chibicc/preprocess.c

// This file implements the C preprocessor.
//
// The preprocessor takes a list of tokens as an input and returns a
// new list of tokens as an output.
//
// The preprocessing language is designed in such a way that that's
// guaranteed to stop even if there is a recursive macro.
// Informally speaking, a macro is applied only once for each token.
// That is, if a macro token T appears in a result of direct or
// indirect macro expansion of T, T won't be expanded any further.
// For example, if T is defined as U, and U is defined as T, then
// token T is expanded to U and then to T and the macro expansion
// stops at that point.
//
// To achieve the above behavior, we attach for each token a set of
// macro names from which the token is expanded. The set is called
// "hideset". Hideset is initially empty, and every time we expand a
// macro, the macro name is added to the resulting tokens' hidesets.
//
// The above macro expansion algorithm is explained in this document
// written by Dave Prossor, which is used as a basis for the
// standard's wording:
// https://github.com/rui314/chibicc/wiki/cpp.algo.pdf

#include "chibicc.h"

typedef struct MacroParam MacroParam;
struct MacroParam {
  MacroParam *next;
  char *name;
};

typedef struct MacroArg MacroArg;
struct MacroArg {
  MacroArg *next;
  char *name;
  Token *tok;
};

typedef struct Macro Macro;
struct Macro {
  Macro *next;
  char *name;
  bool is_objlike; // Object-like or function-like
  MacroParam *params;
  Token *body;
  bool deleted;
};

// `#if` can be nested, so we use a stack to manage nested `#if`s.
typedef struct CondIncl CondIncl;
struct CondIncl {
  CondIncl *next;
  enum { IN_THEN, IN_ELIF, IN_ELSE } ctx;
  Token *tok;
  bool included;
};

typedef struct Hideset Hideset;
struct Hideset {
  Hideset *next;
  char *name;
};

static Macro *macros;
static CondIncl *cond_incl;

static Token *preprocess2(Token *tok);

static bool is_hash(Token *tok) {
  return tok->at_bol && equal(tok, "#");
}

// Some preprocessor directives such as #include allow extraneous
// tokens before newline. This function skips such tokens.
static Token *skip_line(Token *tok) {
  if (tok->at_bol)
    return tok;
  warn_tok(tok, "extra token");
  while (tok->at_bol)
    tok = tok->next;
  return tok;
}

static Token *copy_token(Token *tok) {
  Token *t = calloc(1, sizeof(Token));
  *t = *tok;
  t->next = NULL;
  return t;
}

static Token *new_eof(Token *tok) {
  Token *t = copy_token(tok);
  t->kind = TK_EOF;
  t->len = 0;
  return t;
}

static Hideset *new_hideset(char *name) {
  Hideset *hs = calloc(1, sizeof(Hideset));
  hs->name = name;
  return hs;
}

static Hideset *hideset_union(Hideset *hs1, Hideset *hs2) {
  Hideset head = {};
  Hideset *cur = &head;

  for (; hs1; hs1 = hs1->next)
    cur = cur->next = new_hideset(hs1->name);
  cur->next = hs2;
  return head.next;
}

static bool hideset_contains(Hideset *hs, char *s, int len) {
  for (; hs; hs = hs->next)
    if (strlen(hs->name) == len && !strncmp(hs->name, s, len))
      return true;
  return false;
}

static Hideset *hideset_intersection(Hideset *hs1, Hideset *hs2) {
  Hideset head = {};
  Hideset *cur = &head;

  for (; hs1; hs1 = hs1->next)
    if (hideset_contains(hs2, hs1->name, strlen(hs1->name)))
      cur = cur->next = new_hideset(hs1->name);
  return head.next;
}

static Token *add_hideset(Token *tok, Hideset *hs) {
  Token head = {};
  Token *cur = &head;

  for (; tok; tok = tok->next) {
    Token *t = copy_token(tok);
    t->hideset = hideset_union(t->hideset, hs);
    cur = cur->next = t;
  }
  return head.next;
}

