Allow Cairo scatter marker conversion to be overridden

[asinghvi17]

Description

This PR allows external packages like MakieTeX to override Makie.to_spritemarker for Cairo and vector formats, while still exploiting Makie.to_spritemarker to convert to images.

The PR also bumps the CairoMakie version to prepare for a release.

Type of change

Delete options that do not apply:

• New feature (non-breaking change which adds functionality)

Checklist

• Added an entry in CHANGELOG.md (for new features and breaking changes)

[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	3.46s (3.37, 3.57) 0.06+-	173.75ms (166.55, 182.42) 6.59+-	482.15ms (469.33, 501.17) 10.39+-	7.71ms (7.47, 8.17) 0.25+-	25.98ms (25.78, 26.16) 0.16+-
master	3.44s (3.38, 3.51) 0.05+-	171.76ms (165.55, 178.51) 4.46+-	475.49ms (467.76, 493.27) 8.51+-	7.52ms (7.33, 7.76) 0.13+-	25.94ms (25.78, 26.26) 0.16+-
evaluation	0.99x invariant, 0.02s (0.31d, 0.57p, 0.06std)	0.99x invariant, 1.99ms (0.35d, 0.52p, 5.52std)	0.99x invariant, 6.66ms (0.70d, 0.21p, 9.45std)	0.98x invariant, 0.19ms (0.93d, 0.12p, 0.19std)	1.00x invariant, 0.04ms (0.26d, 0.63p, 0.16std)
CairoMakie	3.16s (3.10, 3.26) 0.06+-	175.66ms (168.69, 182.88) 5.62+-	108.14ms (103.68, 112.52) 3.57+-	7.86ms (7.54, 8.05) 0.20+-	897.03μs (887.64, 913.81) 8.61+-
master	3.13s (3.09, 3.24) 0.07+-	175.90ms (170.65, 181.69) 4.41+-	106.78ms (103.80, 114.94) 4.20+-	7.61ms (7.47, 7.86) 0.13+-	910.34μs (892.48, 927.87) 16.25+-
evaluation	0.99x invariant, 0.03s (0.42d, 0.45p, 0.06std)	1.00x invariant, -0.24ms (-0.05d, 0.93p, 5.01std)	0.99x invariant, 1.36ms (0.35d, 0.53p, 3.88std)	0.97x slower X, 0.25ms (1.50d, 0.02p, 0.16std)	1.01x invariant, -13.32μs (-1.02d, 0.09p, 12.43std)
WGLMakie	3.79s (3.70, 3.83) 0.04+-	239.00ms (172.29, 401.58) 105.94+-	9.04s (8.90, 9.18) 0.10+-	10.52ms (9.78, 13.97) 1.53+-	69.26ms (68.26, 69.83) 0.55+-
master	3.75s (3.73, 3.79) 0.02+-	395.32ms (390.39, 411.66) 7.34+-	8.99s (8.88, 9.09) 0.08+-	9.83ms (9.66, 9.96) 0.13+-	69.35ms (68.43, 70.19) 0.56+-
evaluation	0.99x invariant, 0.04s (1.06d, 0.08p, 0.03std)	1.65x faster✅, -156.32ms (-2.08d, 0.01p, 56.64std)	0.99x invariant, 0.05s (0.54d, 0.34p, 0.09std)	0.93x noisy🤷‍♀️, 0.69ms (0.64d, 0.28p, 0.83std)	1.00x invariant, -0.09ms (-0.16d, 0.77p, 0.55std)

[asinghvi17]

Tested this with MakieTeX and the API works great. This also removes FFTW as a direct dependency of CairoMakie.

[jkrumbiegel]

Was this already reviewed?

[asinghvi17]

By Simon on Slack, yes