Integrate more chibicc changes

third_party/chibicc/alloc.c

@@ -0,0 +1,44 @@
+/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
+│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8                                :vi│
+╞══════════════════════════════════════════════════════════════════════════════╡
+│ Copyright 2020 Justine Alexandra Roberts Tunney                              │
+│                                                                              │
+│ Permission to use, copy, modify, and/or distribute this software for         │
+│ any purpose with or without fee is hereby granted, provided that the         │
+│ above copyright notice and this permission notice appear in all copies.      │
+│                                                                              │
+│ THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL                │
+│ WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED                │
+│ WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE             │
+│ AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL         │
+│ DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR        │
+│ PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER               │
+│ TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR             │
+│ PERFORMANCE OF THIS SOFTWARE.                                                │
+╚─────────────────────────────────────────────────────────────────────────────*/
+#include "third_party/chibicc/chibicc.h"
+
+long alloc_node_count;
+long alloc_token_count;
+long alloc_obj_count;
+long alloc_type_count;
+
+Node *alloc_node(void) {
+  ++alloc_node_count;
+  return calloc(1, sizeof(Node));
+}
+
+Token *alloc_token(void) {
+  ++alloc_token_count;
+  return calloc(1, sizeof(Token));
+}
+
+Obj *alloc_obj(void) {
+  ++alloc_obj_count;
+  return calloc(1, sizeof(Obj));
+}
+
+Type *alloc_type(void) {
+  ++alloc_type_count;
+  return calloc(1, sizeof(Type));
+}

third_party/chibicc/chibicc.c

@@ -141,7 +141,16 @@ static char *quote_makefile(char *s) {
 static void PrintMemoryUsage(void) {
   struct mallinfo mi;
   mi = mallinfo();
+  fprintf(stderr, "\n");
   fprintf(stderr, "allocated %,ld bytes of memory\n", mi.arena);
+  fprintf(stderr, "allocated %,ld nodes (%,ld bytes)\n", alloc_node_count,
+          sizeof(Node) * alloc_node_count);
+  fprintf(stderr, "allocated %,ld tokens (%,ld bytes)\n", alloc_token_count,
+          sizeof(Token) * alloc_token_count);
+  fprintf(stderr, "allocated %,ld objs (%,ld bytes)\n", alloc_obj_count,
+          sizeof(Obj) * alloc_obj_count);
+  fprintf(stderr, "allocated %,ld types (%,ld bytes)\n", alloc_type_count,
+          sizeof(Type) * alloc_type_count);
 }
 
 static void strarray_push_comma(StringArray *a, char *s) {

third_party/chibicc/chibicc.h

@@ -62,14 +62,14 @@ void strarray_push(StringArray *, char *);
 // tokenize.c
 //
 
-// Token
 typedef enum {
-  TK_RESERVED,  // Keywords or punctuators
+  TK_IDENT,    // Identifiers
-  TK_IDENT,     // Identifiers
+  TK_PUNCT,    // Punctuators
-  TK_STR,       // String literals
+  TK_KEYWORD,  // Keywords
-  TK_NUM,       // Numeric literals
+  TK_STR,      // String literals
-  TK_PP_NUM,    // Preprocessing numbers
+  TK_NUM,      // Numeric literals
-  TK_EOF,       // End-of-file markers
+  TK_PP_NUM,   // Preprocessing numbers
+  TK_EOF,      // End-of-file markers
 } TokenKind;
 
 struct File {
@@ -81,15 +81,14 @@ struct File {
   int line_delta;
 };
 
-// Token type
+struct thatispacked Token {
-struct Token {
+  Token *next;       // Next token
-  TokenKind kind;    // Token kind
   int len;           // Token length
   int line_no;       // Line number
   int line_delta;    // Line number
+  TokenKind kind;    // Token kind
   bool at_bol;       // True if this token is at beginning of line
   bool has_space;    // True if this token follows a space character
-  Token *next;       // Next token
   char *loc;         // Token location
   Type *ty;          // Used if TK_NUM or TK_STR
   File *file;        // Source location
@@ -518,7 +517,7 @@ int encode_utf8(char *, uint32_t);
 uint32_t decode_utf8(char **, char *);
 bool is_ident1(uint32_t);
 bool is_ident2(uint32_t);
-int str_width(char *, int);
+int display_width(char *, int);
 
 //
 // hashmap.c
@@ -564,6 +563,20 @@ extern bool opt_sse4;
 extern bool opt_verbose;
 extern char *base_file;
 
+//
+// alloc.c
+//
+
+extern long alloc_node_count;
+extern long alloc_token_count;
+extern long alloc_obj_count;
+extern long alloc_type_count;
+
+Node *alloc_node(void);
+Token *alloc_token(void);
+Obj *alloc_obj(void);
+Type *alloc_type(void);
+
 COSMOPOLITAN_C_END_
 #endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
 #endif /* COSMOPOLITAN_THIRD_PARTY_CHIBICC_CHIBICC_H_ */

third_party/chibicc/codegen.c

@@ -2219,7 +2219,7 @@ static void emit_data(Obj *prog) {
     int align = (var->ty->kind == TY_ARRAY && var->ty->size >= 16)
                     ? MAX(16, var->align)
                     : var->align;
-    if (opt_common && var->is_tentative && !var->is_tls) {
+    if (opt_common && var->is_tentative) {
       println("\t.comm\t%s,%d,%d", nameof(var), var->ty->size, align);
     } else {
       if (var->section) {
@@ -2410,10 +2410,10 @@ static void emit_text(Obj *prog) {
     // Emit code
     gen_stmt(fn->body);
     assert(!depth);
-    // The C spec defines a special rule for the main function.
+    // [https://www.sigbus.info/n1570#5.1.2.2.3p1] The C spec defines
-    // Reaching the end of the main function is equivalent to
+    // a special rule for the main function. Reaching the end of the
-    // returning 0, even though the behavior is undefined for the
+    // main function is equivalent to returning 0, even though the
-    // other functions. See C11 5.1.2.2.3.
+    // behavior is undefined for the other functions.
     if (strcmp(nameof(fn), "main") == 0) {
       emitlin("\txor\t%eax,%eax");
     }

third_party/chibicc/hashmap.c

@@ -2,7 +2,10 @@
 
 #include "third_party/chibicc/chibicc.h"
 
-#define TOMBSTONE ((void *)-1)  // Represents a deleted hash entry
+#define INIT_SIZE      16            // initial hash bucket size
+#define LOW_WATERMARK  50            // keep usage below 50% after rehashing
+#define HIGH_WATERMARK 70            // perform rehash when usage exceeds 70%
+#define TOMBSTONE      ((void *)-1)  // represents deleted hash table entry
 
 static uint64_t fnv_hash(char *s, int len) {
   uint64_t hash = 0xcbf29ce484222325;
@@ -24,7 +27,8 @@ static void rehash(HashMap *map) {
     }
   }
   size_t cap = map->capacity;
-  while ((nkeys * 100) / cap >= 50) cap = cap * 2;
+  while ((nkeys * 100) / cap >= LOW_WATERMARK) cap = cap * 2;
+  assert(cap > 0);
   // Create a new hashmap and copy all key-values.
   HashMap map2 = {};
   map2.buckets = calloc(cap, sizeof(HashEntry));
@@ -56,9 +60,11 @@ static HashEntry *get_entry(HashMap *map, char *key, int keylen) {
 
 static HashEntry *get_or_insert_entry(HashMap *map, char *key, int keylen) {
   if (!map->buckets) {
-    map->buckets = calloc((map->capacity = 16), sizeof(HashEntry));
+    map->buckets = calloc(INIT_SIZE, sizeof(HashEntry));
+    map->capacity = INIT_SIZE;
+  } else if ((map->used * 100) / map->capacity >= HIGH_WATERMARK) {
+    rehash(map);
   }
-  if ((map->used * 100) / map->capacity >= 70) rehash(map);
   uint64_t hash = fnv_hash(key, keylen);
   for (int i = 0; i < map->capacity; i++) {
     HashEntry *ent = &map->buckets[(hash + i) & (map->capacity - 1)];

third_party/chibicc/parse.c

@@ -25,21 +25,13 @@ typedef struct Scope Scope;
 // Scope for local variables, global variables, typedefs
 // or enum constants
 typedef struct {
-  char *name;
-  int depth;
   Obj *var;
   Type *type_def;
   Type *enum_ty;
   int enum_val;
 } VarScope;
 
-// Scope for struct, union or enum tags
+// Represents a block scope.
-typedef struct {
-  char *name;
-  int depth;
-  Type *ty;
-} TagScope;
-
 struct Scope {
   Scope *next;
   // C has two block scopes; one is for variables/typedefs and
@@ -103,10 +95,6 @@ static Obj *globals;
 
 static Scope *scope = &(Scope){};
 
-// scope_depth is incremented by one at the beginning of a block
-// scope and decremented by one at the end of a block scope.
-static int scope_depth;
-
 // Points to the function object the parser is currently parsing.
 static Obj *current_fn;
 
@@ -173,12 +161,10 @@ static void enter_scope(void) {
   Scope *sc = calloc(1, sizeof(Scope));
   sc->next = scope;
   scope = sc;
-  scope_depth++;
 }
 
 static void leave_scope(void) {
   scope = scope->next;
-  scope_depth--;
 }
 
 // Find a variable by name.
@@ -190,16 +176,16 @@ static VarScope *find_var(Token *tok) {
   return NULL;
 }
 
-static TagScope *find_tag(Token *tok) {
+static Type *find_tag(Token *tok) {
   for (Scope *sc = scope; sc; sc = sc->next) {
-    TagScope *sc2 = hashmap_get2(&sc->tags, tok->loc, tok->len);
+    Type *ty = hashmap_get2(&sc->tags, tok->loc, tok->len);
-    if (sc2) return sc2;
+    if (ty) return ty;
   }
   return NULL;
 }
 
 Node *new_node(NodeKind kind, Token *tok) {
-  Node *node = calloc(1, sizeof(Node));
+  Node *node = alloc_node();
   node->kind = kind;
   node->tok = tok;
   return node;
@@ -252,7 +238,7 @@ static Node *new_vla_ptr(Obj *var, Token *tok) {
 
 Node *new_cast(Node *expr, Type *ty) {
   add_type(expr);
-  Node *node = calloc(1, sizeof(Node));
+  Node *node = alloc_node();
   node->kind = ND_CAST;
   node->tok = expr->tok;
   node->lhs = expr;
@@ -262,8 +248,6 @@ Node *new_cast(Node *expr, Type *ty) {
 
 static VarScope *push_scope(char *name) {
   VarScope *sc = calloc(1, sizeof(VarScope));
-  sc->name = name;
-  sc->depth = scope_depth;
   hashmap_put(&scope->vars, name, sc);
   return sc;
 }
@@ -303,7 +287,7 @@ static Initializer *new_initializer(Type *ty, bool is_flexible) {
 }
 
 static Obj *new_var(char *name, Type *ty) {
-  Obj *var = calloc(1, sizeof(Obj));
+  Obj *var = alloc_obj();
   var->name = name;
   var->ty = ty;
   var->align = ty->align;
@@ -330,9 +314,7 @@ static Obj *new_gvar(char *name, Type *ty) {
 
 static char *new_unique_name(void) {
   static int id = 0;
-  char *buf = calloc(1, 20);
+  return xasprintf(".L..%d", id++);
-  sprintf(buf, ".L..%d", id++);
-  return buf;
 }
 
 static Obj *new_anon_gvar(Type *ty) {
@@ -360,11 +342,7 @@ static Type *find_typedef(Token *tok) {
 }
 
 static void push_tag_scope(Token *tok, Type *ty) {
-  TagScope *sc = calloc(1, sizeof(TagScope));
+  hashmap_put2(&scope->tags, tok->loc, tok->len, ty);
-  sc->name = strndup(tok->loc, tok->len);
-  sc->depth = scope_depth;
-  sc->ty = ty;
-  hashmap_put2(&scope->tags, tok->loc, tok->len, sc);
 }
 
 // Consumes token if equal to STR or __STR__.
@@ -599,9 +577,14 @@ static Token *thing_attributes(Token *tok, void *arg) {
   error_tok(tok, "unknown function attribute");
 }
 
-// typespec = typename typename*
+// declspec = ("void" | "_Bool" | "char" | "short" | "int" | "long"
-// typename = "void" | "_Bool" | "char" | "short" | "int" | "long"
+//             | "typedef" | "static" | "extern" | "inline"
-//            | struct-decl | union-decl | typedef-name
+//             | "_Thread_local" | "__thread"
+//             | "signed" | "unsigned"
+//             | struct-decl | union-decl | typedef-name
+//             | enum-specifier | typeof-specifier
+//             | "const" | "volatile" | "auto" | "register" | "restrict"
+//             | "__restrict" | "__restrict__" | "_Noreturn")+
 //
 // The order of typenames in a type-specifier doesn't matter. For
 // example, `int long static` means the same as `static long int`.
@@ -614,7 +597,7 @@ static Token *thing_attributes(Token *tok, void *arg) {
 // while keeping the "current" type object that the typenames up
 // until that point represent. When we reach a non-typename token,
 // we returns the current type object.
-static Type *typespec(Token **rest, Token *tok, VarAttr *attr) {
+static Type *declspec(Token **rest, Token *tok, VarAttr *attr) {
   // We use a single integer as counters for all typenames.
   // For example, bits 0 and 1 represents how many times we saw the
   // keyword "void" so far. With this, we can use a switch statement
@@ -851,7 +834,7 @@ static Token *static_assertion(Token *tok) {
 }
 
 // func-params = ("void" | param ("," param)* ("," "...")?)? ")"
-// param       = typespec declarator
+// param       = declspec declarator
 static Type *func_params(Token **rest, Token *tok, Type *ty) {
   if (EQUAL(tok, "void") && EQUAL(tok->next, ")")) {
     *rest = tok->next->next;
@@ -868,7 +851,7 @@ static Type *func_params(Token **rest, Token *tok, Type *ty) {
       skip(tok, ')');
       break;
     }
-    Type *ty2 = typespec(&tok, tok, NULL);
+    Type *ty2 = declspec(&tok, tok, NULL);
     ty2 = declarator(&tok, tok, ty2);
     Token *name = ty2->name;
     if (ty2->kind == TY_ARRAY) {
@@ -935,8 +918,8 @@ static Type *declarator(Token **rest, Token *tok, Type *ty) {
   ty = pointers(&tok, tok, ty);
   if (EQUAL(tok, "(")) {
     Token *start = tok;
-    Type ignore = {};
+    Type dummy = {};
-    declarator(&tok, tok->next, &ignore);
+    declarator(&tok, start->next, &dummy);
     tok = skip(tok, ')');
     ty = type_suffix(rest, tok, ty);
     ty = declarator(&tok, start->next, ty);
@@ -959,8 +942,8 @@ static Type *abstract_declarator(Token **rest, Token *tok, Type *ty) {
   ty = pointers(&tok, tok, ty);
   if (EQUAL(tok, "(")) {
     Token *start = tok;
-    Type ignore = {};
+    Type dummy = {};
-    abstract_declarator(&tok, tok->next, &ignore);
+    abstract_declarator(&tok, start->next, &dummy);
     tok = skip(tok, ')');
     ty = type_suffix(rest, tok, ty);
     return abstract_declarator(&tok, start->next, ty);
@@ -968,9 +951,9 @@ static Type *abstract_declarator(Token **rest, Token *tok, Type *ty) {
   return type_suffix(rest, tok, ty);
 }
 
-// type-name = typespec abstract-declarator
+// type-name = declspec abstract-declarator
 static Type *typename(Token **rest, Token *tok) {
-  Type *ty = typespec(&tok, tok, NULL);
+  Type *ty = declspec(&tok, tok, NULL);
   return abstract_declarator(rest, tok, ty);
 }
 
@@ -1003,11 +986,11 @@ static Type *enum_specifier(Token **rest, Token *tok) {
     tok = tok->next;
   }
   if (tag && !EQUAL(tok, "{")) {
-    TagScope *sc = find_tag(tag);
+    Type *ty = find_tag(tag);
-    if (!sc) error_tok(tag, "unknown enum type");
+    if (!ty) error_tok(tag, "unknown enum type");
-    if (sc->ty->kind != TY_ENUM) error_tok(tag, "not an enum tag");
+    if (ty->kind != TY_ENUM) error_tok(tag, "not an enum tag");
     *rest = tok;
-    return sc->ty;
+    return ty;
   }
   tok = skip(tok, '{');
   // Read an enum-list.
@@ -1070,8 +1053,8 @@ static Node *new_alloca(Node *sz) {
   return node;
 }
 
-// declaration = typespec (declarator ("=" expr)? ("," declarator ("="
+// declaration = declspec (declarator ("=" expr)?
-// expr)?)*)? ";"
+//                         ("," declarator ("=" expr)?)*)? ";"
 static Node *declaration(Token **rest, Token *tok, Type *basety,
                          VarAttr *attr) {
   Node head = {};
@@ -1363,9 +1346,11 @@ static void struct_initializer1(Token **rest, Token *tok, Initializer *init) {
 // struct-initializer2 = initializer ("," initializer)*
 static void struct_initializer2(Token **rest, Token *tok, Initializer *init,
                                 Member *mem) {
+  bool first = true;
   for (; mem && !is_end(tok); mem = mem->next) {
     Token *start = tok;
-    if (mem != init->ty->members) tok = skip(tok, ',');
+    if (!first) tok = skip(tok, ',');
+    first = false;
     if (EQUAL(tok, "[") || EQUAL(tok, ".")) {
       *rest = start;
       return;
@@ -1389,6 +1374,7 @@ static void union_initializer(Token **rest, Token *tok, Initializer *init) {
   init->mem = init->ty->members;
   if (EQUAL(tok, "{")) {
     initializer2(&tok, tok->next, init->children[0]);
+    CONSUME(&tok, tok, ",");
     *rest = skip(tok, '}');
   } else {
     initializer2(rest, tok, init->children[0]);
@@ -1769,7 +1755,7 @@ static Node *stmt(Token **rest, Token *tok) {
     brk_label = node->brk_label = new_unique_name();
     cont_label = node->cont_label = new_unique_name();
     if (is_typename(tok)) {
-      Type *basety = typespec(&tok, tok, NULL);
+      Type *basety = declspec(&tok, tok, NULL);
       node->init = declaration(&tok, tok, basety, NULL);
     } else {
       node->init = expr_stmt(&tok, tok);
@@ -1872,7 +1858,7 @@ static Node *compound_stmt(Token **rest, Token *tok) {
   while (!EQUAL(tok, "}")) {
     if (is_typename(tok) && !EQUAL(tok->next, ":")) {
       VarAttr attr = {};
-      Type *basety = typespec(&tok, tok, &attr);
+      Type *basety = declspec(&tok, tok, &attr);
       if (attr.is_typedef) {
         tok = parse_typedef(tok, basety);
         continue;
@@ -2565,30 +2551,14 @@ static Node *mul(Token **rest, Token *tok) {
   }
 }
 
-// compound-literal = initializer "}"
+// cast = "(" type-name ")" cast | unary
-static Node *compound_literal(Token **rest, Token *tok, Type *ty,
-                              Token *start) {
-  if (scope_depth == 0) {
-    Obj *var = new_anon_gvar(ty);
-    gvar_initializer(rest, tok, var);
-    return new_var_node(var, start);
-  }
-  Obj *var = new_lvar(new_unique_name(), ty);
-  Node *lhs = lvar_initializer(rest, tok, var);
-  Node *rhs = new_var_node(var, tok);
-  return new_binary(ND_COMMA, lhs, rhs, tok);
-}
-
-// cast = "(" type-name ")" "{" compound-literal
-//      | "(" type-name ")" cast
-//      | unary
 static Node *cast(Token **rest, Token *tok) {
   if (EQUAL(tok, "(") && is_typename(tok->next)) {
     Token *start = tok;
     Type *ty = typename(&tok, tok->next);
     tok = skip(tok, ')');
     // compound literal
-    if (EQUAL(tok, "{")) return compound_literal(rest, tok, ty, start);
+    if (EQUAL(tok, "{")) return unary(rest, start);
     // type cast
     Node *node = new_cast(cast(rest, tok), ty);
     node->tok = start;
@@ -2612,9 +2582,10 @@ static Node *unary(Token **rest, Token *tok) {
     return new_unary(ND_ADDR, lhs, tok);
   }
   if (EQUAL(tok, "*")) {
-    // [C18 6.5.3.2p4] This is an oddity in the C spec, but dereferencing
+    // [https://www.sigbus.info/n1570#6.5.3.2p4] This is an oddity
-    // a function shouldn't do anything. If foo is a function, `*foo`,
+    // in the C spec, but dereferencing a function shouldn't do
-    // `**foo` or `*****foo` are all equivalent to just `foo`.
+    // anything. If foo is a function, `*foo`, `**foo` or `*****foo`
+    // are all equivalent to just `foo`.
     Node *node = cast(rest, tok->next);
     add_type(node);
     if (node->ty->kind == TY_FUNC) return node;
@@ -2640,14 +2611,14 @@ static Node *unary(Token **rest, Token *tok) {
   return postfix(rest, tok);
 }
 
-// struct-members = (typespec declarator (","  declarator)* ";")*
+// struct-members = (declspec declarator (","  declarator)* ";")*
 static void struct_members(Token **rest, Token *tok, Type *ty) {
   Member head = {};
   Member *cur = &head;
   int idx = 0;
   while (!EQUAL(tok, "}")) {
     VarAttr attr = {};
-    Type *basety = typespec(&tok, tok, &attr);
+    Type *basety = declspec(&tok, tok, &attr);
     bool first = true;
     // Anonymous struct member
     if ((basety->kind == TY_STRUCT || basety->kind == TY_UNION) &&
@@ -2708,8 +2679,8 @@ static Type *struct_union_decl(Token **rest, Token *tok) {
   }
   if (tag && !EQUAL(tok, "{")) {
     *rest = tok;
-    TagScope *sc = find_tag(tag);
+    Type *ty2 = find_tag(tag);
-    if (sc) return sc->ty;
+    if (ty2) return ty2;
     ty->size = -1;
     push_tag_scope(tag, ty);
     return ty;
@@ -2721,10 +2692,10 @@ static Type *struct_union_decl(Token **rest, Token *tok) {
   if (tag) {
     // If this is a redefinition, overwrite a previous type.
     // Otherwise, register the struct type.
-    TagScope *sc = find_tag(tag);
+    Type *ty2 = hashmap_get2(&scope->tags, tag->loc, tag->len);
-    if (sc && sc->depth == scope_depth) {
+    if (ty2) {
-      *sc->ty = *ty;
+      *ty2 = *ty;
-      return sc->ty;
+      return ty2;
     }
     push_tag_scope(tag, ty);
   }
@@ -2837,7 +2808,8 @@ static Node *new_inc_dec(Node *node, Token *tok, int addend) {
                   node->ty);
 }
 
-// postfix = ident "(" func-args ")" postfix-tail*
+// postfix = "(" type-name ")" "{" initializer-list "}"
+//         | ident "(" func-args ")" postfix-tail*
 //         | primary postfix-tail*
 //
 // postfix-tail = "[" expr "]"
@@ -2847,6 +2819,21 @@ static Node *new_inc_dec(Node *node, Token *tok, int addend) {
 //              | "++"
 //              | "--"
 static Node *postfix(Token **rest, Token *tok) {
+  if (EQUAL(tok, "(") && is_typename(tok->next)) {
+    // Compound literal
+    Token *start = tok;
+    Type *ty = typename(&tok, tok->next);
+    tok = skip(tok, ')');
+    if (scope->next == NULL) {
+      Obj *var = new_anon_gvar(ty);
+      gvar_initializer(rest, tok, var);
+      return new_var_node(var, start);
+    }
+    Obj *var = new_lvar("", ty);
+    Node *lhs = lvar_initializer(rest, tok, var);
+    Node *rhs = new_var_node(var, tok);
+    return new_binary(ND_COMMA, lhs, rhs, start);
+  }
   Node *node = primary(&tok, tok);
   for (;;) {
     if (EQUAL(tok, "(")) {
@@ -2961,7 +2948,7 @@ static Node *generic_selection(Token **rest, Token *tok) {
   return ret;
 }
 
-// primary = "(" "{" stmt stmt* "}" ")"
+// primary = "(" "{" stmt+ "}" ")"
 //         | "(" expr ")"
 //         | "sizeof" "(" type-name ")"
 //         | "sizeof" unary
@@ -3367,8 +3354,9 @@ static Token *function(Token *tok, Type *basety, VarAttr *attr) {
     fn->va_area = new_lvar("__va_area__", array_of(ty_char, 136));
   fn->alloca_bottom = new_lvar("__alloca_size__", pointer_to(ty_char));
   tok = skip(tok, '{');
-  // [C18 6.4.2.2] "__func__" is automatically defined as a
+  // [https://www.sigbus.info/n1570#6.4.2.2p1] "__func__" is
-  // local variable containing the current function name.
+  // automatically defined as a local variable containing the
+  // current function name.
   push_scope("__func__")->var =
       new_string_literal(fn->name, array_of(ty_char, strlen(fn->name) + 1));
   // [GNU] __FUNCTION__ is yet another name of __func__.
@@ -3401,7 +3389,7 @@ static Token *global_variable(Token *tok, Type *basety, VarAttr *attr) {
     if (attr->align) var->align = attr->align;
     if (EQUAL(tok, "=")) {
       gvar_initializer(&tok, tok->next, var);
-    } else if (!attr->is_extern) {
+    } else if (!attr->is_extern && !attr->is_tls) {
       var->is_tentative = true;
     }
   }
@@ -3537,7 +3525,7 @@ Obj *parse(Token *tok) {
     }
     VarAttr attr = {};
     tok = attribute_list(tok, &attr, thing_attributes);
-    Type *basety = typespec(&tok, tok, &attr);
+    Type *basety = declspec(&tok, tok, &attr);
     if (attr.is_typedef) {
       tok = parse_typedef(tok, basety);
       continue;

third_party/chibicc/preprocess.c

@@ -96,7 +96,7 @@ static Token *skip_line(Token *tok) {
 }
 
 static Token *copy_token(Token *tok) {
-  Token *t = calloc(1, sizeof(Token));
+  Token *t = alloc_token();
   *t = *tok;
   t->next = NULL;
   return t;
@@ -234,9 +234,8 @@ static Token *copy_line(Token **rest, Token *tok) {
 }
 
 static Token *new_num_token(int val, Token *tmpl) {
-  char buf[30];
+  char *buf = xasprintf("%d\n", val);
-  sprintf(buf, "%d\n", val);
+  return tokenize(new_file(tmpl->file->name, tmpl->file->file_no, buf));
-  return tokenize(new_file(tmpl->file->name, tmpl->file->file_no, strdup(buf)));
 }
 
 static Token *read_const_expr(Token **rest, Token *tok) {
@@ -270,10 +269,10 @@ static long eval_const_expr(Token **rest, Token *tok) {
   Token *expr = read_const_expr(rest, tok->next);
   expr = preprocess2(expr);
   if (expr->kind == TK_EOF) error_tok(start, "no expression");
-  // [C18 6.10.1.4] The standard requires we replace remaining
+  // [https://www.sigbus.info/n1570#6.10.1p4] The standard requires
-  // non-macro identifiers with "0" before evaluating a constant
+  // we replace remaining non-macro identifiers with "0" before
-  // expression. For example, `#if foo` is equivalent to `#if 0`
+  // evaluating a constant expression. For example, `#if foo` is
-  // if foo is not defined.
+  // equivalent to `#if 0` if foo is not defined.
   for (Token *t = expr; t->kind != TK_EOF; t = t->next) {
     if (t->kind == TK_IDENT) {
       Token *next = t->next;
@@ -453,8 +452,7 @@ static Token *stringize(Token *hash, Token *arg) {
 // Concatenate two tokens to create a new token.
 static Token *paste(Token *lhs, Token *rhs) {
   // Paste the two tokens.
-  char *buf = calloc(1, lhs->len + rhs->len + 1);
+  char *buf = xasprintf("%.*s%.*s", lhs->len, lhs->loc, rhs->len, rhs->loc);
-  sprintf(buf, "%.*s%.*s", lhs->len, lhs->loc, rhs->len, rhs->loc);
   // Tokenize the resulting string.
   Token *tok = tokenize(new_file(lhs->file->name, lhs->file->file_no, buf));
   if (tok->next->kind != TK_EOF)
@@ -706,7 +704,7 @@ static char *detect_include_guard(Token *tok) {
   return NULL;
 }
 
-static Token *include_file(Token *tok, char *path) {
+static Token *include_file(Token *tok, char *path, Token *filename_tok) {
   // Check for "#pragma once"
   if (hashmap_get(&pragma_once, path)) return tok;
   // If we read the same file before, and if the file was guarded
@@ -716,7 +714,8 @@ static Token *include_file(Token *tok, char *path) {
   char *guard_name = hashmap_get(&include_guards, path);
   if (guard_name && hashmap_get(&macros, guard_name)) return tok;
   Token *tok2 = tokenize_file(path);
-  if (!tok2) error_tok(tok, "%s: cannot open file: %s", path, strerror(errno));
+  if (!tok2)
+    error_tok(filename_tok, "%s: cannot open file: %s", path, strerror(errno));
   guard_name = detect_include_guard(tok2);
   if (guard_name) hashmap_put(&include_guards, path, guard_name);
   return append(tok2, tok);
@@ -760,19 +759,19 @@ static Token *preprocess2(Token *tok) {
         char *path =
             xasprintf("%s/%s", dirname(strdup(start->file->name)), filename);
         if (fileexists(path)) {
-          tok = include_file(tok, path);
+          tok = include_file(tok, path, start->next->next);
           continue;
         }
       }
       char *path = search_include_paths(filename);
-      tok = include_file(tok, path ? path : filename);
+      tok = include_file(tok, path ? path : filename, start->next->next);
       continue;
     }
     if (EQUAL(tok, "include_next")) {
       bool ignore;
       char *filename = read_include_filename(&tok, tok->next, &ignore);
       char *path = search_include_next(filename);
-      tok = include_file(tok, path ? path : filename);
+      tok = include_file(tok, path ? path : filename, start->next->next);
       continue;
     }
     if (EQUAL(tok, "define")) {
@@ -914,17 +913,13 @@ static char *format_date(struct tm *tm) {
       "Jan", "Feb", "Mar", "Apr", "May", "Jun",
       "Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
   };
-  char buf[30];
+  return xasprintf("\"%s %2d %d\"", mon[tm->tm_mon], tm->tm_mday,
-  sprintf(buf, "\"%s %2d %d\"", mon[tm->tm_mon], tm->tm_mday,
+                   tm->tm_year + 1900);
-          tm->tm_year + 1900);
-  return strdup(buf);
 }
 
 // __TIME__ is expanded to the current time, e.g. "13:34:03".
 static char *format_time(struct tm *tm) {
-  char buf[30];
+  return xasprintf("\"%02d:%02d:%02d\"", tm->tm_hour, tm->tm_min, tm->tm_sec);
-  sprintf(buf, "\"%02d:%02d:%02d\"", tm->tm_hour, tm->tm_min, tm->tm_sec);
-  return strdup(buf);
 }
 
 void init_macros(void) {
@@ -1302,11 +1297,11 @@ static void join_adjacent_string_literals(Token *tok) {
   }
   // Second pass: concatenate adjacent string literals.
   for (Token *tok1 = tok; tok1->kind != TK_EOF;) {
-    Token *tok2 = tok1->next;
+    if (tok1->kind != TK_STR || tok1->next->kind != TK_STR) {
-    if (tok1->kind != TK_STR || tok2->kind != TK_STR) {
       tok1 = tok1->next;
       continue;
     }
+#if 0
     assert(tok1->ty->base->size == tok2->ty->base->size);
     Token *t = copy_token(tok1);
     t->ty =
@@ -1317,6 +1312,25 @@ static void join_adjacent_string_literals(Token *tok) {
            tok2->str, tok2->ty->size);
     t->len = strlen(t->loc);
     *tok1 = *t;
+#else
+    Token *tok2 = tok1->next;
+    while (tok2->kind == TK_STR) tok2 = tok2->next;
+    int len = tok1->ty->array_len;
+    for (Token *t = tok1->next; t != tok2; t = t->next) {
+      len = len + t->ty->array_len - 1;
+    }
+    char *buf = calloc(tok1->ty->base->size, len);
+    int i = 0;
+    for (Token *t = tok1; t != tok2; t = t->next) {
+      memcpy(buf + i, t->str, t->ty->size);
+      i = i + t->ty->size - t->ty->base->size;
+    }
+    *tok1 = *copy_token(tok1);
+    tok1->ty = array_of(tok1->ty->base, len);
+    tok1->str = buf;
+    tok1->next = tok2;
+    tok1 = tok2;
+#endif
   }
 }
 

third_party/chibicc/test/hog_test.c

@@ -1,4 +0,0 @@
-main(void) {
-  void *p;
-  p = "hello";
-}

third_party/chibicc/test/initializer_test.c

@@ -445,6 +445,15 @@ int main() {
            };
            x.a;
          }));
+  ASSERT(1, ({
+           union {
+             int a;
+             char b;
+           } x = {
+               1,
+           };
+           x.a;
+         }));
   ASSERT(2, ({
            enum {
              x,

third_party/chibicc/test/struct_test.c

@@ -392,6 +392,24 @@ int main() {
            } x = {1}, y = {2};
            (0 ? x : y).a;
          }));
+  ASSERT(2, ({
+           struct {
+             int a;
+           } x = {1}, y = {2};
+           (x = y).a;
+         }));
+  ASSERT(1, ({
+           struct {
+             int a;
+           } x = {1}, y = {2};
+           (1 ? x : y).a;
+         }));
+  ASSERT(2, ({
+           struct {
+             int a;
+           } x = {1}, y = {2};
+           (0 ? x : y).a;
+         }));
 
   return 0;
 }

third_party/chibicc/tokenize.c

@@ -38,7 +38,7 @@ static void verror_at(char *filename, char *input, int line_no, char *loc,
   int indent = fprintf(stderr, "%s:%d: ", filename, line_no);
   fprintf(stderr, "%.*s\n", (int)(end - line), line);
   // Show the error message.
-  int pos = str_width(line, loc - line) + indent;
+  int pos = display_width(line, loc - line) + indent;
   fprintf(stderr, "%*s", pos, "");  // print pos spaces.
   fprintf(stderr, "^ ");
   vfprintf(stderr, fmt, ap);
@@ -53,6 +53,7 @@ void error_at(char *loc, char *fmt, ...) {
   va_list ap;
   va_start(ap, fmt);
   verror_at(current_file->name, current_file->contents, line_no, loc, fmt, ap);
+  va_end(ap);
   exit(1);
 }
 
@@ -64,7 +65,7 @@ void error_tok(Token *tok, char *fmt, ...) {
     verror_at(t->file->name, t->file->contents, t->line_no, t->loc, fmt, ap);
     va_end(ap);
   }
-  va_end(va);
+  va_end(ap);
   exit(1);
 }
 
@@ -73,6 +74,7 @@ void warn_tok(Token *tok, char *fmt, ...) {
   va_start(ap, fmt);
   verror_at(tok->file->name, tok->file->contents, tok->line_no, tok->loc, fmt,
             ap);
+  va_end(ap);
 }
 
 static int is_space(int c) {
@@ -103,9 +105,9 @@ Token *skip(Token *tok, char op) {
   }
 }
 
-// Create a new token and add it as the next token of `cur`.
+// Create a new token.
 static Token *new_token(TokenKind kind, char *start, char *end) {
-  Token *tok = calloc(1, sizeof(Token));
+  Token *tok = alloc_token();
   tok->kind = kind;
   tok->loc = start;
   tok->len = end - start;
@@ -117,18 +119,17 @@ static Token *new_token(TokenKind kind, char *start, char *end) {
   return tok;
 }
 
-// Read an identifier and returns a pointer pointing to the end
+// Read an identifier and returns the length of it.
-// of an identifier.
+// If p does not point to a valid identifier, 0 is returned.
-//
+static int read_ident(char *start) {
-// Returns null if p does not point to a valid identifier.
+  char *p = start;
-static char *read_ident(char *p) {
   uint32_t c = decode_utf8(&p, p);
-  if (!is_ident1(c)) return NULL;
+  if (!is_ident1(c)) return 0;
   for (;;) {
     char *q;
     c = decode_utf8(&q, p);
     if (!('a' <= c && c <= 'f') && !is_ident2(c)) {
-      return p;
+      return p - start;
     }
     p = q;
   }
@@ -140,6 +141,19 @@ static int from_hex(char c) {
   return c - 'A' + 10;
 }
 
+// Read a punctuator token from p and returns its length.
+static int read_punct(char *p) {
+  static char *kw[] = {"<<=", ">>=", "...", "==", "!=", "<=", ">=", "->",
+                       "+=",  "-=",  "*=",  "/=", "++", "--", "%=", "&=",
+                       "|=",  "^=",  "&&",  "||", "<<", ">>", "##"};
+  for (int i = 0; i < sizeof(kw) / sizeof(*kw); i++) {
+    if (startswith(p, kw[i])) {
+      return strlen(kw[i]);
+    }
+  }
+  return ispunct(*p) ? 1 : 0;
+}
+
 static bool is_keyword(Token *tok) {
   static HashMap map;
   if (map.capacity == 0) {
@@ -190,6 +204,17 @@ static int read_escaped_char(char **new_pos, char *p) {
     return c;
   }
   *new_pos = p + 1;
+  // Escape sequences are defined using themselves here. E.g.
+  // '\n' is implemented using '\n'. This tautological definition
+  // works because the compiler that compiles our compiler knows
+  // what '\n' actually is. In other words, we "inherit" the ASCII
+  // code of '\n' from the compiler that compiles our compiler,
+  // so we don't have to teach the actual code here.
+  //
+  // This fact has huge implications not only for the correctness
+  // of the compiler but also for the security of the generated code.
+  // For more info, read "Reflections on Trusting Trust" by Ken Thompson.
+  // https://github.com/rui314/chibicc/wiki/thompson1984.pdf
   switch (*p) {
     case 'a':
       return '\a';
@@ -217,7 +242,7 @@ static int read_escaped_char(char **new_pos, char *p) {
 static char *string_literal_end(char *p) {
   char *start = p;
   for (; *p != '"'; p++) {
-    if (*p == '\0') error_at(start, "unclosed string literal");
+    if (*p == '\n' || *p == '\0') error_at(start, "unclosed string literal");
     if (*p == '\\') p++;
   }
   return p;
@@ -225,7 +250,7 @@ static char *string_literal_end(char *p) {
 
