diff --git a/codeGenCpu6502/src/prog8/codegen/cpu6502/AsmOptimizer.kt b/codeGenCpu6502/src/prog8/codegen/cpu6502/AsmOptimizer.kt
index c820e49ab..7aea2813c 100644
--- a/codeGenCpu6502/src/prog8/codegen/cpu6502/AsmOptimizer.kt
+++ b/codeGenCpu6502/src/prog8/codegen/cpu6502/AsmOptimizer.kt
@@ -583,8 +583,8 @@ private fun optimizeJsrRtsAndOtherCombinations(linesByFour: Sequence<List<Indexe
         fun sameLabel(branchInstr: String, jumpInstr: String, labelInstr: String): Boolean {
             if('(' in jumpInstr) return false       // indirect jump cannot be replaced
             val label = labelInstr.trimEnd().substringBefore(':').substringBefore(' ').substringBefore('\t')
-            println("label=$label")
-            return true
+            val branchLabel = branchInstr.trimStart().substring(3).trim()
+            return label==branchLabel
         }
 
         // beq Label + jmp Addr + Label  -> bne Addr
diff --git a/codeGenCpu6502/src/prog8/codegen/cpu6502/FunctionCallAsmGen.kt b/codeGenCpu6502/src/prog8/codegen/cpu6502/FunctionCallAsmGen.kt
index 869249fc5..00d1ed15f 100644
--- a/codeGenCpu6502/src/prog8/codegen/cpu6502/FunctionCallAsmGen.kt
+++ b/codeGenCpu6502/src/prog8/codegen/cpu6502/FunctionCallAsmGen.kt
@@ -82,7 +82,7 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
             is PtAddressOf -> false
             is PtIdentifier -> false
             is PtIrRegister -> false
-            is PtMemoryByte -> return usesOtherRegistersWhileEvaluating(arg.address)
+            is PtMemoryByte -> true     // TODO might not actually need extra registers if the value has to end up in A
             is PtNumber -> false
             is PtBool -> false
             else -> true
@@ -110,9 +110,9 @@ internal class FunctionCallAsmGen(private val program: PtProgram, private val as
                         throw AssemblyError("call argument evaluation problem: can't save cpu statusregister parameter ${call.position}")
                     }
                     else {
-                        if(usedX()) asmgen.saveRegisterStack(CpuRegister.X, false)
-                        if(usedY()) asmgen.saveRegisterStack(CpuRegister.Y, false)
-                        if(usedA()) asmgen.saveRegisterStack(CpuRegister.A, false)
+                        if(usedX()) asmgen.saveRegisterStack(CpuRegister.X, usedA())
+                        if(usedY()) asmgen.saveRegisterStack(CpuRegister.Y, usedA())
+                        if(usedA()) asmgen.saveRegisterStack(CpuRegister.A, usedA())
                         val used = argumentViaRegister(sub, IndexedValue(it, param.second), arg)
                         if(usedA()) asmgen.restoreRegisterStack(CpuRegister.A, false)
                         if(usedY()) asmgen.restoreRegisterStack(CpuRegister.Y, true)
diff --git a/docs/source/todo.rst b/docs/source/todo.rst
index b9cdb647a..fbb36e2b8 100644
--- a/docs/source/todo.rst
+++ b/docs/source/todo.rst
@@ -1,9 +1,7 @@
 TODO
 ====
 
-https://github.com/irmen/prog8/issues/136 (string.find register order issue)
-
-other issues on github.
+See open issues on github.
 
 optimize signed byte/word division by powers of 2 (and shift right?), it's now using divmod routine.  (also % ?)
     see inplacemodificationByteVariableWithLiteralval() and inplacemodificationSomeWordWithLiteralval()
@@ -30,6 +28,12 @@ optimize signed byte/word division by powers of 2 (and shift right?), it's now u
          lda	$4
 
 
+Improve register load order in subroutine call args assignments:
+in certain situations, the "wrong" order of evaluation of function call arguments is done which results
+in overwriting registers that already got their value, which requires a lot of stack juggling (especially on plain 6502 cpu!)
+Maybe this routine can be made more intelligent.  See usesOtherRegistersWhileEvaluating() and argumentsViaRegisters().
+
+
 Future Things and Ideas
 ^^^^^^^^^^^^^^^^^^^^^^^
 Compiler: