Fix camera observable caching and cleanup

[ffreyer]

Description

Fixes #3639 by

• detaching the cached scaled resolution from plot to avoid plot based cleanup
• adding a dublication observable to avoid robj based cleanup

Changes to camera matrix caching

The calculation of cached matrices depends on existing matrices in scene.camera which are unrelated to the plot or renderobject, and on plot.space which may change dynamically. Therefore cached camera matrices

• must use a static space
• must not get deleted when a plot or renderobject is deleted
• must be created dynamically based on the plots space

TODO:

• re-enable camera matrix caching
• check WGLMakie
• add tests

Type of change

Delete options that do not apply:

• Bug fix (non-breaking change which fixes an issue)

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(fig)

	using	create	display	create	display
GLMakie	3.33s (3.30, 3.36) 0.02+-	380.82ms (378.25, 382.63) 1.72+-	470.67ms (463.43, 505.85) 15.53+-	6.96ms (6.86, 7.02) 0.07+-	25.25ms (25.16, 25.34) 0.07+-
master	3.32s (3.30, 3.35) 0.02+-	392.59ms (389.81, 395.16) 2.01+-	465.76ms (461.49, 469.78) 2.97+-	7.04ms (6.92, 7.10) 0.06+-	25.22ms (25.09, 25.30) 0.07+-
evaluation	1.00x invariant, 0.01s (0.46d, 0.41p, 0.02std)	1.03x faster ✓, -11.77ms (-6.30d, 0.00p, 1.86std)	0.99x invariant, 4.92ms (0.44d, 0.44p, 9.25std)	1.01x faster ✓, -0.08ms (-1.18d, 0.05p, 0.06std)	1.00x invariant, 0.04ms (0.55d, 0.32p, 0.07std)
CairoMakie	3.12s (2.99, 3.22) 0.08+-	337.51ms (320.83, 350.21) 10.32+-	150.37ms (141.71, 157.99) 5.98+-	7.65ms (7.16, 8.04) 0.27+-	621.99μs (618.13, 626.55) 3.00+-
master	3.15s (3.09, 3.19) 0.03+-	339.63ms (331.35, 361.43) 10.10+-	153.36ms (142.81, 176.00) 10.84+-	7.76ms (7.48, 7.90) 0.14+-	619.28μs (617.63, 621.03) 1.13+-
evaluation	1.01x invariant, -0.03s (-0.55d, 0.34p, 0.06std)	1.01x invariant, -2.11ms (-0.21d, 0.71p, 10.21std)	1.02x invariant, -2.99ms (-0.34d, 0.54p, 8.41std)	1.01x invariant, -0.11ms (-0.50d, 0.38p, 0.21std)	1.00x invariant, 2.71μs (1.19d, 0.06p, 2.06std)
WGLMakie	3.83s (3.80, 3.87) 0.03+-	336.52ms (331.28, 341.39) 4.17+-	9.42s (9.25, 9.62) 0.15+-	9.71ms (9.48, 9.84) 0.16+-	69.25ms (67.98, 70.80) 0.97+-
master	3.83s (3.80, 3.86) 0.03+-	340.03ms (327.89, 347.06) 7.30+-	9.38s (9.27, 9.50) 0.07+-	10.07ms (9.55, 12.13) 0.91+-	69.08ms (68.48, 70.13) 0.58+-
evaluation	1.00x invariant, 0.0s (0.10d, 0.85p, 0.03std)	1.01x invariant, -3.51ms (-0.59d, 0.30p, 5.73std)	1.00x invariant, 0.04s (0.36d, 0.52p, 0.11std)	1.04x invariant, -0.36ms (-0.55d, 0.34p, 0.54std)	1.00x invariant, 0.18ms (0.22d, 0.69p, 0.78std)

[ffreyer]

Hmm, I'm getting more allocations with this than master... Was this actually useful when you added it @SimonDanisch?