 static Token *read_string_literal(char *start, char *quote) {
   char *end = string_literal_end(quote + 1);
-  char *buf = calloc(1, end - quote);
+  char *buf = calloc(2, end - quote);
   int len = 0;
   for (char *p = quote + 1; p < end;) {
     if (*p == '\\')
@@ -409,7 +434,7 @@ static void convert_pp_number(Token *tok) {
 void convert_pp_tokens(Token *tok) {
   for (Token *t = tok; t->kind != TK_EOF; t = t->next) {
     if (is_keyword(t))
-      t->kind = TK_RESERVED;
+      t->kind = TK_KEYWORD;
     else if (t->kind == TK_PP_NUM)
       convert_pp_number(t);
   }
@@ -546,34 +571,17 @@ Token *tokenize(File *file) {
       continue;
     }
     // Identifier or keyword
-    char *q;
+    int ident_len = read_ident(p);
-    if ((q = read_ident(p)) != NULL) {
+    if (ident_len) {
-      cur = cur->next = new_token(TK_IDENT, p, q);
+      cur = cur->next = new_token(TK_IDENT, p, p + ident_len);
-      p = q;
+      p += cur->len;
       continue;
     }
-    // Three-letter punctuators
+    // Punctuators
-    if (LOOKINGAT(p, "<<=") || LOOKINGAT(p, ">>=") || LOOKINGAT(p, "...")) {
+    int punct_len = read_punct(p);
-      cur = cur->next = new_token(TK_RESERVED, p, p + 3);
+    if (punct_len) {
-      p += 3;
+      cur = cur->next = new_token(TK_PUNCT, p, p + punct_len);
-      continue;
+      p += cur->len;
-    }
-    // Two-letter punctuators
-    if (LOOKINGAT(p, "==") || LOOKINGAT(p, "!=") || LOOKINGAT(p, "<=") ||
-        LOOKINGAT(p, ">=") || LOOKINGAT(p, "->") || LOOKINGAT(p, "+=") ||
-        LOOKINGAT(p, "-=") || LOOKINGAT(p, "*=") || LOOKINGAT(p, "/=") ||
-        LOOKINGAT(p, "++") || LOOKINGAT(p, "--") || LOOKINGAT(p, "%=") ||
-        LOOKINGAT(p, "&=") || LOOKINGAT(p, "|=") || LOOKINGAT(p, "^=") ||
-        LOOKINGAT(p, "&&") || LOOKINGAT(p, "||") || LOOKINGAT(p, "<<") ||
-        LOOKINGAT(p, ">>") || LOOKINGAT(p, "##")) {
-      cur = cur->next = new_token(TK_RESERVED, p, p + 2);
-      p += 2;
-      continue;
-    }
-    // Single-letter punctuators
-    if (ispunct(*p)) {
-      cur = cur->next = new_token(TK_RESERVED, p, p + 1);
-      p++;
       continue;
     }
     error_at(p, "invalid token");
@@ -665,6 +673,7 @@ static void remove_backslash_newline(char *p) {
       p[j++] = p[i++];
     }
   }
+  for (; n > 0; n--) p[j++] = '\n';
   p[j] = '\0';
 }
 
@@ -710,6 +719,11 @@ static void convert_universal_chars(char *p) {
 Token *tokenize_file(char *path) {
   char *p = read_file(path);
   if (!p) return NULL;
+  // UTF-8 texts may start with a 3-byte "BOM" marker sequence.
+  // If exists, just skip them because they are useless bytes.
+  // (It is actually not recommended to add BOM markers to UTF-8
+  // texts, but it's not uncommon particularly on Windows.)
+  if (!memcmp(p, "\xef\xbb\xbf", 3)) p += 3;
   canonicalize_newline(p);
   remove_backslash_newline(p);
   convert_universal_chars(p);

third_party/chibicc/type.c

@@ -18,7 +18,7 @@ Type ty_double[1] = {{TY_DOUBLE, 8, 8}};
 Type ty_ldouble[1] = {{TY_LDOUBLE, 16, 16}};
 
 static Type *new_type(TypeKind kind, int size, int align) {
-  Type *ty = calloc(1, sizeof(Type));
+  Type *ty = alloc_type();
   ty->kind = kind;
   ty->size = size;
   ty->align = align;
@@ -77,7 +77,7 @@ bool is_compatible(Type *t1, Type *t2) {
 }
 
 Type *copy_type(Type *ty) {
-  Type *ret = calloc(1, sizeof(Type));
+  Type *ret = alloc_type();
   *ret = *ty;
   ret->origin = ty;
   return ret;

third_party/chibicc/unicode.c

@@ -66,9 +66,9 @@ static bool in_range(uint32_t *range, uint32_t c) {
   return false;
 }
 
-// C11 allows not only ASCII but some multibyte characters in certan
+// [https://www.sigbus.info/n1570#D] C11 allows not only ASCII but
-// Unicode ranges to be used in an identifier. See C11 Annex D for the
+// some multibyte characters in certan Unicode ranges to be used in an
-// details.
+// identifier.
 //
 // This function returns true if a given character is acceptable as
 // the first character of an identifier.
@@ -108,7 +108,7 @@ bool is_ident2(uint32_t c) {
 
 // Returns the number of columns needed to display a given
 // string in a fixed-width font.
-int str_width(char *p, int len) {
+int display_width(char *p, int len) {
   char *start = p;
   int w = 0;
   while (p - start < len) {