diff --git a/InstructionSets/ARM/Executor.hpp b/InstructionSets/ARM/Executor.hpp
index 6294235cd..76096a58a 100644
--- a/InstructionSets/ARM/Executor.hpp
+++ b/InstructionSets/ARM/Executor.hpp
@@ -52,6 +52,10 @@ struct Executor {
 						registers_.pc(4) :
 						registers_[fields.shift_register()];
 
+				// "The amount by which the register should be shifted may be contained in
+				// ... **the bottom byte** of another register".
+				shift_amount &= 0xff;
+
 				// A register shift amount of 0 has a different meaning than an in-instruction
 				// shift amount of 0.
 				if(!shift_amount) {
@@ -98,6 +102,12 @@ struct Executor {
 			operand2 = fields.immediate();
 			if(fields.rotate()) {
 				shift<ShiftType::RotateRight, shift_sets_carry>(operand2, fields.rotate(), rotate_carry);
+			} else {
+				// This is possibly clarified by later data sheets; take carry as if a rotate by 32
+				// had occurred.
+				if constexpr (shift_sets_carry) {
+					rotate_carry = operand2 & 0x8000'0000;
+				}
 			}
 		} else {
 			operand2 = decode_shift<true, shift_sets_carry>(fields, rotate_carry, shift_by_register ? 8 : 4);
diff --git a/OSBindings/Mac/Clock SignalTests/ARMDecoderTests.mm b/OSBindings/Mac/Clock SignalTests/ARMDecoderTests.mm
index 5fa9b2690..cd34246d3 100644
--- a/OSBindings/Mac/Clock SignalTests/ARMDecoderTests.mm	
+++ b/OSBindings/Mac/Clock SignalTests/ARMDecoderTests.mm	
@@ -383,22 +383,6 @@ struct MemoryLedger {
 			uint32_t r15_mask = 0xffff'ffff;
 			bool ignore_test = false;
 			switch(instruction) {
-				case 0x03110002:
-					// tsteq r1, #2; per my reading this is LSL#0 so the original
-					// carry value should be preserved. The test set doesn't seem
-					// to agree. Until I can reconcile them, don't test carry.
-					r15_mask &= ~ConditionCode::Carry;
-				break;
-
-				case 0x11800215:
-					// orrne r0, r0, r5, lsl r2
-					// The test set applies the rule 'lsl r2 % 32' whereas the data
-					// sheet specifically says for a shift-by-register that
-					// "LSL by more than 32 has result zero, carry out zero" so I think
-					// the test set is adrift on the following:
-					ignore_test = regs[2] >= 32;
-				break;
-
 				case 0xe090e00f:
 					// adds lr, r0, pc
 					// The test set comes from an ARM that doesn't multiplex flags
@@ -407,6 +391,11 @@ struct MemoryLedger {
 					regs[15] &= 0x03ff'fffc;
 				break;
 
+				// TODO:
+				//	* adds to R15: e090f00e, e090f00f; possibly to do with non-multiplexing original?
+				//	* movs to PC: e1b0f00e; as above?
+				//	* some test set teqs — e33ff3c3, e33ff343, e33ef000 — seem to advance the PC twice?
+
 				default: break;
 			}
 
@@ -429,10 +418,10 @@ struct MemoryLedger {
 				if(ignore_test) {
 					continue;
 				}
-
-				if(instruction == 0xe33ef000 && test_count == 1) {
-					printf("");
-				}
+//
+//				if(instruction == 0x03110002 && test_count == 5) {
+//					printf("");
+//				}
 
 				execute(instruction, *test);
 				NSMutableString *error = nil;