Fix coredump with mul_I -> div_I ops

Fixes R#2280 rakudo/rakudo#2280

The issue in the ticket stems from -2147483648 being big enough to fit into
32bit int, yet its negation, 2147483648, being 1 bit too big. The assumption
in the small-int portion of the div_I op is that if we divide MVMint32 by
MVMint32, we'll always get an MVMint32-sized result, however if we divide
-2147483648 by -1—both of which are MVMint32-sized, we get 2147483648 that's
1 bit too big to fit into the result.

We fix that issue by using MVMint64 for the result along with casts to ensure
we upgrade to big-enough size soon enough.

Next, comes the question of why
`nqp::div_I(nqp::mul_I(1073741824, -2, Int), -1, Int)` crashes, yet if you
manually calculate `nqp::div_I(-2147483648, -1, Int)`, you no longer get the
crash. This is because in the former case, we start with two small-ints,
the mul_I op process them and since the -2147483648 result is big enough
to fit into a 32-bit int, it gives a small-int back.

In the latter case, however, we start off with a 2147483648—a bit larger than
a small-int, so it's a big-int—and feed it to the negation op, which simply sets
the sign of the big-int and gives us a big-int back. Now at div_I op, instead
of two small-ints, we have one big-int, so we go through the big-int path
and we never hit the small-int path that has a bug.

src/math/bigintops.c

@@ -647,7 +647,7 @@ MVMObject *MVM_bigint_div(MVMThreadContext *tc, MVMObject *result_type, MVMObjec
     } else {
         MVMint32 num   = ba->u.smallint.value;
         MVMint32 denom = bb->u.smallint.value;
-        MVMint32 value;
+        MVMint64 value;
         if ((cmp_a == MP_LT) ^ (cmp_b == MP_LT)) {
             if (denom == 0) {
                 MVM_exception_throw_adhoc(tc, "Division by zero");
@@ -658,7 +658,7 @@ MVMObject *MVM_bigint_div(MVMThreadContext *tc, MVMObject *result_type, MVMObjec
                 value = num / denom;
             }
         } else {
-            value = num / denom;
+            value = (MVMint64)num / denom;
         }
         store_int64_result(bc, value);
     }