#pragma GCC optimize("Ofast,no-stack-protector")
#pragma GCC optimize("O3")
#pragma GCC optimize("fast-math")
#pragma GCC target("avx2")
#pragma GCC target("sse,sse2,sse3,ssse3,sse4,popcnt,abm,mmx,avx")
#pragma GCC optimize("unroll-loops")
#pragma GCC target("popcnt,lzcnt,abm,bmi,bmi2")
#include <bits/stdc++.h>
#define fr first
#define sc second
#define all(a) (a).begin(), (a).end()
#define unique(a) a.resize(unique(a.begin(), a.end()) - a.begin())
#define popcnt(x) __builtin_popcountll(x)
using namespace std;
#ifdef ONPC
mt19937 rnd(223);
mt19937 rndll(223);
#else
mt19937 rnd(chrono::high_resolution_clock::now()
.time_since_epoch().count());
mt19937 rndll(223);
#endif
#define TIME (clock() * 1.0 / CLOCKS_PER_SEC)
using ll = long long;
using ld = double;
const int maxn = 2e4 + 100, inf = 1e9 + 100;
namespace segtree {
// This implementation is disgusting, but it seems to work and do it faster than previous version.
template<typename Item>
Item tree_merge(const Item& a, const Item& b) {
Item i;
i.update(a, b);
return i;
}
template<typename Item, bool lazy>
struct Pusher {};
template<typename Item>
struct Pusher<Item, false> {
void push(const vector<Item>&, int, int, int) {}
Item ask_on_segment(const vector<Item>& tree, int n, int l, int r) {
l |= n;
r |= n;
Item resl, resr;
while (l <= r) {
if (l & 1) {
resl = tree_merge(resl, tree[l]);
++l;
}
if (!(r & 1)) {
resr = tree_merge(tree[r], resr);
--r;
}
l >>= 1;
r >>= 1;
}
return tree_merge(resl, resr);
}
void push_point(const vector<Item>&, int, int) {}
template<typename P>
int lower_bound(const vector<Item>& tree, int n, int l, P p) {
Item cur;
if (p(cur)) return l - 1;
l |= n;
int r = n | (n - 1);
// carefully go up
while (true) {
if (p(tree_merge(cur, tree[l]))) {
break;
}
if (l == r) return n;
if (l & 1) {
cur = tree_merge(cur, tree[l]);
++l;
}
l >>= 1;
r >>= 1;
}
// usual descent from l
while (l < n) {
if (p(tree_merge(cur, tree[l * 2]))) {
l = l * 2;
} else {
cur = tree_merge(cur, tree[l * 2]);
l = l * 2 + 1;
}
}
return (l ^ n);
}
template<typename P>
int lower_bound_rev(const vector<Item>& tree, int n, int r, P p) {
Item cur;
if (p(cur)) return r + 1;
r |= n;
int l = n;
// carefully go up
while (true) {
if (p(tree_merge(tree[r], cur))) {
break;
}
if (l == r) return -1;
if (!(r & 1)) {
cur = tree_merge(tree[r], cur);
--r;
}
l >>= 1;
r >>= 1;
}
// usual descent from r
while (r < n) {
if (p(tree_merge(tree[r * 2 + 1], cur))) {
r = r * 2 + 1;
} else {
cur = tree_merge(tree[r * 2 + 1], cur);
r = r * 2;
}
}
return (r ^ n);
}
};
template<typename Item>
struct Pusher<Item, true> {
void push(vector<Item>& tree, int ind, int l, int r) {
tree[ind].push(tree[ind * 2], tree[ind * 2 + 1], l, r);
}
Item ask_on_segment(vector<Item>& tree, int n, int l, int r) {
int vl = 0, vr = n - 1;
int i = 1;
Item result;
while (vl != vr) {
int m = (vl + vr) / 2;
if (l > m) {
push(tree, i, vl, vr);
i = i * 2 + 1;
vl = m + 1;
} else if (r <= m) {
push(tree, i, vl, vr);
