Optimize code size a bit

src/hal/eeprom.h

@@ -1,6 +1,7 @@
 /// @file eeprom.h
 #pragma once
 #include <stdint.h>
+#include <stddef.h>
 
 namespace hal {
 
@@ -9,7 +10,11 @@ namespace eeprom {
 
 class EEPROM {
 public:
+#ifdef UNITTEST
+    using addr_t = size_t;
+#else
     using addr_t = uint16_t;
+#endif
 
     static void WriteByte(addr_t addr, uint8_t value);
     static void UpdateByte(addr_t addr, uint8_t value);

src/logic/feed_to_bondtech.cpp

@@ -29,8 +29,7 @@ void logic::FeedToBondtech::GoToPushToNozzle() {
     state = PushingFilamentIntoNozzle;
 }
 
-void FeedToBondtech::UpdateBowdenLength(int32_t feedEnd_mm) {
+void FeedToBondtech::UpdateBowdenLength(int32_t measuredBowdenLength) {
-    int32_t measuredBowdenLength = abs(feedEnd_mm - feedStart_mm);
     if (measuredBowdenLength < config::maximumBowdenLength.v) { // is the measured length any valid/acceptable?
         static_assert(config::maximumBowdenLength.v <= 65535, "Max bowden length too long");
         int16_t mbl = (int16_t)measuredBowdenLength;
@@ -121,7 +120,7 @@ bool FeedToBondtech::Step() {
             dbg_logic_P(PSTR("Feed to Bondtech --> Idler disengaged"));
             dbg_logic_fP(PSTR("Pulley end steps %u"), mpu::pulley.CurrentPosition_mm());
             state = OK;
-            UpdateBowdenLength(mpu::pulley.CurrentPosition_mm());
+            UpdateBowdenLength(abs(mpu::pulley.CurrentPosition_mm() - feedStart_mm));
             ml::leds.SetMode(mg::globals.ActiveSlot(), ml::green, ml::on);
         }
         return false;

src/logic/feed_to_bondtech.h

@@ -52,7 +52,7 @@ struct FeedToBondtech {
 
 private:
     /// Update bowden length if changed significantly
-    void UpdateBowdenLength(int32_t feedEnd_mm);
+    void UpdateBowdenLength(int32_t measuredBowdenLength);
 
     /// Common processing of pushing filament into fsensor (reused by multiple states)
     //    bool PushingFilament();

src/modules/permanent_storage.cpp

@@ -53,7 +53,7 @@ constexpr const uint16_t eepromBowdenLenMaximum = config::maximumBowdenLength.v;
 
 namespace ee = hal::eeprom;
 
-#define EEOFFSET(x) reinterpret_cast<size_t>(&(x))
+#define EEOFFSET(x) reinterpret_cast<ee::EEPROM::addr_t>(&(x))
 
 void Init() {
     if (ee::EEPROM::ReadByte(ee::EEPROM::End()) != layoutVersion) {
@@ -93,11 +93,10 @@ static bool validBowdenLen(const uint16_t BowdenLength) {
 /// @return stored bowden length
 uint16_t BowdenLength::Get(uint8_t slot) {
     if (validFilament(slot)) {
-        // @@TODO these reinterpret_cast expressions look horrible but I'm keeping them almost intact to respect the original code from MM_control_01
+        uint16_t bowdenLength = ee::EEPROM::ReadWord(EEOFFSET(eepromBase->eepromBowdenLen[slot]));
-        uint16_t bowdenLength = ee::EEPROM::ReadWord(reinterpret_cast<size_t>(&(eepromBase->eepromBowdenLen[slot])));
 
         if (eepromEmpty == bowdenLength) {
-            const uint8_t LengthCorrectionLegacy = ee::EEPROM::ReadByte(reinterpret_cast<size_t>(&(eepromBase->eepromLengthCorrection)));
+            const uint8_t LengthCorrectionLegacy = ee::EEPROM::ReadByte(EEOFFSET(eepromBase->eepromLengthCorrection));
             if (LengthCorrectionLegacy <= 200) {
                 bowdenLength = eepromLengthCorrectionBase + LengthCorrectionLegacy * 10;
             }

src/modules/pulley.h

@@ -27,7 +27,9 @@ public:
     void PlanMove(mm::P_pos_t delta, mm::P_speed_t feed_rate, mm::P_speed_t end_rate = { 0 });
 
