change to uint8_t for bytecode to reduce wasted bytes

null_test.ar

@@ -1 +0,0 @@
-let x(hello,lol,world, WORLD,HELLOOOO,LOLLLLLL, WORLD)=let f(x)="bruh"

src/main.c

@@ -13,12 +13,15 @@
 #include <stdint.h>
 #include <stdio.h>
 #include <string.h>
+#include <time.h>
 #include <unistd.h>
 
 const char FILE_IDENTIFIER[] = "ARBI";
-const uint64_t version_number = 0;
+const uint32_t version_number = 0;
 
 int main(int argc, char *argv[]) {
+  clock_t start,end;
+  double time_spent;
   setlocale(LC_ALL, "");
   if (argc <= 1)
     return -1;
@@ -35,37 +38,48 @@ int main(int argc, char *argv[]) {
   }
 
   LexerState state = {path, file, 0, 0, &tokens};
+  start = clock();
   lexer(state);
+  end = clock();
+  time_spent = (double)(end - start) / CLOCKS_PER_SEC;
+  printf("Lexer time taken: %f seconds\n", time_spent);
   fclose(state.file);
 
   DArray ast;
 
   darray_init(&ast, sizeof(ParsedValue));
 
+
+  start = clock();
   parser(path, &ast, &tokens, false);
+  end = clock();
+  time_spent = (double)(end - start) / CLOCKS_PER_SEC;
+  printf("Parser time taken: %f seconds\n", time_spent);
   darray_free(&tokens, free_token);
 
   Translated translated = init_translator();
 
+
+  start = clock();
   translate(&translated, &ast);
+  end = clock();
+  time_spent = (double)(end - start) / CLOCKS_PER_SEC;
+  printf("Translation time taken: %f seconds\n", time_spent);
 
   darray_free(&ast, free_parsed);
 
   file = fopen("out.arbin", "wb");
 
-  uint64_t regCount = (uint64_t)translated.registerCount;
   uint64_t constantsSize = (uint64_t)translated.constants.size;
   uint64_t bytecodeSize = (uint64_t)translated.bytecode.size;
 
-  uint64_t version_number_htole64ed = htole64(version_number);
+  uint32_t version_number_htole32ed = htole32(version_number);
-  regCount = htole64(regCount);
-  regCount = htole64(regCount);
   constantsSize = htole64(constantsSize);
   bytecodeSize = htole64(bytecodeSize);
 
   fwrite(&FILE_IDENTIFIER, sizeof(char), strlen(FILE_IDENTIFIER), file);
-  fwrite(&version_number_htole64ed, sizeof(uint64_t), 1, file);
+  fwrite(&version_number_htole32ed, sizeof(uint32_t), 1, file);
-  fwrite(&regCount, sizeof(uint64_t), 1, file);
+  fwrite(&translated.registerCount, sizeof(uint8_t), 1, file);
   fwrite(&constantsSize, sizeof(uint64_t), 1, file);
   fwrite(&bytecodeSize, sizeof(uint64_t), 1, file);
   fwrite(translated.constants.data, 1, translated.constants.size, file);
@@ -77,8 +91,15 @@ int main(int argc, char *argv[]) {
   generate_siphash_key();
 
   init_types();
+
+  start = clock();
   runtime(translated);
 
+  end = clock();
+  time_spent = (double)(end - start) / CLOCKS_PER_SEC;
+
+  printf("Execution time taken: %f seconds\n", time_spent);
+
   free_translator(&translated);
   return 0;
 }

src/runtime/runtime.c

@@ -14,6 +14,14 @@
 #include <stdlib.h>
 #include <unistd.h>
 
+uint64_t bytes_to_uint64(const uint8_t bytes[8]) {
+  uint64_t value = 0;
+  for (int i = 0; i < 8; i++) {
+    value |= ((uint64_t)bytes[i]) << (i * 8);
+  }
+  return value;
+}
+
 void init_types() {
   BASE_CLASS = init_argon_class("BASE_CLASS");
 
@@ -25,14 +33,22 @@ void init_types() {
   init_base_field();
 }
 
+
+uint8_t pop_byte(Translated *translated, RuntimeState *state) {
+  return *((uint8_t *)darray_get(&translated->bytecode, state->head++));
+}
+
 uint64_t pop_bytecode(Translated *translated, RuntimeState *state) {
-  uint64_t *instruction = darray_get(&translated->bytecode, state->head++);
+  uint8_t bytes[8];
-  return *instruction;
+  for (size_t i = 0; i < sizeof(bytes); i++) {
+    bytes[i] = *((uint8_t *)darray_get(&translated->bytecode, state->head++));
+  }
+  return bytes_to_uint64(bytes);
 }
 
 void load_const(Translated *translated, RuntimeState *state) {
-  uint64_t to_register = pop_bytecode(translated, state);
+  uint64_t to_register = pop_byte(translated, state);
-  types type = pop_bytecode(translated, state);
+  types type = pop_byte(translated, state);
   size_t length = pop_bytecode(translated, state);
   uint64_t offset = pop_bytecode(translated, state);
 