i = i * 2;
vr = m;
} else {
break;
}
}
if (l == vl && r == vr) {
return tree[i];
}
push(tree, i, vl, vr);
// left
{
int ind = i * 2;
int L = vl, R = (vl + vr) / 2;
while (l != L) {
int m = (L + R) / 2;
push(tree, ind, L, R);
if (l <= m) {
result = tree_merge(tree[ind * 2 + 1], result);
ind *= 2;
R = m;
} else {
ind = ind * 2 + 1;
L = m + 1;
}
}
result = tree_merge(tree[ind], result);
}
// right
{
int ind = i * 2 + 1;
int L = (vl + vr) / 2 + 1, R = vr;
while (r != R) {
int m = (L + R) / 2;
push(tree, ind, L, R);
if (r > m) {
result = tree_merge(result, tree[ind * 2]);
ind = ind * 2 + 1;
L = m + 1;
} else {
ind = ind * 2;
R = m;
}
}
result = tree_merge(result, tree[ind]);
}
return result;
}
void push_point(vector<Item>& tree, int n, int ind) {
int l = 0, r = n - 1;
int i = 1;
while (l != r) {
push(tree, i, l, r);
int m = (l + r) / 2;
if (ind <= m) {
r = m;
i *= 2;
} else {
l = m + 1;
i = i * 2 + 1;
}
}
}
template<typename P>
pair<int, Item> _lower_bound(vector<Item>& tree, int l, P p, Item cur, int i, int vl, int vr) {
if (vl == vr) {
if (p(tree_merge(cur, tree[i]))) {
return {vl, tree[i]};
} else {
return {vl + 1, tree[i]};
}
}
push(tree, i, vl, vr);
int m = (vl + vr) / 2;
if (l > m) {
return _lower_bound(tree, l, p, cur, i * 2 + 1, m + 1, vr);
} else if (l <= vl) {
if (!p(tree_merge(cur, tree[i]))) {
return {vr + 1, tree_merge(cur, tree[i])};
}
if (p(tree_merge(cur, tree[i * 2]))) {
return _lower_bound(tree, l, p, cur, i * 2, vl, m);
} else {
return _lower_bound(tree, l, p, tree_merge(cur, tree[i * 2]), i * 2 + 1, m + 1, vr);
}
} else {
auto [ind, it] = _lower_bound(tree, l, p, cur, i * 2, vl, m);
if (ind <= m) return {ind, it};
return _lower_bound(tree, l, p, it, i * 2 + 1, m + 1, vr);
}
}
template<typename P>
int lower_bound(vector<Item>& tree, int n, int l, P p) {
Item cur;
if (p(cur)) return l - 1;
return _lower_bound(tree, l, p, cur, 1, 0, n - 1).first;
}
template<typename P>
pair<int, Item> _lower_bound_rev(vector<Item>& tree, int r, P p, Item cur, int i, int vl, int vr) {
if (vl == vr) {
if (p(tree_merge(tree[i], cur))) {
return {vl, tree[i]};
} else {
return {vl - 1, tree[i]};
}
}
push(tree, i, vl, vr);
int m = (vl + vr) / 2;
if (r <= m) {
return _lower_bound_rev(tree, r, p, cur, i * 2, vl, m);
} else if (r >= vr) {
if (!p(tree_merge(tree[i], cur))) {
return {vl - 1, tree_merge(cur, tree[i])};
}
if (p(tree_merge(tree[i * 2 + 1], cur))) {
return _lower_bound_rev(tree, r, p, cur, i * 2 + 1, m + 1, vr);
} else {
return _lower_bound_rev(tree, r, p, tree_merge(tree[i * 2 + 1], cur), i * 2, vl, m);
}
} else {
auto [ind, it] = _lower_bound_rev(tree, r, p, cur, i * 2 + 1, m + 1, vr);
if (ind > m) return {ind, it};
return _lower_bound_rev(tree, r, p, it, i * 2, vl, m);
}
}
template<typename P>
int lower_bound_rev(vector<Item>& tree, int n, int r, P p) {
Item cur;
if (p(cur)) return r + 1;
return _lower_bound_rev(tree, r, p, cur, 1, 0, n - 1).first;
}
};
template<typename Item, bool lazy = false>
