Fix invalid range logscale

[SimonDanisch]

Continuation of #1671
Throws proper error for:

lines(1:10,1:10, axis=(; limits=(1,1,1,5))

[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	14.35s (14.22, 14.55) 0.11+-	1.27s (1.25, 1.33) 0.03+-	878.60ms (852.83, 924.19) 24.23+-	12.83ms (12.72, 13.07) 0.12+-	112.15ms (110.58, 113.54) 1.16+-
master	14.39s (14.18, 14.70) 0.17+-	1.32s (1.31, 1.34) 0.01+-	862.41ms (850.98, 875.66) 8.47+-	12.77ms (12.72, 12.84) 0.04+-	111.08ms (109.79, 112.46) 0.96+-
evaluation	-0.23%, -0.03s invariant (-0.24d, 0.67p, 0.14std)	-3.70%, -0.05s faster ✓ (-2.28d, 0.00p, 0.02std)	+1.84%, 16.2ms invariant (0.89d, 0.14p, 16.35std)	+0.44%, 0.06ms invariant (0.61d, 0.29p, 0.08std)	+0.95%, 1.07ms invariant (1.00d, 0.09p, 1.06std)
CairoMakie	11.01s (10.84, 11.37) 0.19+-	1.10s (1.08, 1.11) 0.01+-	232.64ms (226.46, 242.48) 5.96+-	10.96ms (10.70, 11.38) 0.22+-	6.58ms (6.42, 6.75) 0.11+-
master	11.08s (10.92, 11.49) 0.20+-	1.12s (1.10, 1.15) 0.02+-	221.50ms (214.63, 228.32) 5.71+-	10.99ms (10.78, 11.35) 0.20+-	6.64ms (6.50, 6.78) 0.11+-
evaluation	-0.68%, -0.08s invariant (-0.38d, 0.49p, 0.20std)	-1.82%, -0.02s invariant (-1.19d, 0.05p, 0.02std)	+4.79%, 11.14ms slower X (1.91d, 0.00p, 5.84std)	-0.30%, -0.03ms invariant (-0.16d, 0.77p, 0.21std)	-0.94%, -0.06ms invariant (-0.56d, 0.32p, 0.11std)
WGLMakie	15.94s (15.58, 16.36) 0.26+-	1.46s (1.41, 1.53) 0.04+-	14.29s (14.02, 14.57) 0.21+-	19.38ms (16.49, 28.49) 4.14+-	1.44s (1.40, 1.48) 0.03+-
master	16.10s (15.71, 16.51) 0.26+-	1.49s (1.42, 1.57) 0.06+-	14.53s (14.28, 14.84) 0.22+-	17.49ms (16.98, 18.40) 0.57+-	1.42s (1.38, 1.47) 0.03+-
evaluation	-1.03%, -0.16s invariant (-0.62d, 0.27p, 0.26std)	-1.89%, -0.03s invariant (-0.52d, 0.35p, 0.05std)	-1.66%, -0.24s invariant (-1.08d, 0.07p, 0.22std)	+9.75%, 1.89ms noisy🤷‍♀️ (0.64d, 0.28p, 2.35std)	+1.44%, 0.02s invariant (0.66d, 0.24p, 0.03std)

[SimonDanisch]

Ah, this still fails (should fail), since it changes the limits for plots with only one point, e.g. text(Point2f(0,0); text="hi")

[SimonDanisch]

Ok, i think this is the wrong approach!
Plot limits from e.g. text(Point2f(0,0); text="hi") already get expanded...
The question is, should we error in the limit setting code for dmin==dmax, or just also expand it, like its done for a plot with only one point?
Will close this PR for now, since a fix will look pretty different!