next/prevfloat(::BFloat16,::Integer) with tests

[milankl]

fixes #48

[codecov]

Codecov Report

Attention: 3 lines in your changes are missing coverage. Please review.

Comparison is base (630428a) 52.17% compared to head (e10709c) 65.41%.
Report is 7 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##           master      #49       +/-   ##
===========================================
+ Coverage   52.17%   65.41%   +13.23%     
===========================================
  Files           3        3               
  Lines         138      133        -5     
===========================================
+ Hits           72       87       +15     
+ Misses         66       46       -20     

Files	Coverage Δ
src/bfloat16.jl	72.26% <91.89%> (+4.97%)	⬆️

... and 2 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[maleadt]

A recursive implementation like that isn't great, I think it's better to copy what Base has:

julia> @benchmark nextfloat(BFloat16(0), 9999)
BenchmarkTools.Trial: 10000 samples with 1 evaluation.
 Range (min … max):  89.875 μs … 114.750 μs  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     90.083 μs               ┊ GC (median):    0.00%
 Time  (mean ± σ):   90.405 μs ±   1.361 μs  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▂█▄▄▆▅▃▁                   ▁▂  ▁                             ▁
  ████████▆▅▃▃▁▃▁▁▁▁▁▁▃▁▁▁▁▁▇██▅▇█▆▃▅▅▅▄▄▃▄▄▃▃▁▃▁▁▅▁▃▄▁▁▃▃▁▁▆▆ █
  89.9 μs       Histogram: log(frequency) by time      96.4 μs <

 Memory estimate: 0 bytes, allocs estimate: 0.

julia> @benchmark nextfloat(BFloat16(0), 9999)
BenchmarkTools.Trial: 10000 samples with 1000 evaluations.
 Range (min … max):  1.167 ns … 15.167 ns  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     1.250 ns              ┊ GC (median):    0.00%
 Time  (mean ± σ):   1.259 ns ±  0.273 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

                     ▂                  █                  ▂ ▁
  ▃▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█ █
  1.17 ns      Histogram: log(frequency) by time     1.29 ns <

 Memory estimate: 0 bytes, allocs estimate: 0.

[milankl]

Awesome! Thanks Tim!