#include<bits/stdc++.h>
#define int long long
#define ff first
#define ss second
#define ve vector
#define pb push_back
#define en end
#define be begin
#pragma GGC optimize("O3,unroll-loops")
#pragma GGC target("avx2,bmi,bmi2,lzcnt,popcnt")
#define ppb pop_back
using namespace std;
bool comp(pair<float,int>a,pair<float,int>b){
if (a.ff>=b.ff) return 1;
else return 0;
}
int n;
long long m;
int32_t main(){
ios_base::sync_with_stdio(0);
cin.tie(0);
cin>>n>>m;
ve<pair<double,double>>r;
ve<pair<float,int>>v;
for (int i=0; i<n; i++){
int x,y;
cin>>x>>y;
double dx=x,dy=y;
float z=dx/dy;
// cout<<fixed<<" "<<setprecision(9)<<" "<<dx/dy<<endl;
r.pb({x,y});
v.pb({z,i});
}
ve<int>z;
sort(v.be(),v.en(),comp);
priority_queue<pair<int,int>>qu;
int sm=0;
int mx=0;
ve<int>a[n];
ve<int>b[n];
double mxx;
int t=0;
for (int i=0; i<n; i++){
double k=v[i].ff;
if (qu.size()>0){
while (sm*k>m) sm-=qu.top().ff,b[i].pb(qu.top().ss),qu.pop();
}
if ((sm+r[v[i].ss].ss)*k<=m) {sm+=r[v[i].ss].ss; qu.push({r[v[i].ss].ss,v[i].ss}); a[i].pb(v[i].ss);}
else if (qu.top().ff>r[v[i].ss].ss) {sm-=qu.top().ff; b[i].pb(qu.top().ss); qu.pop(); qu.push({r[v[i].ss].ss,v[i].ss}); sm+=r[v[i].ss].ss; a[i].pb(v[i].ss);}
if (mx<qu.size()) mxx=1e18;
if (mx<qu.size() || (mx==qu.size() && mxx<sm*k)){
t=i;
}
mxx=max(mxx,sm*k);
mx=max(mx,(int)qu.size());
// cout<<setprecision(9)<<fixed<<k<<" "<<sm<<endl;
}
ve<int>c(n,0);
for (int i=0; i<=t; i++){
for (auto g:a[i]){
c[g]=1;
//cout<<"a: "<<g<<endl;
}
for (auto g:b[i]){
c[g]=0;
// cout<<"b: "<<g<<endl;
}
}
cout<<mx<<endl;
//cout<<t<<endl;
for (int i=0; i<n; i++){
if (c[i]==1) cout<<i+1<<endl;
}
}
