Reduce the minimum compile time a bit

[Ogeon]

Reduces the minimum compile time a bit by reducing the feature set of syn, making more dependencies optional and changing to dependency resolver version 2. I got inspired by #285 to at least make the build time somewhat better. This reduces it by about 30% on my slow laptop (from ~15 seconds to ~10 seconds), but your mileage may vary and it also depends on what other dependencies are in the mix.

Breaking Change

• Increases MSRV to 1.60.0 to get the ? syntax in cargo features. Required for making approx optional.
• approx is now optional, but still enabled by default.
• The crate will no longer find itself by default if it's renamed in Cargo.tom. It requires the find-crate feature.

Slow Laptop Benchmarks

Slow laptop type    Microsoft Surface Pro 7 i5
Slow OS setup       Ubuntu in WSL2 in Windows 11
Processor           Intel(R) Core(TM) i5-1035G4 CPU @ 1.10GHz   1.50 GHz
Installed RAM       8,00 GB (7,60 GB usable)
System type         64-bit operating system, x64-based processor

rustc 1.62.0 (a8314ef7d 2022-06-27)

Commands

Debug mode:

cargo clean
cargo check --timings --no-default-features --features=std -p palette

Release mode:

cargo clean
cargo check --timings --no-default-features --features=std -p palette --release

Before

Debug mode	Release mode

After

Debug mode	Release mode

[github-actions]

Benchmark for d5591c5

Click to view benchmark

Test	Base	PR	%
Cie family/lab to lch	3.1±0.23µs	3.0±0.23µs	-3.23%
Cie family/lab to xyz	788.1±55.90ns	799.6±49.45ns	+1.46%
Cie family/lch to lab	2.5±0.17µs	2.5±0.16µs	0.00%
Cie family/linsrgb to xyz	4.0±0.27µs	3.9±0.30µs	-2.50%
Cie family/linsrgb to xyz - wide::f32x8	789.6±52.35ns	1263.8±81.30ns	+60.06%
Cie family/xyz to lab	18.7±1.62µs	19.8±1.41µs	+5.88%
Cie family/xyz to yxy	615.0±31.74ns	578.1±39.15ns	-6.00%
Cie family/yxy to xyz	603.3±37.38ns	603.5±45.17ns	+0.03%
Matrix functions/matrix_inverse	11.1±0.88ns	11.6±0.85ns	+4.50%
Matrix functions/multiply_3x3	10.6±0.71ns	10.7±0.47ns	+0.94%
Matrix functions/multiply_rgb_to_xyz	5.7±0.36ns	5.9±0.39ns	+3.51%
Matrix functions/multiply_xyz	5.9±0.92ns	5.9±0.35ns	0.00%
Matrix functions/multiply_xyz_to_rgb	5.7±0.34ns	5.5±0.36ns	-3.51%
Matrix functions/rgb_to_xyz_matrix	22.3±1.79ns	23.1±1.35ns	+3.59%
Rgb family/hsl to hsv	668.9±52.27ns	731.6±34.56ns	+9.37%
Rgb family/hsl to linear hsl	14.4±1.02µs	14.2±0.98µs	-1.39%
Rgb family/hsl to rgb	2.3±0.12µs	2.4±0.17µs	+4.35%
Rgb family/hsv to hsl	1053.1±52.17ns	1113.8±68.67ns	+5.76%
Rgb family/hsv to hwb	222.8±17.22ns	210.5±10.56ns	-5.52%
Rgb family/hsv to linear hsv	12.5±0.87µs	12.2±0.83µs	-2.40%
Rgb family/hsv to rgb	2.2±0.29µs	2.3±0.16µs	+4.55%
Rgb family/hwb to hsv	554.1±29.91ns	559.2±26.30ns	+0.92%
Rgb family/hwb to linear hwb	12.4±0.75µs	12.8±0.93µs	+3.23%
Rgb family/linear hsl to hsl	12.3±0.70µs	12.2±0.83µs	-0.81%
Rgb family/linear hsv to hsv	10.8±0.68µs	11.3±0.87µs	+4.63%
Rgb family/linear hwb to hwb	11.7±0.71µs	11.4±0.83µs	-2.56%
Rgb family/linsrgb to rgb	6.3±0.39µs	6.1±0.36µs	-3.17%
Rgb family/linsrgb to rgb - wide::f32x8	8.1±0.56µs	8.4±0.38µs	+3.70%
Rgb family/linsrgb_f32 to rgb_u8	943.5±55.75ns	907.2±59.08ns	-3.85%
Rgb family/linsrgb_f64 to rgb_u8	1090.2±66.88ns	1153.5±62.96ns	+5.81%
Rgb family/rgb to hsl	933.1±51.86ns	918.3±70.38ns	-1.59%
Rgb family/rgb to hsl - wide::f32x8	600.2±25.30ns	596.6±43.03ns	-0.60%
Rgb family/rgb to hsv	599.5±39.07ns	625.6±34.43ns	+4.35%
Rgb family/rgb to hsv - wide::f32x8	562.7±31.15ns	566.9±44.38ns	+0.75%
Rgb family/rgb to linsrgb	8.8±0.55µs	8.8±0.86µs	0.00%
Rgb family/rgb to linsrgb - wide::f32x8	7.0±0.56µs	6.1±0.43µs	-12.86%
Rgb family/rgb_u8 to linsrgb_f32	315.6±17.69ns	322.3±28.50ns	+2.12%
Rgb family/rgb_u8 to linsrgb_f64	302.2±20.93ns	334.8±17.51ns	+10.79%
Rgb family/xyz to linsrgb	7.6±0.47µs	7.4±0.53µs	-2.63%
Rgb family/xyz to linsrgb - wide::f32x8	1098.4±55.21ns	1035.6±57.50ns	-5.72%

[Ogeon]

bors r+

[bors]

Build succeeded:

• ci