Fix performance regression with BlockArrays

[tkf]

Fix #42 by updating to JuliaFolds/Transducers.jl#417

[codecov]

Codecov Report

Merging #49 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master      #49   +/-   ##
=======================================
  Coverage   96.01%   96.01%           
=======================================
  Files           4        4           
  Lines         427      427           
=======================================
  Hits          410      410           
  Misses         17       17           

Flag	Coverage Δ
#unittests	96.01% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c4aa6db...fdbcbea. Read the comment docs.

[tkf]

ID	time ratio	memory ratio
["parallel_triangular_sum", "sequential"]	0.20 (5%) ✅	0.89 (1%) ✅
["parallel_triangular_sum", "threaded"]	0.21 (5%) ✅	0.89 (1%) ✅
["parallel_triangular_sum", "threaded_nest", ":nestlevel => 1"]	0.22 (5%) ✅	0.89 (1%) ✅
["parallel_triangular_sum", "threaded_nest", ":nestlevel => 2"]	0.26 (5%) ✅	0.90 (1%) ✅
["parallel_triangular_sum", "threaded_nest", ":nestlevel => 3"]	0.33 (5%) ✅	0.90 (1%) ✅
["sum", ":label => \"BlockVector\"", ":impl => \"floop\""]	0.04 (5%) ✅	0.00 (1%) ✅
["sum", ":label => \"filter\"", ":impl => \"floop\""]	0.90 (5%) ✅	1.00 (1%)
["sum", ":label => \"flatten\"", ":impl => \"for\""]	1.05 (5%) ❌	1.00 (1%)

https://github.com/JuliaFolds/FLoops-data/blob/benchmark-results/2020/09/06/063739/result.md