Move recursion to data_limits

[ffreyer]

Description

I thought about the changes in #3179 some more and came to the conclusion that moving the recursion to data_limits instead of point_iterator isn't something we need or should hold back. It's just a simple internal change that is pretty isolated from other bounding box and transformation changes.

This rolls back the data_limits -> point_iterator changes from #3179 and instead updates bbox recursion.

TODO:

• check errorbars, i.e. fix error bar range in recipes #3179
• check bracket
• check hexbin

Type of change

• Internal clean up

Checklist

• Added an entry in NEWS.md (for new features and breaking changes)
• Added or changed relevant sections in the documentation
• Added unit tests for new algorithms, conversion methods, etc.
• Added reference image tests for new plotting functions, recipes, visual options, etc.

[MakieBot]

Compile Times benchmark

Note, that these numbers may fluctuate on the CI servers, so take them with a grain of salt. All benchmark results are based on the mean time and negative percent mean faster than the base branch. Note, that GLMakie + WGLMakie run on an emulated GPU, so the runtime benchmark is much slower. Results are from running:

using_time = @ctime using Backend
# Compile time
create_time = @ctime fig = scatter(1:4; color=1:4, colormap=:turbo, markersize=20, visible=true)
display_time = @ctime Makie.colorbuffer(display(fig))
# Runtime
create_time = @benchmark fig = scatter(1:4; color=1:4, colormap=:turbo, markersize=20, visible=true)
display_time = @benchmark Makie.colorbuffer(display(fig))

	using	create	display	create	display
GLMakie	14.05s (13.93, 14.19) 0.10+-	1.35s (1.34, 1.37) 0.01+-	0.94s (0.91, 1.03) 0.04+-	12.49ms (12.35, 12.66) 0.13+-	166.83ms (164.43, 168.83) 1.84+-
master	14.11s (13.84, 14.45) 0.24+-	1.34s (1.28, 1.41) 0.04+-	927.82ms (909.46, 953.07) 13.55+-	12.62ms (12.46, 12.86) 0.15+-	167.32ms (164.52, 170.86) 2.32+-
evaluation	-0.43%, -0.06s invariant (-0.32d, 0.56p, 0.17std)	+1.10%, 0.01s invariant (0.48d, 0.40p, 0.03std)	+0.85%, 0.01s invariant (0.26d, 0.64p, 0.03std)	-1.04%, -0.13ms invariant (-0.94d, 0.11p, 0.14std)	-0.29%, -0.49ms invariant (-0.23d, 0.67p, 2.08std)
CairoMakie	11.39s (11.05, 11.99) 0.38+-	1.21s (1.14, 1.29) 0.06+-	234.22ms (218.01, 250.31) 11.60+-	11.53ms (11.19, 12.12) 0.36+-	6.93ms (6.59, 7.20) 0.22+-
master	11.31s (10.90, 11.73) 0.27+-	1.18s (1.13, 1.24) 0.04+-	225.93ms (214.73, 238.80) 9.46+-	11.31ms (10.94, 11.59) 0.27+-	6.88ms (6.50, 7.18) 0.22+-
evaluation	+0.71%, 0.08s invariant (0.24d, 0.66p, 0.33std)	+2.51%, 0.03s invariant (0.61d, 0.28p, 0.05std)	+3.54%, 8.29ms invariant (0.78d, 0.17p, 10.53std)	+1.93%, 0.22ms invariant (0.70d, 0.21p, 0.31std)	+0.72%, 0.05ms invariant (0.23d, 0.68p, 0.22std)
WGLMakie	14.89s (14.42, 15.33) 0.33+-	1.46s (1.42, 1.59) 0.06+-	13.18s (12.86, 13.61) 0.30+-	18.05ms (17.07, 19.17) 0.71+-	1.24s (1.18, 1.36) 0.08+-
master	14.88s (14.47, 15.46) 0.30+-	1.46s (1.41, 1.52) 0.05+-	13.07s (12.87, 13.36) 0.17+-	17.86ms (17.00, 18.89) 0.72+-	1.24s (1.17, 1.29) 0.04+-
evaluation	+0.09%, 0.01s invariant (0.04d, 0.94p, 0.32std)	+0.05%, 0.0s invariant (0.01d, 0.98p, 0.05std)	+0.86%, 0.11s invariant (0.46d, 0.41p, 0.23std)	+1.02%, 0.18ms invariant (0.26d, 0.64p, 0.71std)	-0.10%, -0.0s invariant (-0.02d, 0.97p, 0.06std)

[SimonDanisch]

This breaks latex hlines in axis: