symbol interning

.editorconfig

@@ -0,0 +1,13 @@
+root = true
+
+[*]
+end_of_line = lf
+insert_final_newline = true
+
+[*.{c,h}]
+indent_style = space
+indent_size = 2
+
+[Makefile]
+indent_style = tab
+indent_size = 4

Makefile

@@ -1,6 +1,9 @@
-CFLAGS := -std=c99 -Og -g -Wpedantic -Wall
+CFLAGS	:= -std=c99 -Og -g -Wpedantic -Wall
-.PHONY: clean
+OBJS	:= main.o sym.o
 
-wscm: wscm.o
+wscm: $(OBJS)
+	$(CC) $(OBJS) -o wscm
+
+.PHONY: clean
 clean:
-	rm -f wscm wscm.o
+	rm -f wscm main.o sym.o

main.c

@@ -0,0 +1,12 @@
+#include <stdio.h>
+#include "wscm.h"
+
+int main(void) {
+  const S *hello = intern("hello", -1);
+  printf("hello  = %p\n", (void *)hello);
+
+  const S *hello2 = intern("hello", -1);
+  printf("hello2 = %p (should be equal to hello)\n", (void *)hello2);
+
+  return 0;
+}

sym.c

@@ -0,0 +1,81 @@
+#include <stdlib.h>
+#include <string.h>
+
+#include "wscm.h"
+
+St syms = {0, 0, NULL};
+
+static S *findsym(const char *str, Z len, U32 hash) {
+  if (syms.capacity == 0)
+    return NULL;
+
+  Z ix = hash % syms.capacity;
+  for (Z i = 0; i < syms.capacity; i++) {
+    S *s = syms.data[ix];
+    if (!s)
+      return NULL;
+    if (s->hash == hash && s->len == len)
+      return s;
+    ix = (ix + 1) % syms.capacity;
+  }
+  return NULL;
+}
+
+static void symtabresize(void) {
+  Z cap = syms.capacity;
+  if (cap == 0) {
+    syms.capacity = 16;
+  } else {
+    syms.capacity *= 2;
+  }
+  S **nb = calloc(syms.capacity, sizeof(S *));
+  for (Z i = 0; i < cap; i++) {
+    if (syms.data[i]) {
+      S *s = syms.data[i];
+      Z ix = s->hash % syms.capacity;
+      while (nb[ix])
+        ix = (ix + 1) % syms.capacity;
+      nb[ix] = s;
+    }
+  }
+  if (syms.data != NULL)
+    free(syms.data);
+  syms.data = nb;
+}
+
+U32 hashstring(const char *data, Z len) {
+  U32 hash = 2166136261u;
+  for (Z i = 0; i < len; i++) {
+    hash ^= (uint8_t)data[i];
+    hash *= 16777619u;
+  }
+  return hash;
+}
+
+S *intern(const char *str, I len) {
+  if (len < 0)
+    len = strlen(str);
+
+  U32 hash = hashstring(str, len);
+  S *s = findsym(str, len, hash);
+  if (s)
+    return s;
+
+  s = malloc(sizeof(S));
+  s->data = malloc(len);
+  memcpy(s->data, str, len);
+  s->len = len;
+  s->hash = hash;
+
+  if (syms.count + 1 > syms.capacity)
+    symtabresize();
+
+  Z ix = hash % syms.capacity;
+  while (syms.data[ix])
+    ix = (ix + 1) % syms.capacity;
+
+  syms.data[ix] = s;
+  syms.count++;
+
+  return s;
+}

wscm.c

@@ -1,6 +0,0 @@
-#include <stdio.h>
-
-int main(void) {
-  printf("Hello, world!\n");
-  return 0;
-}

wscm.h

@@ -0,0 +1,61 @@
+#include <stddef.h>
+#include <stdint.h>
+
+// common types
+typedef uintptr_t U;
+typedef intptr_t I;
+
+typedef uint8_t U8;
+typedef uint32_t U32;
+typedef int32_t I32;
+typedef size_t Z;
+
+// objects
+typedef uintptr_t O;
+
+// cons pair
+typedef struct C C;
+struct C {
+  O head, tail;
+};
+
+// symbol
+typedef struct S S;
+struct S {
+  U8 *data;
+  U32 hash;
+  Z len;
+};
+
+// symbol table
+typedef struct St St;
+struct St {
+  I count;
+  Z capacity;
+  S **data;
+};
+
+// gc header
+typedef struct H H;
+struct H {
+  I type;
+  Z len;
+};
+
+// heap
+typedef struct E E;
+struct E {
+  struct {
+    U8 *start, *end;
+    U8 *free;
+  } from, to;
+
+  I root_count;
+  Z root_capacity;
+  O **roots;
+};
+
+extern E heap;
+extern St syms;
+
+S *intern(const char *str, I len);