// Append tok2 to the end of tok1.
static Token *append(Token *tok1, Token *tok2) {
  if (tok1->kind == TK_EOF)
    return tok2;

  Token head = {};
  Token *cur = &head;

  for (; tok1->kind != TK_EOF; tok1 = tok1->next)
    cur = cur->next = copy_token(tok1);
  cur->next = tok2;
  return head.next;
}

static Token *skip_cond_incl2(Token *tok) {
  while (tok->kind != TK_EOF) {
    if (is_hash(tok) &&
        (equal(tok->next, "if") || equal(tok->next, "ifdef") ||
         equal(tok->next, "ifndef"))) {
      tok = skip_cond_incl2(tok->next->next);
      continue;
    }
    if (is_hash(tok) && equal(tok->next, "endif"))
      return tok->next->next;
    tok = tok->next;
  }
  return tok;
}

// Skip until next `#else`, `#elif` or `#endif`.
// Nested `#if` and `#endif` are skipped.
static Token *skip_cond_incl(Token *tok) {
  while (tok->kind != TK_EOF) {
    if (is_hash(tok) &&
        (equal(tok->next, "if") || equal(tok->next, "ifdef") ||
         equal(tok->next, "ifndef"))) {
      tok = skip_cond_incl2(tok->next->next);
      continue;
    }

    if (is_hash(tok) &&
        (equal(tok->next, "elif") || equal(tok->next, "else") ||
         equal(tok->next, "endif")))
      break;
    tok = tok->next;
  }
  return tok;
}

// Copy all tokens until the next newline, terminate them with
// an EOF token and then returns them. This function is used to
// create a new list of tokens for `#if` arguments.
static Token *copy_line(Token **rest, Token *tok) {
  Token head = {};
  Token *cur = &head;

  for (; !tok->at_bol; tok = tok->next)
    cur = cur->next = copy_token(tok);

  cur->next = new_eof(tok);
  *rest = tok;
  return head.next;
}

// Read and evaluate a constant expression.
static long eval_const_expr(Token **rest, Token *tok) {
  Token *start = tok;
  Token *expr = copy_line(rest, tok->next);
  expr = preprocess2(expr);

  if (expr->kind == TK_EOF)
    error_tok(start, "no expression");

  Token *rest2;
  long val = const_expr(&rest2, expr);
  if (rest2->kind != TK_EOF)
    error_tok(rest2, "extra token");
  return val;
}

static CondIncl *push_cond_incl(Token *tok, bool included) {
  CondIncl *ci = calloc(1, sizeof(CondIncl));
  ci->next = cond_incl;
  ci->ctx = IN_THEN;
  ci->tok = tok;
  ci->included = included;
  cond_incl = ci;
  return ci;
}

static Macro *find_macro(Token *tok) {
  if (tok->kind != TK_IDENT)
    return NULL;

  for (Macro *m = macros; m; m = m->next)
    if (strlen(m->name) == tok->len && !strncmp(m->name, tok->loc, tok->len))
      return m->deleted ? NULL : m;
  return NULL;
}

static Macro *add_macro(char *name, bool is_objlike, Token *body) {
  Macro *m = calloc(1, sizeof(Macro));
  m->next = macros;
  m->name = name;
  m->is_objlike = is_objlike;
  m->body = body;
  macros = m;
  return m;
}

static MacroParam *read_macro_params(Token **rest, Token *tok) {
  MacroParam head = {};
  MacroParam *cur = &head;

  while (!equal(tok, ")")) {
    if (cur != &head)
      tok = skip(tok, ",");

    if (tok->kind != TK_IDENT)
      error_tok(tok, "expected an identifier");
    MacroParam *m = calloc(1, sizeof(MacroParam));
    m->name = strndup(tok->loc, tok->len);
    cur = cur->next = m;
    tok = tok->next;
  }
  *rest = tok->next;
  return head.next;
}

