Segmentation fault while doing DFS on a coloured graph - c++

Question

Here is my Code:
#include <bits/stdc++.h>
using namespace std;
#define limt 101
//edge stores the edges for each node
//since it is defined outside main, i don't think it is the cause for
//segmentation fault
vector<int> edge[limt][limt];
int vist[limt], tgt;
//dfs call for a node numbered <i>u</i>
//and for only the edges colored <i>c</i>
int dfs(int u, int c)
{
if (u == tgt) return 1;
vist[u] = 1;
for (int i = 0; i < edge[u][c].size(); ++i) {
if (dfs(edge[u][c][i], c)) return 1;
}
vist[u] = 0;
return 0;
}
int main()
{
//n : number of nodes
//m: number of edges
//c: color of edge
int n, m;
int u, v, c;
cin >> n >> m;
for (int i = 0; i < m; ++i) {
cin >> u >> v >> c;
edge[u][c].push_back(v);
edge[v][c].push_back(u);
}
//q: number of queries/test-cases
int q, x, y;
cin >> q;
while (q--) {
cin >> x >> y;
memset(vist, 0, sizeof(vist));
int ans = 0;
vist[x] = 1;
tgt = y;
//program crashes for second test-case after reaching here
//i - the color of edge
//j - all the edges from x of color i
for (int i = 0; i < limt; ++i) {
for (int j = 0; j < edge[x][i].size(); ++j) {
ans += dfs(edge[x][i][j], i);
}
}
vist[x] = 0;
cout << ans << endl;
}
return 0;
}
The input i am giving is:
4 5
1 2 1
1 2 2
2 3 1
2 3 3
2 4 3
3
1 2
3 4
1 4
Here q is the number of test cases, the first case runs fine and gives the output 2 but on the second case program crashes with segmentation fault. I don't know what is the bug in it. I am sure i am not causing any overflow as the declarations are outside main().

Answer 1

In the second case, the dfs function causes infinite recursion.
The first call is dfs(u=2, c=1); which calls dfs(u=1, c=1); which calls dfs(u=2, c=1); and so on ad nauseum.
If you run your code in a debugger you should be able to see what happened by outputting the stack trace at the point of the segmentation fault.
Here is a picture of the graph generated by your data (where the colours are 1=red, 2=green, 3=blue):
Your loop inside main() starts at 3 and follows Red to find 2, then calls dfs(2, red); . (This was already a bit strange as you could have just called dfs(3, red) to get the same result). But then the dfs function just follows the first Red edge it finds so it ends up shuttling back and forth between 2 and 1.
To avoid this case you will need to add a way for dfs to know not to revisit a node it already visited. In fact you keep track of vist already but do not check it in the dfs function.
I think that just adding if ( vist[u] ) return 0; to the start of that function will fix this problem for this graph. However I think you should reset all of vist to 0 for each colour, not just for each testcase. For example in this graph if there were a blue edge from 2 to 4 but not from 3 to 4, the search would never find it because 2 was already marked visited during the red search.