#pragma GCC optimize ("Ofast")
#include<bits/stdc++.h>
using namespace std;
void*wmem;
char memarr[96000000];
template<class T> inline void walloc1d(T **arr, int x, void **mem = &wmem){
  static int skip[16] = {0, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1};
  (*mem) = (void*)( ((char*)(*mem)) + skip[((unsigned long long)(*mem)) & 15] );
  (*arr)=(T*)(*mem);
  (*mem)=((*arr)+x);
}
inline int my_getchar_unlocked(){
  static char buf[1048576];
  static int s = 1048576;
  static int e = 1048576;
  if(s == e && e == 1048576){
    e = fread_unlocked(buf, 1, 1048576, stdin);
    s = 0;
  }
  if(s == e){
    return EOF;
  }
  return buf[s++];
}
inline void rd(int &x){
  int k;
  int m=0;
  x=0;
  for(;;){
    k = my_getchar_unlocked();
    if(k=='-'){
      m=1;
      break;
    }
    if('0'<=k&&k<='9'){
      x=k-'0';
      break;
    }
  }
  for(;;){
    k = my_getchar_unlocked();
    if(k<'0'||k>'9'){
      break;
    }
    x=x*10+k-'0';
  }
  if(m){
    x=-x;
  }
}
struct MY_WRITER{
  char buf[1048576];
  int s;
  int e;
  MY_WRITER(){
    s = 0;
    e = 1048576;
  }
  ~MY_WRITER(){
    if(s){
      fwrite_unlocked(buf, 1, s, stdout);
    }
  }
}
;
MY_WRITER MY_WRITER_VAR;
void my_putchar_unlocked(int a){
  if(MY_WRITER_VAR.s == MY_WRITER_VAR.e){
    fwrite_unlocked(MY_WRITER_VAR.buf, 1, MY_WRITER_VAR.s, stdout);
    MY_WRITER_VAR.s = 0;
  }
  MY_WRITER_VAR.buf[MY_WRITER_VAR.s++] = a;
}
inline void wt_L(char a){
  my_putchar_unlocked(a);
}
inline void wt_L(int x){
  int s=0;
  int m=0;
  char f[10];
  if(x<0){
    m=1;
    x=-x;
  }
  while(x){
    f[s++]=x%10;
    x/=10;
  }
  if(!s){
    f[s++]=0;
  }
  if(m){
    my_putchar_unlocked('-');
  }
  while(s--){
    my_putchar_unlocked(f[s]+'0');
  }
}
template<class S> inline void arrInsert(const int k, int &sz, S a[], const S aval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  a[k] = aval;
}
template<class S, class T> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  a[k] = aval;
  b[k] = bval;
}
template<class S, class T, class U> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval, U c[], const U cval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  for(i=sz-1;i>k;i--){
    c[i] = c[i-1];
  }
  a[k] = aval;
  b[k] = bval;
  c[k] = cval;
}
template<class S, class T, class U, class V> inline void arrInsert(const int k, int &sz, S a[], const S aval, T b[], const T bval, U c[], const U cval, V d[], const V dval){
  int i;
  sz++;
  for(i=sz-1;i>k;i--){
    a[i] = a[i-1];
  }
  for(i=sz-1;i>k;i--){
    b[i] = b[i-1];
  }
  for(i=sz-1;i>k;i--){
    c[i] = c[i-1];
  }
  for(i=sz-1;i>k;i--){
    d[i] = d[i-1];
  }
  a[k] = aval;
  b[k] = bval;
  c[k] = cval;
  d[k] = dval;
}
struct graph{
  int N;
  int*es;
  int**edge;
  void setEdge(int N__, int M, int A[], int B[], void **mem = &wmem){
    int i;
    N = N__;
    walloc1d(&es, N, mem);
    walloc1d(&edge, N, mem);
    for(i=(0);i<(N);i++){
      es[i] = 0;
    }
    for(i=(0);i<(M);i++){
      es[A[i]]++;
      es[B[i]]++;
    }
    for(i=(0);i<(N);i++){
      walloc1d(&edge[i], es[i], mem);
    }
    for(i=(0);i<(N);i++){
      es[i] = 0;
    }
    for(i=(0);i<(M);i++){
      edge[A[i]][es[A[i]]++] = B[i];
      edge[B[i]][es[B[i]]++] = A[i];
    }
  }
  void getDist(int root, int res[], void *mem = wmem){
    int i;
    int j;
    int k;
    int*q;
    int s;
    int z;
    walloc1d(&q, N, &mem);
    for(i=(0);i<(N);i++){
      res[i]=-1;
    }
    res[root]=0;
    s=0;
    z=1;
    q[0]=root;
    while(z){
      i=q[s++];
      z--;
      for(j=(0);j<(es[i]);j++){
        k=edge[i][j];
        if(res[k]>=0){
          continue;
        }
        res[k]=res[i]+1;
        q[s+z++]=k;
      }
    }
  }
  int getDist(int a, int b, void *mem = wmem){
    int i;
    int j;
    int k;
    int*q;
    int s;
    int z;
    int*d;
    if(a==b){
      return 0;
    }
    walloc1d(&d, N, &mem);
    walloc1d(&q, N, &mem);
    for(i=(0);i<(N);i++){
      d[i] = -1;
    }
    d[a] = 0;
    s = 0;
    z = 1;
    q[0] = a;
    while(z){
      i = q[s++];
      z--;
      for(j=(0);j<(es[i]);j++){
        k = edge[i][j];
        if(d[k] >= 0){
          continue;
        }
        d[k] = d[i] + 1;
        if(k==b){
          return d[k];
        }
        q[s+z++] = k;
      }
    }
    return -1;
  }
}
;
int N;
int M;
int A;
int B;
int L[300000];
int R[300000];
graph g;
int dis[300000];
int main(){
  int i;
  wmem = memarr;
  int st;
  int ed;
  int res;
  rd(N);N += (1);
  rd(M);
  rd(A);A += (-1);
  rd(B);
  {
    int Lj4PdHRW;
    for(Lj4PdHRW=(0);Lj4PdHRW<(M);Lj4PdHRW++){
      rd(L[Lj4PdHRW]);L[Lj4PdHRW] += (-1);
      rd(R[Lj4PdHRW]);
    }
  }
  st = N++;
  ed = N++;
  for(i=(0);i<(A+1);i++){
    arrInsert(M,M,L,st,R,i);
  }
  for(i=(B);i<(N-2);i++){
    arrInsert(M,M,L,i,R,ed);
  }
  g.setEdge(N,M,L,R);
  res = g.getDist(st, ed) - 2;
  if(res<0){
    wt_L(-1);
    wt_L('\n');
  }
  else{
    wt_L(res);
    wt_L('\n');
  }
  return 0;
}
// cLay varsion 20200926-1

// --- original code ---
// int N, M, A, B, L[3d5], R[3d5];
// graph g;
// int dis[3d5];
// {
//   int st, ed, res;
//   rd(N++,M,A--,B,(L--,R)(M));
//   st = N++;
//   ed = N++;
//   rep(i,A+1) arrInsert(M,M,L,st,R,i);
//   rep(i,B,N-2) arrInsert(M,M,L,i,R,ed);
//   g.setEdge(N,M,L,R);
//   res = g.getDist(st, ed) - 2;
//   wt(if[res<0, -1, res]);
// }