Strip line info from default exprs

[jkrumbiegel]

Description

Fixes #2926

[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	39.74s (39.39, 40.11) 0.26+-	18.18s (17.93, 18.68) 0.25+-	16.50s (16.21, 16.94) 0.25+-	11.84ms (11.54, 11.96) 0.14+-	138.66ms (135.02, 142.18) 2.86+-
master	39.68s (39.17, 40.10) 0.31+-	18.07s (17.85, 18.44) 0.21+-	16.54s (16.34, 16.85) 0.21+-	11.84ms (11.58, 12.11) 0.18+-	137.88ms (135.84, 138.72) 1.06+-
evaluation	+0.14%, 0.06s invariant (0.19d, 0.72p, 0.29std)	+0.58%, 0.11s invariant (0.45d, 0.41p, 0.23std)	-0.26%, -0.04s invariant (-0.19d, 0.74p, 0.23std)	+0.07%, 0.01ms invariant (0.05d, 0.93p, 0.16std)	+0.56%, 0.78ms invariant (0.36d, 0.52p, 1.96std)
CairoMakie	33.50s (33.40, 33.56) 0.06+-	18.25s (18.08, 18.36) 0.10+-	2.68s (2.65, 2.73) 0.03+-	10.45ms (10.30, 10.62) 0.12+-	4.84ms (4.67, 4.99) 0.13+-
master	33.67s (33.56, 33.76) 0.09+-	18.28s (18.10, 18.42) 0.11+-	2.68s (2.65, 2.73) 0.02+-	10.50ms (10.34, 10.74) 0.13+-	5.07ms (4.89, 5.26) 0.14+-
evaluation	-0.51%, -0.17s faster ✓ (-2.27d, 0.00p, 0.07std)	-0.19%, -0.03s invariant (-0.31d, 0.57p, 0.11std)	-0.06%, -0.0s invariant (-0.06d, 0.92p, 0.03std)	-0.53%, -0.06ms invariant (-0.44d, 0.42p, 0.13std)	-4.72%, -0.23ms faster ✓ (-1.73d, 0.01p, 0.13std)
WGLMakie	45.33s (44.82, 46.09) 0.42+-	22.47s (22.06, 22.76) 0.26+-	22.76s (22.28, 23.20) 0.33+-	14.25ms (13.84, 15.15) 0.52+-	802.19ms (760.73, 835.00) 21.90+-
master	45.02s (44.55, 45.28) 0.26+-	22.63s (22.29, 22.91) 0.24+-	22.93s (22.58, 23.29) 0.26+-	14.44ms (13.81, 15.29) 0.50+-	800.17ms (763.13, 819.63) 18.05+-
evaluation	+0.69%, 0.31s invariant (0.90d, 0.12p, 0.34std)	-0.71%, -0.16s invariant (-0.63d, 0.26p, 0.25std)	-0.72%, -0.16s invariant (-0.54d, 0.33p, 0.30std)	-1.32%, -0.19ms invariant (-0.37d, 0.51p, 0.51std)	+0.25%, 2.01ms invariant (0.10d, 0.85p, 19.97std)

[SimonDanisch]

Could it work to do this without MacroTools, by just having:

striplines(@nospecialize(expr)) = expr
function striplines(expr::Expr)
    args = [striplines(elem) for elem in expr.args if !(elem isa LineNumberNode)]
    return Expr(expr.head, args...)
end

For a simple test, this seems to work fine!

[jkrumbiegel]

yeah sure, was just to lazy to copy paste and thought MacroTools is always an indirect dep anyway

[SimonDanisch]

Ok, fair enough, maybe we want to use MacroTools for our macros more anyways!