Possible problem with gadgets

[alexloboda]

Hi,

There is one more test case Nodes:

0 100
1 -50
2 100
3 -1000
4 -1
5 -1
6 -1000
7 100
8 -5
9 100

Edges:

0 1
1 2
0 3
1 3
2 3
4 3
5 3
0 4
2 5
4 7
4 6
5 6
5 9
7 6
8 6
9 6
7 8
8 9

The point in that heinz give two different answers with/without preprocessing. But seems like that problem more global. It is a graph with 2 tricomponents. And all 3 rooted solutions match in each tricomponent. But answer contains only part of these solutions. That's why these tricomponent can't be replaced by gadget.

Possibly that exists another test case when preprocessing can't transform source graph to solvable by solver without preprocessing.