diff --git a/codeGeneration/src/prog8/compiler/target/cpu6502/codegen/assignment/AugmentableAssignmentAsmGen.kt b/codeGeneration/src/prog8/compiler/target/cpu6502/codegen/assignment/AugmentableAssignmentAsmGen.kt
index c51f332b9..1564a233d 100644
--- a/codeGeneration/src/prog8/compiler/target/cpu6502/codegen/assignment/AugmentableAssignmentAsmGen.kt
+++ b/codeGeneration/src/prog8/compiler/target/cpu6502/codegen/assignment/AugmentableAssignmentAsmGen.kt
@@ -364,6 +364,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
 
     private fun inplaceModification_byte_value_to_pointer(pointervar: IdentifierReference, operator: String, value: Expression) {
         asmgen.assignExpressionToVariable(value, "P8ZP_SCRATCH_B1", DataType.UBYTE, null)
+        val sourceName = asmgen.loadByteFromPointerIntoA(pointervar)
         when (operator) {
             // note: ** (power) operator requires floats.
             "+" -> asmgen.out("  clc |  adc  P8ZP_SCRATCH_B1")
@@ -394,8 +395,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             "^", "xor" -> asmgen.out(" eor  P8ZP_SCRATCH_B1")
             else -> throw AssemblyError("invalid operator for in-place modification $operator")
         }
-        // TODO THIS IS WRONG:????
-        val sourceName = asmgen.loadByteFromPointerIntoA(pointervar)
         asmgen.storeAIntoZpPointerVar(sourceName)
     }
 
diff --git a/docs/source/todo.rst b/docs/source/todo.rst
index 8f90aede8..5dbf3feb2 100644
--- a/docs/source/todo.rst
+++ b/docs/source/todo.rst
@@ -3,7 +3,7 @@ TODO
 
 For next compiler release (7.5)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-check correctness of inplaceModification_byte_value_to_pointer()
+- allow cx16.rX as loop variable in forloops (now requires unqualified identifiername)
 ...
 
 
diff --git a/examples/test.p8 b/examples/test.p8
index a47227daf..bca37af23 100644
--- a/examples/test.p8
+++ b/examples/test.p8
@@ -3,11 +3,27 @@
 
 main {
     sub start() {
-        word llw = 300
-        cx16.r0s = 9 * 2 * 10 * llw
-        cx16.r1s = llw * 9 * 2 * 10
-        cx16.r3s = llw / 30 / 3
-        cx16.r4s = llw / 2 * 10
-        cx16.r5s = llw * 90 / 5     ; not optimized because of loss of integer division precision
+        ubyte @zp xx
+        for xx in 0 to 10 {
+            txt.print_ub(xx)
+            txt.spc()
+        }
+        txt.nl()
+
+        for cx16.r0L in 0 to 10 {
+            txt.print_ub(cx16.r0L)
+            txt.spc()
+        }
+        txt.nl()
+
+        for main.derp.xx in 0 to 10 {
+            txt.print_ub(main.derp.xx)
+            txt.spc()
+        }
+        txt.nl()
+    }
+
+    sub derp() {
+        ubyte xx
     }
 }