Support printing floats

Author: Dylan McKay

avrlit.py.in

@@ -151,7 +151,8 @@ def link(objects, config):
   executablePath = "{}/{}.elf".format(config.output_dir, "executable")
   libavrlit_path = os.path.join(libavrlit_dir, config.board, "libavrlit.a")
   arguments = objects + ["-mmcu=" + config.mcu,
-                        "-o", executablePath, libavrlit_path]
+                        "-o", executablePath, libavrlit_path,
+                        "-Wl,-u,vfprintf", "-lprintf_flt", "-lm"]
   runCommand(config.ld, arguments, config)
   return executablePath
 

libavrlit/src/avrlit.cpp

@@ -11,6 +11,7 @@
 #include "serial/serial.h"
 
 #include <stdlib.h>
+#include <stdio.h>
 
 namespace {
 
@@ -27,6 +28,18 @@ void reportInteger(const char *name, T result, bool includeColon) {
   avrlit::SerialImpl::GetCurrent()->send(buffer);
 }
 
+void reportFloat(const char *name, float f, bool includeColon) {
+  avrlit::SerialImpl::GetCurrent()->send(name);
+
+  if (includeColon)
+    avrlit::SerialImpl::GetCurrent()->send(": ");
+
+  char buffer[40];
+  sprintf(buffer, "%f", (double)f);
+
+  avrlit::SerialImpl::GetCurrent()->send(buffer);
+}
+
 template<typename T>
 void reportArgumentInteger(const char *name, uint8_t idx, T result) {
   if (idx > 0) {
@@ -35,12 +48,24 @@ void reportArgumentInteger(const char *name, uint8_t idx, T result) {
 
   reportInteger(name, result, true);
 }
+void reportArgumentFloat(const char *name, uint8_t idx, float result) {
+  if (idx > 0) {
+    avrlit::SerialImpl::GetCurrent()->send(", ");
+  }
+
+  reportFloat(name, result, true);
+}
 
 template<typename T>
 void reportResultInteger(T result) {
   reportInteger("return ", result, false);
   avrlit::SerialImpl::GetCurrent()->send('\n');
 }
+
+void reportResultFloat(float result) {
+  reportFloat("return ", result, false);
+  avrlit::SerialImpl::GetCurrent()->send('\n');
+}
 }
 
 extern "C" {
@@ -79,14 +104,14 @@ extern "C" {
   void avr_instrumentation_argument_i32(const char *argName, uint8_t idx, uint32_t val) { reportArgumentInteger(argName, idx, val); }
   void avr_instrumentation_argument_i64(const char *argName, uint8_t idx, uint64_t val) { reportArgumentInteger(argName, idx, val); }
 
-  void avr_instrumentation_argument_f32(const char *argName, uint8_t idx, float val) { }
+  void avr_instrumentation_argument_f32(const char *argName, uint8_t idx, float val) { reportArgumentFloat(argName, idx, val); }
 
   void avr_instrumentation_result_i8(uint8_t result) { reportResultInteger(result); }
   void avr_instrumentation_result_i16(uint16_t result) { reportResultInteger(result); }
   void avr_instrumentation_result_i32(uint32_t result) { reportResultInteger(result); }
   void avr_instrumentation_result_i64(uint64_t result) { reportResultInteger(result); }
 
-  void avr_instrumentation_result_f32(float result) {  }
+  void avr_instrumentation_result_f32(float result) { reportResultFloat(result); }
 }
 
 extern "C" {