@@ -49,10 +65,10 @@ void load_const(Translated *translated, RuntimeState *state) {
 
 void run_instruction(Translated *translated, RuntimeState *state,
                      struct Stack stack) {
-  OperationType opcode = pop_bytecode(translated, state);
+  OperationType opcode = pop_byte(translated, state);
   switch (opcode) {
   case OP_LOAD_NULL:
-    state->registers[pop_bytecode(translated, state)] = ARGON_NULL;
+    state->registers[pop_byte(translated, state)] = ARGON_NULL;
     break;
   case OP_LOAD_CONST:
     load_const(translated, state);

src/translator/bytecode_spec.md

@@ -1,13 +1,15 @@
 # Bytecode Specification
 
+all opcodes are uint8_t, and all operands are uint64_t unless marked with an asterisk (*), where it is marked as uint8_t
+
 ## OP_LOAD_CONST
 
 loads and initialises a value from the constant buffer into the provided register.
 
 this operation 4 operands.
 
-1. the register to write to.
+1. the register to write to. (*)
-2. the type of data from the constant buffer.
+2. the type of data from the constant buffer. (*)
 3. the length of the data in the constant buffer.
 4. the offset in the constant buffer.
 
@@ -19,7 +21,7 @@ this operation takes 3 operands.
 
 1. the length of the variable name.
 2. the offset in the constant buffer of the variable name.
-3. the register of the given value
+3. the register of the given value (*)
 
 ## OP_LOAD_NULL
 
@@ -27,7 +29,7 @@ sets a given register to null.
 
 this operation takes 1 operand.
 
-1. the register to set to null.
+1. the register to set to null. (*)
 
 ## OP_LOAD_FUNCTION
 

src/translator/declaration/declaration.c

@@ -17,14 +17,14 @@ size_t translate_parsed_declaration(Translated *translated,
     size_t offset =
         arena_push(&translated->constants, singleDeclaration->name, length);
 
-    push_instruction_code(translated, OP_DECLARE);
+    push_instruction_byte(translated, OP_DECLARE);
     push_instruction_code(translated, length);
     push_instruction_code(translated, offset);
-    push_instruction_code(translated, 0);
+    push_instruction_byte(translated, 0);
   }
   if (delcarations.size != 1) {
-    push_instruction_code(translated, OP_LOAD_NULL);
+    push_instruction_byte(translated, OP_LOAD_NULL);
-    push_instruction_code(translated, 0);
+    push_instruction_byte(translated, 0);
   }
   return first;
 }

src/translator/function/function.c

@@ -1,22 +1,24 @@
 #include "function.h"
 #include "../translator.h"
 #include <stddef.h>
+#include <stdint.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
 size_t translate_parsed_function(Translated *translated,
                                  ParsedFunction *parsedFunction) {
-  DArray temp_bytecode;
-  darray_init(&temp_bytecode, sizeof(uint64_t));
   DArray main_bytecode = translated->bytecode;
-  translated->bytecode = temp_bytecode;
+  DArray _temp_bytecode;
+  darray_init(&_temp_bytecode, sizeof(uint8_t));
+  translated->bytecode = _temp_bytecode;
   translate_parsed(translated, parsedFunction->body);
   size_t function_bytecode_offset =
       arena_push(&translated->constants, translated->bytecode.data,
                  translated->bytecode.size*translated->bytecode.element_size);
   size_t function_bytecode_length = translated->bytecode.size;
+  darray_free(&translated->bytecode, NULL);
   translated->bytecode = main_bytecode;
-  darray_free(&temp_bytecode, NULL);
+  size_t start = push_instruction_byte(translated, OP_LOAD_FUNCTION);
-  size_t start = push_instruction_code(translated, OP_LOAD_FUNCTION);
   size_t offset = arena_push(&translated->constants, parsedFunction->name,
                              strlen(parsedFunction->name));
   push_instruction_code(translated, offset);

src/translator/number/number.c

@@ -10,10 +10,10 @@ size_t translate_parsed_number(Translated *translated, char *number_str, size_t
   size_t number_pos = arena_push(&translated->constants, number_str, length);
   set_registers(translated, to_register+1);
   
-  size_t start = push_instruction_code(translated, OP_LOAD_CONST);
+  size_t start = push_instruction_byte(translated, OP_LOAD_CONST);
-  push_instruction_code(translated, to_register);
+  push_instruction_byte(translated, to_register);
 