static void read_macro_definition(Token **rest, Token *tok) {
  if (tok->kind != TK_IDENT)
    error_tok(tok, "macro name must be an identifier");
  char *name = strndup(tok->loc, tok->len);
  tok = tok->next;

  if (!tok->has_space && equal(tok, "(")) {
    // Function-like macro
    MacroParam *params = read_macro_params(&tok, tok->next);
    Macro *m = add_macro(name, false, copy_line(rest, tok));
    m->params = params;
  } else {
    // Object-like macro
    add_macro(name, true, copy_line(rest, tok));
  }
}

static MacroArg *read_macro_arg_one(Token **rest, Token *tok) {
  Token head = {};
  Token *cur = &head;
  int level = 0;

  while (level > 0 || (!equal(tok, ",") && !equal(tok, ")"))) {
    if (tok->kind == TK_EOF)
      error_tok(tok, "premature end of input");

    if (equal(tok, "("))
      level++;
    else if (equal(tok, ")"))
      level--;

    cur = cur->next = copy_token(tok);
    tok = tok->next;
  }

  cur->next = new_eof(tok);

  MacroArg *arg = calloc(1, sizeof(MacroArg));
  arg->tok = head.next;
  *rest = tok;
  return arg;
}

static MacroArg *read_macro_args(Token **rest, Token *tok, MacroParam *params) {
  Token *start = tok;
  tok = tok->next->next;

  MacroArg head = {};
  MacroArg *cur = &head;

  MacroParam *pp = params;
  for (; pp; pp = pp->next) {
    if (cur != &head)
      tok = skip(tok, ",");
    cur = cur->next = read_macro_arg_one(&tok, tok);
    cur->name = pp->name;
  }

  if (pp)
    error_tok(start, "too many arguments");
  skip(tok, ")");
  *rest = tok;
  return head.next;
}

static MacroArg *find_arg(MacroArg *args, Token *tok) {
  for (MacroArg *ap = args; ap; ap = ap->next)
    if (tok->len == strlen(ap->name) && !strncmp(tok->loc, ap->name, tok->len))
      return ap;
  return NULL;
}

// Replace func-like macro parameters with given arguments.
static Token *subst(Token *tok, MacroArg *args) {
  Token head = {};
  Token *cur = &head;

  while (tok->kind != TK_EOF) {
    MacroArg *arg = find_arg(args, tok);

    // Handle a macro token. Macro arguments are completely macro-expanded
    // before they are substituted into a macro body.
    if (arg) {
      Token *t = preprocess2(arg->tok);
      for (; t->kind != TK_EOF; t = t->next)
        cur = cur->next = copy_token(t);
      tok = tok->next;
      continue;
    }

    // Handle a non-macro token.
    cur = cur->next = copy_token(tok);
    tok = tok->next;
    continue;
  }

  cur->next = tok;
  return head.next;
}

// If tok is a macro, expand it and return true.
// Otherwise, do nothing and return false.
static bool expand_macro(Token **rest, Token *tok) {
  if (hideset_contains(tok->hideset, tok->loc, tok->len))
    return false;

  Macro *m = find_macro(tok);
  if (!m)
    return false;

  // Object-like macro application
  if (m->is_objlike) {
    Hideset *hs = hideset_union(tok->hideset, new_hideset(m->name));
    Token *body = add_hideset(m->body, hs);
    *rest = append(body, tok->next);
    return true;
  }

  // If a funclike macro token is not followed by an argument list,
  // treat it as a normal identifier.
  if (!equal(tok->next, "("))
    return false;

  // Function-like macro application
  Token *macro_token = tok;
  MacroArg *args = read_macro_args(&tok, tok, m->params);
  Token *rparen = tok;

  // Tokens that consist a func-like macro invocation may have different
  // hidesets, and if that's the case, it's not clear what the hideset
  // for the new tokens should be. We take the interesection of the
  // macro token and the closing parenthesis and use it as a new hideset
  // as explained in the Dave Prossor's algorithm.
  Hideset *hs = hideset_intersection(macro_token->hideset, rparen->hideset);
  hs = hideset_union(hs, new_hideset(m->name));