struct Segtree {
vector<Item> tree;
Pusher<Item, lazy> pusher;
int n;
int n0;
Segtree(int n = 0) {
build(n);
}
template<typename U>
Segtree(const vector<U>& v) {
build(v);
}
void build(int n) {
this->n0 = n;
while (n & (n - 1)) ++n;
this->n = n;
tree.assign(n * 2, {});
}
template<typename U>
void build(const vector<U>& v) {
build(v.size());
for (int i = 0; i < v.size(); ++i) {
tree[n | i].init(v[i], i);
}
build();
}
void build() {
for (int i = n - 1; i >= 1; --i) {
tree[i].update(tree[i * 2], tree[i * 2 + 1]);
}
}
void push(int ind, int l, int r) {
pusher.push(tree, ind, l, r);
}
template<typename T>
void set(int ind, const T& t) {
pusher.push_point(tree, n, ind);
ind |= n;
tree[ind].init(t, ind ^ n);
ind >>= 1;
while (ind) {
tree[ind].update(tree[ind * 2], tree[ind * 2 + 1]);
ind >>= 1;
}
}
template<typename T>
void update(int ind, const T& t) {
pusher.push_point(tree, n, ind);
ind |= n;
tree[ind].update(t, ind ^ n);
ind >>= 1;
while (ind) {
tree[ind].update(tree[ind * 2], tree[ind * 2 + 1]);
ind >>= 1;
}
}
Item& ith(int ind) {
static_assert(!lazy, "don't use this method with lazy propagation, unless you're sure you need it");
return tree[ind | n];
}
const Item& root() const {
return tree[1];
}
Item ask(int l, int r) {
l = max(l, 0);
r = min(r, n - 1);
if (l > r) return {};
return pusher.ask_on_segment(tree, n, l, r);
}
template<typename T>
void modify(int l, int r, const T& t) {
static_assert(lazy, "lazy must be set to true to use this function");
l = max(l, 0);
r = min(r, n - 1);
if (l > r) return;
int vl = 0, vr = n - 1;
int i = 1;
while (vl != vr) {
int m = (vl + vr) / 2;
if (l > m) {
push(i, vl, vr);
i = i * 2 + 1;
vl = m + 1;
} else if (r <= m) {
push(i, vl, vr);
i = i * 2;
vr = m;
} else {
break;
}
}
if (l == vl && r == vr) {
tree[i].modify(t, l, r);
} else {
push(i, vl, vr);
// left
{
int ind = i * 2;
int L = vl, R = (vl + vr) / 2;
while (l != L) {
int m = (L + R) / 2;
push(ind, L, R);
if (l <= m) {
tree[ind * 2 + 1].modify(t, m + 1, R);
ind *= 2;
R = m;
} else {
ind = ind * 2 + 1;
L = m + 1;
}
}
tree[ind].modify(t, L, R);
ind >>= 1;
while (ind != i) {
tree[ind].update(tree[ind * 2], tree[ind * 2 + 1]);
ind >>= 1;
}
}
// right
{
int ind = i * 2 + 1;
int L = (vl + vr) / 2 + 1, R = vr;
while (r != R) {
int m = (L + R) / 2;
push(ind, L, R);
if (r > m) {
tree[ind * 2].modify(t, L, m);
ind = ind * 2 + 1;
L = m + 1;
} else {
ind = ind * 2;
R = m;
}
}
tree[ind].modify(t, L, R);
ind >>= 1;
while (ind != i) {
tree[ind].update(tree[ind * 2], tree[ind * 2 + 1]);
ind >>= 1;
}
}
tree[i].update(tree[i * 2], tree[i * 2 + 1]);
}
i >>= 1;
while (i) {
tree[i].update(tree[i * 2], tree[i * 2 + 1]);
i >>= 1;
}
}
// first index r such that p(tree.ask(l, r)) == true
// if p() is true for empty item, return l-1
// if p() is never true, returns n
template<typename P>
int lower_bound(int l, P p) {
l = max(l, 0);
if (l >= n0) return n0;
return min(n0, pusher.lower_bound(tree, n, l, p));
}
