REL: yt 4.0.2 meta issue

[neutrinoceros]

Here's a list of existing pull requests that need back porting for yt's next bug fix release
Other devs, please feel free to edit this list if you find it's missing anything
Yt 4.0.1 was released on July 21, so I'm going to add every user-facing bug fix that was merged since then

• fix a bug with DatasetSeries instances created via __new__ #3238
• BUG: fix a bug in AMRVAC frontend w/ relative parfiles #3343
• Checking if extrema are all None in PhasePlot; fixes #3431 #3432
• BUG: add support for setting up derived fields from lambdas #3440
• Fix for 2D AMReX datasets in which 3 values are in the amr.n_cell entry in the job_info file. #3445
• BUG: fix a broken error message for yt.load in case of FileNotFoundError #3446
• ENH: raise a clear error message when calling unsupported plot callbacks #3453
• Make sure we add the Parameters group to hvals if it's present #3455
• fix combining FITS images #3456
• Changed _project_coords to create new array instead of modifying in-place #3463
• Add support for global_startindex != 0 to covering_grid #3468
• Don't get domain dimensions from cparam file for Boxlib frontend. #3470
• DOC: use absolute links instead of relative path for quickstart guard linking with README.md #3474 (this one is docs, but affects the PyPI page)
• BUG: fix mpl data path handling for Windows #3480
• BUG: fix a broken call to warnings.warn #3485
• [bugfix] Get order of axes right for FITSParticleProjection #3488
• BUG: graciously fallback to a default repr for yt arrays in jupyter notebooks #3517
• BUG: Fix re-allocation of octree #3524
• BUG: fix incorrect default axis labels in 'r' oriented slice plots for spherical data #3532
• BUG: fix incorrect default limits for slice plots along curvilinear directions (phi, theta) #3533
• BUG: fix colorbar minor ticks drawing for symlog norms #3536
• RAMSES: Allow to read non-standard info files #3544
• BUG: remove broken logging messages in lens module #3548 + BUG: cleanup a broken debug logging message #3569
• BUG: implement AthenaDataset._is_valid #3424
• ENH: loosen header matching for Athena vtk files to allow compatibility with single variable 'vtk outputs' #3490
• BUG: fix error message in lookup_on_disk_data #3465
• BUG: fix two bugs in yt.visualization.eps_writer #3496
• ⭐ BUG: fix default math font in yt plots against matplotlib >= 3.4 #3520
• ⭐ BUG: Set the display name for new fields #3282
• BLD: Fix compilation on FreeBSD #3605
• BLD: Fix build on i386 #3606
• BUG: override image suffix with matplotlib argument in case of conflict #3607
• BUG: fix a bug where 2D plots in spherical geometries would always use code-units in plot axes #3618
• [gdf] dimensionality doesn't imply which dimension is reduced #3630
• MNT: migrate away from packaging.version.parse #3494
• BUG: fix set_zlim method with symlog norm #3538
• BUG: fix two bugs with symlog colorbar ticks for matplotlib 3.5 #3556
• BUG: fix for matplotlib 3.5 + symlog + linthresh value outside data range #3565
• ⭐ BUG: fix LaTeX typesetting for fractional ylabels in LinePlot #3637
• BUG: fix an off-by-one error in ColorTransferFunction.to_map_colormap method #3661

candidates for back porting (currently not merged)

• BUG: fixes for Hammer-Aitoff projection (r-normal) 2D viz in spherical geometries #3628
• BUG: fix figure layout in case the long axis is vertical #3633

Notes:
We can strike an item here as soon has it has a backport PR, even unmerged.

items marked with "⭐" contain answer-store updates. These will likely be harder to back port unless we're cautious to reproduce the order in which they were merged.
It would be a good idea to try to reproduce the merging order for the whole list, though it may be hard to do it exactly.

[matthewturk]

3410 is a bit of a bigger deal than bugfix, isn't it?

[neutrinoceros]

IMHO it's not, but feel free to remove it if you're not confortable with it. The bug it resolved was probably not ever encountered in production anyway.

[neutrinoceros]

⚠️ PRs requiring manual backports:
https://github.com/yt-project/yt/issues?q=label%3A%22Still+Needs+Manual+Backport%22+is%3Aclosed

[neutrinoceros]

@matthewturk on second thought I think #3410 is not worth the risk of backporting it manually, so I'm taking it off the list.

[neutrinoceros]

We got to a state where PRs can be back ported immediately after they are merged, which is much easier to manage, so I don't think it's worth maintaining this meta issue anymore. The release isn't ready yet but there is no reason left to keep this open.