Feature/optimize human printer

[arichardet]

Optimized human printer and added benchmarks to tests

Stats on improvements:

name                       old time/op    new time/op    delta
DrawNodeNoChanges             108µs ± 2%      62µs ± 8%  -42.47%        (p=0.000 n=20+19)
DrawNodeNoChangesFiltered    6.63µs ± 5%    6.55µs ±10%     ~           (p=0.091 n=19+20)
DrawNodeMetaFiltered         6.65µs ± 3%    6.42µs ± 2%   -3.49%        (p=0.000 n=19+20)
DrawNodeChanges               122µs ± 2%      59µs ± 2%  -51.69%        (p=0.000 n=18+20)
DrawNodeError                 124µs ±13%      59µs ± 1%  -52.43%        (p=0.000 n=19+20)

name                       old alloc/op   new alloc/op   delta
DrawNodeNoChanges            18.2kB ± 0%     6.2kB ± 0%  -66.09%        (p=0.000 n=20+20)
DrawNodeNoChangesFiltered    3.86kB ± 0%    3.86kB ± 0%     ~     (all samples are equal)
DrawNodeMetaFiltered         3.86kB ± 0%    3.86kB ± 0%     ~     (all samples are equal)
DrawNodeChanges              19.9kB ± 0%     6.5kB ± 0%  -67.52%        (p=0.000 n=20+20)
DrawNodeError                20.2kB ± 0%     6.5kB ± 0%  -68.03%        (p=0.000 n=20+20)

name                       old allocs/op  new allocs/op  delta
DrawNodeNoChanges               381 ± 0%       122 ± 0%  -67.98%        (p=0.000 n=20+20)
DrawNodeNoChangesFiltered      82.0 ± 0%      82.0 ± 0%     ~     (all samples are equal)
DrawNodeMetaFiltered           82.0 ± 0%      82.0 ± 0%     ~     (all samples are equal)
DrawNodeChanges                 417 ± 0%       123 ± 0%  -70.50%        (p=0.000 n=20+20)
DrawNodeError                   425 ± 0%       123 ± 0%  -71.06%        (p=0.000 n=20+20)

[BrianHicks]

is out ever used? It doesn't look like it, here.

[BrianHicks]

What happens if you cache the rendered templates instead of just the template functions?

[arichardet]

Addressed feedback and made changes to human printer tests.
Attempted to cache for further improvements; however, the changes were either negligible or worse.