math.big: fix Integer.bit_len() when there are no digits in the number

vlib/math/big/big_test.v

@@ -433,7 +433,7 @@ fn test_set_bit() {
 	assert a == b
 }
 
-fn test_bit_len() {
+fn test_bit_len() ? {
 	assert big.zero_int.bit_len() == 0
 	assert big.one_int.bit_len() == 1
 
@@ -441,7 +441,10 @@ fn test_bit_len() {
 
 	assert big.one_int.lshift(1239).bit_len() == 1240
 
-	assert big.integer_from_string('4338476092346017364013796407961305761039463198075691378460917856') or {
+	mut num := big.integer_from_string('4338476092346017364013796407961305761039463198075691378460917856')?
-		panic('Could not read from decimal')
+	assert num.bit_len() == 212
-	}.bit_len() == 212
+	for num.bit_len() > 0 {
+		num = num.rshift(8)
+		assert true
+	}
 }

vlib/math/big/integer.v

@@ -960,6 +960,8 @@ pub fn (x Integer) bit_len() int {
 	if x.signum == 0 {
 		return 0
 	}
-
+	if x.digits.len == 0 {
+		return 0
+	}
 	return x.digits.len * 32 - bits.leading_zeros_32(x.digits.last())
 }