compiler now compiles some things

src/chunk.c

@@ -1,6 +1,37 @@
+#include <stdio.h>
+#include <stdlib.h>
+
 #include "chunk.h"
+
 #include "vendor/yar.h"
 
+Bc *chunk_new(V) {
+  Bc *chunk = calloc(1, sizeof(Bc));
+  chunk->ref = 1;
+  return chunk;
+}
+
+V chunk_acquire(Bc *chunk) {
+#if CHUNK_DEBUG
+  fprintf(stderr, "DEBUG: acquiring chunk at %p\n", (V *)chunk);
+#endif
+  chunk->ref++;
+}
+V chunk_release(Bc *chunk) {
+#if CHUNK_DEBUG
+  fprintf(stderr, "DEBUG: releasing chunk at %p\n", (V *)chunk);
+#endif
+
+  if (--chunk->ref == 0) {
+#if CHUNK_DEBUG
+    fprintf(stderr, "DEBUG: freeing chunk at %p\n", (V *)chunk);
+#endif
+    yar_free(&chunk->constants);
+    yar_free(chunk);
+    free(chunk);
+  }
+}
+
 V chunk_emit_byte(Bc *chunk, U8 byte) { *yar_append(chunk) = byte; }
 
 V chunk_emit_sleb128(Bc *chunk, I num) {
@@ -22,8 +53,3 @@ I chunk_add_constant(Bc *chunk, O value) {
   *yar_append(&chunk->constants) = value;
   return mark;
 }
-
-V chunk_free(Bc *chunk) {
-  yar_free(&chunk->constants);
-  yar_free(chunk);
-}

src/chunk.h

@@ -1,11 +1,14 @@
 #ifndef CHUNK_H
 #define CHUNK_H
 
+#define CHUNK_DEBUG 0
+
 #include "common.h"
 #include "object.h"
 
 /** Bytecode chunk */
 typedef struct Bc {
+  I ref;
   U8 *items;
   Z count, capacity;
   struct {
@@ -14,9 +17,12 @@ typedef struct Bc {
   } constants;
 } Bc;
 
+Bc *chunk_new(V);
+V chunk_acquire(Bc *);
+V chunk_release(Bc *);
+
 V chunk_emit_byte(Bc *, U8);
 V chunk_emit_sleb128(Bc *, I);
 I chunk_add_constant(Bc *, O);
-V chunk_free(Bc *);
 
 #endif

src/compile.c

@@ -1,6 +1,116 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "chunk.h"
 #include "compile.h"
+#include "debug.h"
+#include "gc.h"
+#include "object.h"
+#include "vm.h"
 
+#include "vendor/mpc.h"
 
-I compile_program(mpc_ast_t *ast) {
+// clang-format off
+struct {
+  const char *name;
+  U8 opcode;
+} primitives[] = {
+  {"+", OP_ADD},
+  {"call", OP_APPLY},
+  {NULL, 0},
+};
+// clang-format on
 
+static I compile_expr(Cm *cm, mpc_ast_t *curr, mpc_ast_trav_t **next);
+static I compile_constant(Cm *cm, O value) {
+  I idx = chunk_add_constant(cm->chunk, value);
+  chunk_emit_byte(cm->chunk, OP_CONST);
+  chunk_emit_sleb128(cm->chunk, idx);
+  return 1;
+}
+
+static I compile_quotation(Cm *cm, mpc_ast_t *curr, mpc_ast_trav_t **next) {
+  Cm inner = {0};
+  inner.chunk = chunk_new();
+  inner.gc = cm->gc;
+  inner.dictionary = cm->dictionary;
+
+  (void)mpc_ast_traverse_next(next); // skip opening bracket
+  curr = mpc_ast_traverse_next(next);
+  while (curr != NULL) {
+    if (strcmp(curr->tag, "char") == 0 && strcmp(curr->contents, "]") == 0)
+      break;
+    I res = compile_expr(&inner, curr, next);
+    if (!res)
+      return res;
+    curr = mpc_ast_traverse_next(next);
+  }
+  chunk_emit_byte(inner.chunk, OP_RETURN);
+
+  Hd *hd = gc_alloc(cm->gc, sizeof(Hd) + sizeof(Bc *));
+  hd->type = OBJ_QUOT;
+  Bc **chunk_ptr = (Bc **)(hd + 1);
+  *chunk_ptr = inner.chunk;
+
+  O quot = BOX(hd);
+  compile_constant(cm, quot);
+
+  return 1;
+}
+
+static I compile_expr(Cm *cm, mpc_ast_t *curr, mpc_ast_trav_t **next) {
+  if (strstr(curr->tag, "expr|number") != NULL) {
+    I num = strtol(curr->contents, NULL, 0);
+    return compile_constant(cm, NUM(num));
+  } else if (strstr(curr->tag, "expr|word") != NULL) {
+    for (Z i = 0; primitives[i].name != NULL; i++) {
+      if (strcmp(curr->contents, primitives[i].name) == 0) {
+        chunk_emit_byte(cm->chunk, primitives[i].opcode);
+        return 1;
+      }
+    }
+    fprintf(stderr, "compiler: dictionary nyi\n");
+    return 0;
+  } else if (strstr(curr->tag, "expr|quotation") != NULL) {
+    return compile_quotation(cm, curr, next);
+  } else {
+    fprintf(stderr, "compiler: \"%s\" nyi\n", curr->tag);
+    return 0;
+  }
+
+  return 1;
+}
+
+static I compile_ast(Cm *cm, mpc_ast_t *curr, mpc_ast_trav_t **next) {
+  (void)mpc_ast_traverse_next(next);
+  curr = mpc_ast_traverse_next(next);
+  while (curr != NULL) {
+    if (strcmp(curr->tag, "regex") == 0 && strcmp(curr->contents, "") == 0)
+      break;
+    I res = compile_expr(cm, curr, next);
+    if (!res)
+      return res;
+    curr = mpc_ast_traverse_next(next);
+  }
+
+  return 1;
+}
+
+Bc *compile_program(Gc *gc, mpc_ast_t *ast) {
+  Cm cm = {0};
+  cm.chunk = chunk_new();
+  cm.gc = gc;
+
+  mpc_ast_trav_t *next = mpc_ast_traverse_start(ast, mpc_ast_trav_order_pre);
+  mpc_ast_t *curr = mpc_ast_traverse_next(&next); // Begin traversal
+
+  if (!compile_ast(&cm, curr, &next)) {
+    chunk_release(cm.chunk);
+    return NULL;
+  }
+
+  Bc *chunk = cm.chunk;
+  chunk_emit_byte(chunk, OP_RETURN);
+  return chunk;
 }

src/compile.h

@@ -1,4 +1,24 @@
 #include "common.h"
+
+#include "chunk.h"
+#include "gc.h"
+
 #include "vendor/mpc.h"
 
-I compile_program(mpc_ast_t *);
+/** Compiler dictionary */
+typedef struct Cd Cd;
+struct Cd {
+  Cd *child[4];
+  const char *name;
+  Z offset;
+};
+
+/** Compiler context */
+typedef struct Cm {
+  Gc *gc;
+  Bc *chunk;
+  Cd *dictionary;
+} Cm;
+
+// The chunk returned by `compile_program` is owned by the caller.
+Bc *compile_program(Gc *, mpc_ast_t *);

src/debug.c

@@ -0,0 +1,81 @@
+#include <stdio.h>
+
+#include "debug.h"
+#include "print.h"
+#include "vm.h"
+
+static I decode_sleb128(U8 *ptr, Z *bytes_read) {
+  I result = 0;
+  I shift = 0;
+  U8 byte;
+  Z count = 0;
+  do {
+    byte = ptr[count++];
+    result |= (I)(byte & 0x7F) << shift;
+    shift += 7;
+  } while (byte & 0x80);
+  if ((shift < 64) && (byte & 0x40))
+    result |= -(1LL << shift);
+  *bytes_read = count;
+  return result;
+}
+
+V disassemble(Bc *chunk, const char *name) {
+  printf("=== %s ===\n", name);
+  Z offset = 0;
+  while (offset < chunk->count) {
+    offset = disassemble_instruction(chunk, offset);
+  }
+}
+
+Z disassemble_instruction(Bc *chunk, Z offset) {
+  printf("%04zu ", offset);
+  U8 opcode = chunk->items[offset++];
+  switch (opcode) {
+  case OP_NOP:
+    printf("NOP\n");
+    return offset;
+  case OP_CONST: {
+    Z bytes_read;
+    I idx = decode_sleb128(&chunk->items[offset], &bytes_read);
+    printf("CONST %ld", idx);
+    if (idx >= 0 && idx < (I)chunk->constants.count) {
+      printf(" (");
+      print(chunk->constants.items[idx]);
+      printf(")");
+    }
+    printf("\n");
+    return offset + bytes_read;
+  }
+  case OP_JUMP: {
+    Z bytes_read;
+    I ofs = decode_sleb128(&chunk->items[offset], &bytes_read);
+    printf("JUMP %ld -> %zu\n", ofs, offset + bytes_read + ofs);
+    return offset + bytes_read;
+  }
+  case OP_JUMP_IF_NIL: {
+    Z bytes_read;
+    I ofs = decode_sleb128(&chunk->items[offset], &bytes_read);
+    printf("JUMP_IF_NIL %ld -> %zu\n", ofs, offset + bytes_read + ofs);
+    return offset + bytes_read;
+  }
+  case OP_CALL: {
+    Z bytes_read;
+    I ofs = decode_sleb128(&chunk->items[offset], &bytes_read);
+    printf("CALL %ld\n", ofs);
+    return offset + bytes_read;
+  }
+  case OP_APPLY:
+    printf("APPLY\n");
+    return offset;
+  case OP_RETURN:
+    printf("RETURN\n");
+    return offset;
+  case OP_ADD:
+    printf("ADD\n");
+    return offset;
+  default:
+    printf("? (%d)\n", opcode);
+    return offset;
+  }
+}

src/debug.h

@@ -0,0 +1,5 @@
+#include "chunk.h"
+#include "common.h"
+
+V disassemble(Bc *, const char *);
+Z disassemble_instruction(Bc *, Z);

src/gc.c

@@ -79,7 +79,8 @@ V gc_collect(Gc *gc) {
     switch (hdr->type) {
       // TODO: the rest of the owl
     case OBJ_QUOT: {
-      Bc *chunk = (Bc *)(hdr + 1);
+      Bc **chunk_ptr = (Bc **)(hdr + 1);
+      Bc *chunk = *chunk_ptr;
       for (Z i = 0; i < chunk->constants.count; i++)
         chunk->constants.items[i] = forward(gc, chunk->constants.items[i]);
       break;
@@ -93,6 +94,23 @@ V gc_collect(Gc *gc) {
     scan += ALIGN(hdr->size);
   }
 
+  scan = gc->from.start;
+  while (scan < gc->from.free) {
+    Hd *hdr = (Hd *)scan;
+    if (hdr->type != OBJ_FWD) {
+      switch (hdr->type) {
+      case OBJ_QUOT: {
+        Bc **chunk_ptr = (Bc **)(hdr + 1);
+        chunk_release(*chunk_ptr);
+        break;
+      }
+      default:
+        break;
+      }
+    }
+    scan += ALIGN(hdr->size);
+  }
+
   Gs tmp = gc->from;
   gc->from = gc->to;
   gc->to = tmp;
@@ -103,7 +121,22 @@ V gc_collect(Gc *gc) {
 #endif
 }
 
-void gc_init(Gc *gc) {
+Hd *gc_alloc(Gc *gc, Z sz) {
+  sz = ALIGN(sz);
+  if (gc->from.free + sz > gc->from.end) {
+    gc_collect(gc);
+    if (gc->from.free + sz > gc->from.end) {
+      fprintf(stderr, "out of memory (requested %" PRIdPTR "bytes\n", sz);
+      abort();
+    }
+  }
+  Hd *hdr = (Hd *)gc->from.free;
+  gc->from.free += sz;
+  hdr->size = sz;
+  return hdr;
+}
+
+V gc_init(Gc *gc) {
   gc->from.start = malloc(HEAP_BYTES);
   if (!gc->from.start)
     goto fatal;
@@ -126,7 +159,7 @@ fatal:
   abort();
 }
 
-void gc_deinit(Gc *gc) {
+V gc_deinit(Gc *gc) {
   gc_collect(gc);
   free(gc->from.start);
   free(gc->to.start);

src/main.c

@@ -4,73 +4,59 @@
 #include "common.h"
 
 #include "chunk.h"
-#include "gc.h"
+#include "compile.h"
+#include "debug.h"
 #include "parser.h"
-#include "vendor/mpc.h"
 #include "vm.h"
 
-void dump(const V *data, Z size) {
-  char ascii[17];
-  Z i, j;
-  ascii[16] = '\0';
-  for (i = 0; i < size; ++i) {
-    printf("%02X ", ((unsigned char *)data)[i]);
-    if (((unsigned char *)data)[i] >= ' ' &&
-        ((unsigned char *)data)[i] <= '~') {
-      ascii[i % 16] = ((unsigned char *)data)[i];
-    } else {
-      ascii[i % 16] = '.';
-    }
-    if ((i + 1) % 8 == 0 || i + 1 == size) {
-      printf(" ");
-      if ((i + 1) % 16 == 0) {
-        printf("|  %s \n", ascii);
-      } else if (i + 1 == size) {
-        ascii[(i + 1) % 16] = '\0';
-        if ((i + 1) % 16 <= 8) {
-          printf(" ");
-        }
-        for (j = (i + 1) % 16; j < 16; ++j) {
-          printf("   ");
-        }
-        printf("|  %s \n", ascii);
-      }
-    }
-  }
-}
+#include "vendor/mpc.h"
 
 I repl(void) {
-  Bc chunk = {0};
   Vm vm = {0};
-
   vm_init(&vm);
 
-  I idx = chunk_add_constant(&chunk, NUM(10));
-  chunk_emit_byte(&chunk, OP_CONST);
-  chunk_emit_sleb128(&chunk, idx);
-  chunk_emit_byte(&chunk, OP_RETURN);
+  Bc *chunk = chunk_new();
 
-  vm_run(&vm, &chunk, 0);
+  I idx = chunk_add_constant(chunk, NUM(10));
+  chunk_emit_byte(chunk, OP_CONST);
+  chunk_emit_sleb128(chunk, idx);
+  chunk_emit_byte(chunk, OP_CONST);
+  chunk_emit_sleb128(chunk, idx);
+  chunk_emit_byte(chunk, OP_ADD);
+  chunk_emit_byte(chunk, OP_RETURN);
 
-  return 0;
+  disassemble(chunk, "test chunk");
+  I res = vm_run(&vm, chunk, 0);
+
+  chunk_release(chunk);
+  vm_deinit(&vm);
+  return !res;
 }
 
 I loadfile(const char *fname) {
-  Gc gc = {0};
-  gc_init(&gc);
+  Vm vm = {0};
+  vm_init(&vm);
 
   mpc_result_t res;
   if (!mpc_parse_contents(fname, Program, &res)) {
     mpc_err_print_to(res.error, stderr);
     mpc_err_delete(res.error);
-    gc_deinit(&gc);
     return 1;
   }
 
-  mpc_ast_print(res.output);
+  Bc *chunk = compile_program(&vm.gc, res.output);
   mpc_ast_delete(res.output);
-  gc_deinit(&gc);
-  return 0;
+
+  if (chunk != NULL) {
+    disassemble(chunk, fname);
+    I res = vm_run(&vm, chunk, 0);
+    chunk_release(chunk);
+    vm_deinit(&vm);
+    return !res;
+  } else {
+    vm_deinit(&vm);
+    return 1;
+  }
 }
 
 int main(int argc, const char *argv[]) {

src/object.h

@@ -19,6 +19,7 @@ enum {
   TYPE_NIL = 0,
   TYPE_NUM = 1,
   TYPE_FWD = OBJ_FWD,
+  TYPE_QUOT = OBJ_QUOT,
 };
 
 typedef uintptr_t O;

src/vm.c

@@ -1,8 +1,10 @@
-#include "vm.h"
-#include "gc.h"
-#include "print.h"
 #include <stdio.h>
 
+#include "gc.h"
+#include "object.h"
+#include "print.h"
+#include "vm.h"
+
 static I decode_sleb128(U8 **ptr) {
   I result = 0;
   I shift = 0;
@@ -25,6 +27,7 @@ static I decode_sleb128(U8 **ptr) {
 V vm_init(Vm *vm) {
   vm->sp = vm->stack;
   vm->rsp = vm->rstack;
+  vm->chunk = NULL;
   gc_init(&vm->gc);
 
   for (Z i = 0; i < STACK_SIZE; i++) {
@@ -33,16 +36,39 @@ V vm_init(Vm *vm) {
   }
 }
 
+V vm_deinit(Vm *vm) { gc_deinit(&vm->gc); }
+
 V vm_push(Vm *vm, O o) { *vm->sp++ = o; }
 O vm_pop(Vm *vm) { return *--vm->sp; }
 O vm_peek(Vm *vm) { return *(vm->sp - 1); }
 
-V vm_run(Vm *vm, Bc *chunk, I offset) {
+V vm_rpush(Vm *vm, Bc *chunk, U8 *ip) {
+  vm->rsp->chunk = chunk;
+  vm->rsp->ip = ip;
+  vm->rsp++;
+}
+Fr vm_rpop(Vm *vm) { return *--vm->rsp; }
+
+I vm_run(Vm *vm, Bc *chunk, I offset) {
   I mark = gc_mark(&vm->gc);
   for (Z i = 0; i < chunk->constants.count; i++)
     gc_addroot(&vm->gc, &chunk->constants.items[i]);
 