-  push_instruction_code(translated, TYPE_OP_NUMBER);
+  push_instruction_byte(translated, TYPE_OP_NUMBER);
   push_instruction_code(translated,length);
   push_instruction_code(translated, number_pos);
   return start;

src/translator/string/string.c

@@ -7,9 +7,9 @@
 size_t translate_parsed_string(Translated *translated, ParsedString parsedString) {
   size_t string_pos = arena_push(&translated->constants, parsedString.string, parsedString.length);
   set_registers(translated, 1);
-  size_t start = push_instruction_code(translated, OP_LOAD_CONST);
+  size_t start = push_instruction_byte(translated, OP_LOAD_CONST);
-  push_instruction_code(translated, 0);
+  push_instruction_byte(translated, 0);
-  push_instruction_code(translated, TYPE_OP_STRING);
+  push_instruction_byte(translated, TYPE_OP_STRING);
   push_instruction_code(translated,parsedString.length);
   push_instruction_code(translated, string_pos);
   return start;

src/translator/translator.c

@@ -9,6 +9,12 @@
 #include <stdlib.h>
 #include <string.h>
 
+void uint64_to_bytes(uint64_t value, uint8_t bytes[8]) {
+  for (int i = 0; i < 8; i++) {
+    bytes[i] = (value >> (i * 8)) & 0xFF;
+  }
+}
+
 void arena_init(ConstantArena *arena) {
   arena->data = checked_malloc(CHUNK_SIZE);
   arena->capacity = CHUNK_SIZE;
@@ -56,25 +62,29 @@ size_t arena_push(ConstantArena *arena, const void *data, size_t length) {
 Translated init_translator() {
   Translated translated;
   translated.registerCount = 0;
-  darray_init(&translated.bytecode, sizeof(uint64_t));
+  darray_init(&translated.bytecode, sizeof(uint8_t));
   arena_init(&translated.constants);
   return translated;
 }
 
-void set_instruction_code(Translated *translator, size_t offset,
-                          uint64_t code) {
-  size_t *ptr = (translator->bytecode.data + offset);
-  *ptr = code;
-}
 
-size_t push_instruction_code(Translated *translator, uint64_t code) {
+size_t push_instruction_byte(Translated *translator, uint8_t byte) {
-  code = htole64(code);
   size_t offset = translator->bytecode.size;
-  darray_push(&translator->bytecode, &code);
+  darray_push(&translator->bytecode, &byte);
   return offset;
 }
 
-void set_registers(Translated *translator, size_t count) {
+size_t push_instruction_code(Translated *translator, uint64_t code) {
+  size_t offset = translator->bytecode.size;
+  uint8_t bytes[8];
+  uint64_to_bytes(code, bytes);
+  for (size_t i = 0; i < sizeof(bytes); i++) {
+    darray_push(&translator->bytecode, &(bytes[i]));
+  }
+  return offset;
+}
+
+void set_registers(Translated *translator, uint8_t count) {
   if (count > translator->registerCount)
     translator->registerCount = count;
 }
@@ -91,8 +101,8 @@ size_t translate_parsed(Translated *translated, ParsedValue *parsedValue) {
     return translate_parsed_number(translated, (char *)parsedValue->data, 0);
   case AST_NULL:
     set_registers(translated, 1);
-    size_t output = push_instruction_code(translated, OP_LOAD_NULL);
+    size_t output = push_instruction_byte(translated, OP_LOAD_NULL);
-    push_instruction_code(translated, 0);
+    push_instruction_byte(translated, 0);
     return output;
   case AST_FUNCTION:
     return translate_parsed_function(translated,

src/translator/translator.h

@@ -7,8 +7,8 @@
 #include <stddef.h>
 #include <stdint.h>
 
-typedef enum { OP_LOAD_CONST = 255, OP_DECLARE, OP_LOAD_NULL, OP_LOAD_FUNCTION } OperationType;
+typedef enum { OP_LOAD_CONST, OP_DECLARE, OP_LOAD_NULL, OP_LOAD_FUNCTION } OperationType;
-typedef enum { TYPE_OP_STRING = 255, TYPE_OP_NUMBER } types;
+typedef enum { TYPE_OP_STRING, TYPE_OP_NUMBER } types;
 
 typedef struct {
   void *data;
@@ -17,7 +17,7 @@ typedef struct {
 } ConstantArena;
 
 typedef struct {
-  size_t registerCount;
+  uint8_t registerCount;
   DArray bytecode;
   ConstantArena constants;
 } Translated;
@@ -26,11 +26,11 @@ void *arena_get(ConstantArena *arena, size_t offset);
 
 size_t arena_push(ConstantArena *arena, const void *data, size_t length);
 
-void set_instruction_code(Translated * translator, size_t offset, uint64_t code);
+size_t push_instruction_byte(Translated *translator, uint8_t byte);
 
 size_t push_instruction_code(Translated *translator, uint64_t code);
 
-void set_registers(Translated *translator, size_t count);
+void set_registers(Translated *translator, uint8_t count);
 
 Translated init_translator();
 

test.ar

@@ -1,6 +1,3 @@
-null
-
-
 "h"
 "e"
 "ll"