dedicated functions for logging events adds even more code size on 32bit

Signed-off-by: xermicus <cyrill@parity.io>

crates/llvm-context/src/polkavm/context/mod.rs

@@ -1185,6 +1185,12 @@ where
     /// However, we still protect against this possibility here. Heap index
     /// offsets are generally untrusted and potentially represent valid
     /// (but wrong) pointers when truncated.
+    ///
+    /// TODO: Splitting up into a dedicated function
+    /// could potentially decrease code sizes (LLVM can still decide to inline).
+    /// However, passing i256 parameters is counter productive and
+    /// I've found that splitting it up actualy increases code size.
+    /// Should be reviewed after 64bit support.
     pub fn safe_truncate_int_to_xlen(
         &self,
         value: inkwell::values::IntValue<'ctx>,