chore: remove unsafe code.

This may be causing a heap corruption, due to `rule_id` being out of range. By removing the unsafe code we can debug the issue better.

lib/src/scanner/context.rs

@@ -278,18 +278,13 @@ impl ScanContext<'_> {
     /// increased by the time elapsed since `rule_execution_start_time`.
     #[cfg(feature = "rules-profiling")]
     pub(crate) fn update_time_spent_in_rule(&mut self, rule_id: RuleId) {
-        // SAFETY: it's safe to call `get_unchecked_mut`, the size of the
-        // `time_spent_in_rule` vector is guaranteed to be the number of
-        // existing rules. Therefore, the rule ID can be used as an index
-        // in this vector.
-        unsafe {
-            self.time_spent_in_rule
-                .get_unchecked_mut::<usize>(rule_id.into())
-                .add_assign(self.clock.delta_as_nanos(
-                    self.rule_execution_start_time,
-                    self.clock.raw(),
-                ));
-        }
+        self.time_spent_in_rule
+            .get_mut::<usize>(rule_id.into())
+            .unwrap()
+            .add_assign(self.clock.delta_as_nanos(
+                self.rule_execution_start_time,
+                self.clock.raw(),
+            ));
     }
 
     /// Called during the scan process when a rule didn't match.