The improves[1] on the original implementation by not requiring to store the traversed trie path in list.
P.S:
Oh Yes, I did run the tests and i think the program would be better commented.
[1]: This view comes with a bias from author ;p. I would love to see the benchmarks with the old one.
Coverage decreased (-0.7%) to 98.601% when pulling b003ad169d97de2b0f6e9e4f590c6a91555c8d81 on rmNULL:master into 5591859aabe3da37499a20d0d0d6dd77e480ed8d on vi3k6i5:master.
The improves[1] on the original implementation by not requiring to store the traversed trie path in list.
P.S: Oh Yes, I did run the tests and i think the program would be better commented. [1]: This view comes with a bias from author ;p. I would love to see the benchmarks with the old one.