+#define BINOP(op)                                                              \
+  {                                                                            \
+    O b = vm_pop(vm);                                                          \
+    O a = vm_pop(vm);                                                          \
+    if (!IMM(a) || !IMM(b)) {                                                  \
+      fprintf(stderr, "vm: arithmetic on non-number objects\n");               \
+      return 0;                                                                \
+    }                                                                          \
+    vm_push(vm, NUM(ORD(a) op ORD(b)));                                        \
+    break;                                                                     \
+  }
+
   vm->ip = chunk->items + offset;
+  vm->chunk = chunk;
+
   for (;;) {
     U8 opcode;
     switch (opcode = *vm->ip++) {
@@ -50,20 +76,65 @@ V vm_run(Vm *vm, Bc *chunk, I offset) {
       continue;
     case OP_CONST: {
       I idx = decode_sleb128(&vm->ip);
-      vm_push(vm, chunk->constants.items[idx]);
+      vm_push(vm, vm->chunk->constants.items[idx]);
+      break;
+    }
+    case OP_JUMP: {
+      I ofs = decode_sleb128(&vm->ip);
+      vm->ip += ofs;
+      break;
+    }
+    case OP_JUMP_IF_NIL: {
+      I ofs = decode_sleb128(&vm->ip);
+      if (vm_pop(vm) == NIL)
+        vm->ip += ofs;
+      break;
+    }
+    case OP_CALL: {
+      I ofs = decode_sleb128(&vm->ip);
+      vm_rpush(vm, vm->chunk, vm->ip);
+      vm->ip = chunk->items + ofs;
+      break;
+    }
+    case OP_APPLY: {
+      O quot = vm_pop(vm);
+      if (type(quot) == TYPE_QUOT) {
+        Bc **ptr = (Bc **)(UNBOX(quot) + 1);
+        Bc *chunk = *ptr;
+        vm_rpush(vm, vm->chunk, vm->ip);
+        vm->chunk = chunk;
+        vm->ip = chunk->items;
+      } else {
+        fprintf(stderr, "vm: attempt to apply non-quotation object\n");
+        return 0;
+      }
       break;
     }
     case OP_RETURN:
-      goto done;
+      if (vm->rsp != vm->rstack) {
+        Fr frame = vm_rpop(vm);
+        vm->chunk = frame.chunk;
+        vm->ip = frame.ip;
+      } else {
+        goto done;
+      }
+      break;
+    case OP_ADD:
+      BINOP(+);
+    default:
+      fprintf(stderr, "unknown opcode %d\n", opcode);
+      return 0;
     }
   }
 
 done:
   gc_reset(&vm->gc, mark);
-  // print stack :3
-  for (O *i = vm->stack; i < vm->sp; i++) {
-    print(*i);
-    putchar(' ');
+  if (vm->sp != vm->stack) {
+    for (O *i = vm->stack; i < vm->sp; i++) {
+      print(*i);
+      putchar(' ');
+    }
+    putchar('\n');
   }
-  putchar('\n');
+  return 1;
 }

src/vm.h

@@ -10,25 +10,38 @@
 enum {
   OP_NOP = 0,
   OP_CONST,       // Push constant to stack
+  OP_DROP,
+  OP_DUP,
+  OP_SWAP,
   OP_JUMP,        // Relative jump
   OP_JUMP_IF_NIL, // Relative jump if top-of-stack is nil
-  OP_DOWORD,
   OP_CALL,
+  OP_APPLY,
   OP_RETURN,
+  OP_ADD,
 };
 
 #define STACK_SIZE 256
 
+typedef struct Fr {
+  Bc *chunk;
+  U8 *ip;
+} Fr;
+
 typedef struct Vm {
   Gc gc;
   O stack[256], *sp;
-  U rstack[256], *rsp;
+  Fr rstack[256], *rsp;
   U8 *ip;
+  Bc *chunk;
 } Vm;
 
 V vm_init(Vm *);
+V vm_deinit(Vm *);
+
 V vm_push(Vm *, O);
 O vm_pop(Vm *);
 O vm_peek(Vm *);
-V vm_run(Vm *, Bc *, I);
+
+I vm_run(Vm *, Bc *, I);
 #endif