     // NOTE: always_inline is required here to force gcc <= 7.x to evaluate each call at compile time
-    void __attribute__((always_inline)) PlanMove(unit::U_mm delta, unit::U_mm_s feed_rate, unit::U_mm_s end_rate = { 0 }) {
+    // But it increases the code size with gcc 7.x by ~200B when PlanMove is called with non-constant feed rates
+    // -> so we don't know yet how to avoid those stupid to/from float conversions (due to C++ unit system), but we'll find a way around ;)
+    void /*__attribute__((always_inline))*/ PlanMove(unit::U_mm delta, unit::U_mm_s feed_rate, unit::U_mm_s end_rate = { 0 }) {
         PlanMove(mm::unitToAxisUnit<mm::P_pos_t>(delta),
             mm::unitToAxisUnit<mm::P_speed_t>(feed_rate),
             mm::unitToAxisUnit<mm::P_speed_t>(end_rate));

src/registers.cpp

@@ -244,7 +244,7 @@ static const RegisterRec registers[] /*PROGMEM*/ = {
     RegisterRec(false, &project_build_number),
     // 0x04
     RegisterRec( // MMU errors
-        []() -> uint16_t { return mg::globals.DriveErrors(); },
+        []() -> uint16_t { return mg::globals.DriveErrors(); }, // compiles to: <{lambda()#1}::_FUN()>: jmp <modules::permanent_storage::DriveError::get()>
         // [](uint16_t) {}, // @@TODO think about setting/clearing the error counter from the outside
         2),
     // 0x05
@@ -269,18 +269,18 @@ static const RegisterRec registers[] /*PROGMEM*/ = {
     RegisterRec([]() -> uint16_t { return static_cast<uint16_t>(mg::globals.MotorsStealth()); }, 1),
     // 0xb extra load distance after fsensor triggered (30mm default) [mm] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.FSensorToNozzle_mm().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.FSensorToNozzle_mm()); },
         [](uint16_t d) { mg::globals.SetFSensorToNozzle_mm(d); },
         1),
     // 0x0c fsensor unload check distance (40mm default) [mm] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.FSensorUnloadCheck_mm().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.FSensorUnloadCheck_mm()); },
         [](uint16_t d) { mg::globals.SetFSensorUnloadCheck_mm(d); },
         1),
 
     // 0xd 2 Pulley unload feedrate [mm/s] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.PulleyUnloadFeedrate_mm_s().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.PulleyUnloadFeedrate_mm_s()); },
         [](uint16_t d) { mg::globals.SetPulleyUnloadFeedrate_mm_s(d); },
         2),
 
@@ -302,23 +302,23 @@ static const RegisterRec registers[] /*PROGMEM*/ = {
 
     // 0x11 Pulley load feedrate [mm/s] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.PulleyLoadFeedrate_mm_s().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.PulleyLoadFeedrate_mm_s()); },
         [](uint16_t d) { mg::globals.SetPulleyLoadFeedrate_mm_s(d); },
         2),
     // 0x12 Selector nominal feedrate [mm/s] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.SelectorFeedrate_mm_s().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.SelectorFeedrate_mm_s()); },
         [](uint16_t d) { mg::globals.SetSelectorFeedrate_mm_s(d); },
         2),
     // 0x13 Idler nominal feedrate [deg/s] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.IdlerFeedrate_deg_s().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.IdlerFeedrate_deg_s()); },
         [](uint16_t d) { mg::globals.SetIdlerFeedrate_deg_s(d); },
         2),
 
     // 0x14 Pulley slow load to fsensor feedrate [mm/s] RW
     RegisterRec(
-        []() -> uint16_t { return mg::globals.PulleySlowFeedrate_mm_s().v; },
+        []() -> uint16_t { return mm::truncatedUnit(mg::globals.PulleySlowFeedrate_mm_s()); },
         [](uint16_t d) { mg::globals.SetPulleySlowFeedrate_mm_s(d); },
         2),
     // 0x15 Selector homing feedrate [mm/s] RW