// similarly to lower_bound, returns first (largest) l such that p(tree.ask(l, r)) == true
template<typename P>
int lower_bound_rev(int r, P p) {
r = min(r, n0 - 1);
if (r < 0) return -1;
return pusher.lower_bound_rev(tree, n, r, p);
}
};
}
using segtree::Segtree;
#define uint oleg
using uint = uint64_t;
using Mat = array<uint, 64>;
const uint one = 1;
inline bool mult(uint x, uint y) {
return popcnt(x & y) & 1;
}
struct Li {
Mat a{};
uint b{};
void upd(int i, int j, bool w) {
if (w)
a[j] ^= (one << i);
}
uint apply(uint x) const {
uint w = 0;
for (int i = 0; i < 64; i++)
if (mult(x, a[i]))
w ^= (one << i);
return w ^ b;
}
};
Li def() {
Li q{};
for (int i = 0; i < 64; i++)
q.upd(i, i, 1);
return q;
}
Mat c{};
Li comb(Li const &x, Li const &y) {
c = {};
Li q{};
for (int i = 0; i < 64; i++)
for (int j = 0; j < 64; j++)
if (x.a[j] & (one << i))
c[i] ^= (one << j);
for (int i = 0; i < 64; i++)
for (int j = 0; j < 64; j++)
q.upd(i, j, mult(c[i], y.a[j]));
q.b = y.apply(x.b);
return q;
}
Li a[maxn];
#define cin if(1)cin
Li read_hash() {
int m = 64;
cin >> m;
Li q;
for (int it = 0; it < m; it++) {
int s = it, o = rnd() & 1;
uint A = rndll();
cin >> s >> o >> A;
if (o == 0) {
// or
for (int i = 0; i < 64; i++)
if (A & (one << i)) {
q.b ^= (one << i);
}
A = ~A;
}
{
// and
for (int i = 0; i < 64; i++) {
int pe = (i + s) % 64;
q.upd(i, pe, (A >> pe) & 1);
}
}
}
uint b;
cin >> b;
q.b ^= b;
// for (int j = 0; j < 64; j++)
// cerr << q.a[j] << '\n';
// cerr << q.b << '\n';
// cerr << "-----------------\n";
return q;
}
struct Item {
Li w = def();
template<typename T>
void init(const T& t, int ind) {
w = t;
}
void update(const Item& a, const Item& b) {
w = comb(a.w, b.w);
}
//// similar to init, but more convenient for doing a[i] += x, implement only if needed
// template<typename T>
// void update(const T& t, int ind) {}
//// apply here, save for children
// template<typename T>
// void modify(const T& m, int l, int r) {}
// void push(Item& a, Item& b, int l, int r) {
// int m = (l + r) / 2;
// a.modify(mod, l, m);
// b.modify(mod, m + 1, r);
// // reset mod
// }
};
void solve() {
int n, zaps, C;
n = zaps = 20000;
C = 1;
cin >> n >> zaps >> C;
vector<Li> a(n);
for (int i = 0; i < n; i++) {
a[i] = read_hash();
}
cerr << "\n\nreading " << TIME << endl;
Segtree<Item> t(a);
cerr << "\n\nbuilding " << TIME << endl;
while (zaps--) {
int tp;
tp = rnd() & 1;
// tp = 0;
cin >> tp;
if (tp == 0) {
int l, r;
l = rnd() % n + 1;
r = rnd() % n + 1;
if (l > r)
swap(l, r);
uint x = rndll();
cin >> l >> r >> x;
l--;
r--;
// Li z = def();
// for (int i = l; i <= r; i++)
// z = comb(z, a[i]);
x = t.ask(l, r).w.apply(x);
cout << x << '\n';
} else {
int i;
i = rnd() % n + 1;
cin >> i;
i--;
a[i] = read_hash();
t.set(i, a[i]);
}
}
}
int main() {
#ifdef ONPC
freopen("../a.in", "r", stdin);
freopen("../a.out", "w", stdout);
#endif
ios::sync_with_stdio(0);
cin.tie(0);
cout << fixed;
cout.precision(20);
solve();
cerr << "\n\nConsumed " << TIME << endl;
}