[PORT] [BUG] yen_k_shortest_paths always returns at least one path

[etiennedeg]

#1268
The bug was fixed in #1311 in the dev 2.0 branch of LightGraphs, but never made it to master. It was fixed by calling Experimental methods, which is problematic. We should probably revisit this issue, and maybe we can work to move some of the Experimental shortest path algorithms to the standard algorithms if they bring improvements.

[gdalle]

Fixed by #183

julia> g = cycle_digraph(10)
{10, 10} directed simple Int64 graph

julia> yen_k_shortest_paths(g, 2, 1, weights(g), 2, maxdist=2)
Graphs.YenState{Int64, Int64}(Int64[], Vector{Int64}[])