  Token *body = subst(m->body, args);
  body = add_hideset(body, hs);
  *rest = append(body, tok->next);
  return true;
}

// Visit all tokens in `tok` while evaluating preprocessing
// macros and directives.
static Token *preprocess2(Token *tok) {
  Token head = {};
  Token *cur = &head;

  while (tok->kind != TK_EOF) {
    // If it is a macro, expand it.
    if (expand_macro(&tok, tok))
      continue;

    // Pass through if it is not a "#".
    if (!is_hash(tok)) {
      cur = cur->next = tok;
      tok = tok->next;
      continue;
    }

    Token *start = tok;
    tok = tok->next;

    if (equal(tok, "include")) {
      tok = tok->next;

      if (tok->kind != TK_STR)
        error_tok(tok, "expected a filename");

      char *path;
      if (tok->str[0] == '/')
        path = tok->str;
      else
        path = format("%s/%s", dirname(strdup(tok->file->name)), tok->str);

      Token *tok2 = tokenize_file(path);
      if (!tok2)
        error_tok(tok, "%s", strerror(errno));
      tok = skip_line(tok->next);
      tok = append(tok2, tok);
      continue;
    }

    if (equal(tok, "define")) {
      read_macro_definition(&tok, tok->next);
      continue;
    }

    if (equal(tok, "undef")) {
      tok = tok->next;
      if (tok->kind != TK_IDENT)
        error_tok(tok, "macro name must be an identifier");
      char *name = strndup(tok->loc, tok->len);
      tok = skip_line(tok->next);

      Macro *m = add_macro(name, true, NULL);
      m->deleted = true;
      continue;
    }

    if (equal(tok, "if")) {
      long val = eval_const_expr(&tok, tok);
      push_cond_incl(start, val);
      if (!val)
        tok = skip_cond_incl(tok);
      continue;
    }

    if (equal(tok, "ifdef")) {
      bool defined = find_macro(tok->next);
      push_cond_incl(tok, defined);
      tok = skip_line(tok->next->next);
      if (!defined)
        tok = skip_cond_incl(tok);
      continue;
    }

    if (equal(tok, "ifndef")) {
      bool defined = find_macro(tok->next);
      push_cond_incl(tok, !defined);
      tok = skip_line(tok->next->next);
      if (defined)
        tok = skip_cond_incl(tok);
      continue;
    }

    if (equal(tok, "elif")) {
      if (!cond_incl || cond_incl->ctx == IN_ELSE)
        error_tok(start, "stray #elif");
      cond_incl->ctx = IN_ELIF;

      if (!cond_incl->included && eval_const_expr(&tok, tok))
        cond_incl->included = true;
      else
        tok = skip_cond_incl(tok);
      continue;
    }

    if (equal(tok, "else")) {
      if (!cond_incl || cond_incl->ctx == IN_ELSE)
        error_tok(start, "stray #else");
      cond_incl->ctx = IN_ELSE;
      tok = skip_line(tok->next);

      if (cond_incl->included)
        tok = skip_cond_incl(tok);
      continue;
    }

    if (equal(tok, "endif")) {
      if (!cond_incl)
        error_tok(start, "stray #endif");
      cond_incl = cond_incl->next;
      tok = skip_line(tok->next);
      continue;
    }

    // `#`-only line is legal. It's called a null directive.
    if (tok->at_bol)
      continue;

    error_tok(tok, "invalid preprocessor directive");
  }

  cur->next = tok;
  return head.next;
}

// Entry point function of the preprocessor.
Token *preprocess(Token *tok) {
  tok = preprocess2(tok);
  if (cond_incl)
    error_tok(cond_incl->tok, "unterminated conditional directive");
  convert_keywords(tok);
  return tok;
}