Optimize a bit bfs and dfs iterators

[Tortar]

in particular, bfs is like 40% faster

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.31%. Comparing base (43f9f18) to head (7e5f2a8).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #381   +/-   ##
=======================================
  Coverage   97.31%   97.31%           
=======================================
  Files         120      120           
  Lines        6954     6956    +2     
=======================================
+ Hits         6767     6769    +2     
  Misses        187      187           

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

[simonschoelly]

Does that have any influence on performance? Because I feel like expliectly stating that we return nothing in that case makes it easier to understand what is going on.

[simonschoelly]

I see why dfs should always be faster, as there the only improvement there is using @inbound and the rest seems to be renaming. For bfs a statement like makes it 40% faster needs some explanation. At least on what kind of graphs - but better with the whole benchmark setup.

I am saying this, because we also had the case, that people just tested performance on graphs that did not really fit the problem. Therefore I think we should publish the whole methodology when testing for speedup

[Tortar]

closing in favour of #382