        8780     8817      +37     
==========================================
+ Hits         7882     7914      +32     
- Misses        898      903       +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[penelopeysm]

The ubuntu/macOS + Julia 1.6 tests are failing because of image comparison regression tests, specifically on image number 6 with GR backend.

Looking across a bunch of logs, it seems that any runs with GR.jl v0.73.x cause this test to fail, whereas v0.72.x works.

I've extracted the images from the GitHub runner, for good measure:

ubuntu-latest, 1.6

Difference: 0.003957361095933401 tolerance: 0.0005

macOS-13, 1.6

Difference: 0.003957361095933401 tolerance: 0.001

Reference image

Taken from https://github.com/JuliaPlots/PlotReferenceImages.jl/blob/master/Plots/gr/1.36.4/ref006.png

[penelopeysm]

@BeastyBlacksmith, pinging you because you sorted my other PR, I hope that is ok:)

I think this might be the extent to which I feel confident in fixing stuff. The version clash on GR (comment above, Column 2 below) is a bit bizarre; I've tried to track down why it resolves to 0.72.8 sometimes and 0.73.7 other times, but without any success.

If we stick a GR = 0.73 compat entry (Column 3), it will successfully resolve to 0.73.7 on all combinations of OS/Julia versions and we can update the image regression test, but that will obviously break backwards compatibility. We also can't just manually install [email protected] while running tests because of this error.

The Windows + Julia 1.6 tests currently time out because of Kaleido. It could be made to run by bumping the compat on PlotlyKaleido to 1,2. However, that in turn causes weird fluctuations in the resolved version of GR (Column 4).

(If we update compat for both deps (Column 5), and also update the image regression test, then I think all the tests should pass.)

OS/Julia version	With current compat bounds	With GR=0.73	With PlotlyKaleido=1,2	With both
1.6, Ubuntu	0.73.7	0.73.7	0.73.7	0.73.7
1.6, Windows	hangs	hangs	0.73.7	0.73.7
1.6, macOS-13	0.73.7	0.73.7	0.73.7	0.73.7
1.7, Ubuntu	0.72.8	0.73.7	0.73.7	0.73.7
1.8, Ubuntu	0.72.8	0.73.7	0.72.8	0.73.7
1.9, Ubuntu	0.72.8	0.73.7	0.72.8	0.73.7
1.10, Ubuntu	0.72.8	0.73.7	0.73.7	0.73.7
1.10, Windows	0.72.8	0.73.7	0.73.7	0.73.7
1.10, macOS	0.72.8	0.73.7	0.73.7	0.73.7
pre, Ubuntu	0.72.8	0.73.7	0.72.8	0.73.7

[BeastyBlacksmith]

Thank you for taking a look into this.

On GR: given that the reference clearly has artifacts that got fixed it seems reasonable to me to set the compat to 0.73 and update the refrence image.

On PlotlyKaleido: This ones a bit more work. Since, to be compatible with PlotlyKaleido v2 we need to check in the savefig methods for all outputs that use kaleido (prob. everything except for html) if the kaleido server is running and start it before calling the PlotlyKaleido savefig function.
See JuliaPlots/PlotlyKaleido.jl#6

[BeastyBlacksmith]

@penelopeysm would you be willing to do a v2 port of this?

[penelopeysm]

Happy to do that! 🙂