input.cc

早期freebsd实现

// -*- C++ -*-
/* Copyright (C) 1989, 1990, 1991, 1992 Free Software Foundation, Inc.
     Written by James Clark ([e-mail])

This file is part of groff.

groff is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2, or (at your option) any later
version.

groff is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
for more details.

You should have received a copy of the GNU General Public License along
with groff; see the file COPYING.  If not, write to the Free Software
Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */

#include "troff.h"
#include "symbol.h"
#include "dictionary.h"
#include "hvunits.h"
#include "env.h"
#include "request.h"
#include "node.h"
#include "reg.h"
#include "token.h"
#include "div.h"
#include "charinfo.h"
#include "font.h"
#include "searchpath.h"
#include "macropath.h"
#include "defs.h"

// Needed for getpid().
#include "posix.h"

#define USAGE_EXIT_CODE 1
#define MACRO_PREFIX "tmac."
#define STARTUP_FILE "troffrc"
#define DEFAULT_INPUT_STACK_LIMIT 1000

#ifndef DEFAULT_WARNING_MASK
// warnings that are enabled by default
#define DEFAULT_WARNING_MASK \
     (WARN_CHAR|WARN_NUMBER|WARN_BREAK|WARN_SPACE|WARN_FONT)
#endif

// initial size of buffer for reading names; expanded as necessary
#define ABUF_SIZE 16

#ifdef COLUMN
void init_column_requests();
#endif /* COLUMN */

static node *read_draw_node();
void handle_first_page_transition();
static void push_token(const token &);
void copy_file();
#ifdef COLUMN
void vjustify();
#endif /* COLUMN */
void transparent();
void transparent_file();

const char *program_name = 0;
token tok;
int break_flag = 0;
static int backtrace_flag = 0;
char *pipe_command = 0;
charinfo *charset_table[256];

static int warning_mask = DEFAULT_WARNING_MASK;
static int inhibit_errors = 0;

static void enable_warning(const char *);
static void disable_warning(const char *);

static int escape_char = '\\';
static symbol end_macro_name;
static int compatible_flag = 0;
static void process_input_stack();
int ascii_output_flag = 0;
int suppress_output_flag = 0;

int tcommand_flag = 0;

static int get_copy(node**, int = 0);
static symbol read_escape_name();
static void interpolate_string(symbol);
static void interpolate_macro(symbol);
static void interpolate_number_format(symbol);
static void interpolate_environment_variable(symbol);

static void interpolate_arg(symbol);
static request_or_macro *lookup_request(symbol);
static int get_delim_number(units *, int);
static int get_delim_number(units *, int, units);
static int get_line_arg(units *res, int si, charinfo **cp);
static int read_size(int *);
static symbol get_delim_name();
static void init_registers();

struct input_iterator;
input_iterator *make_temp_iterator(const char *);
const char *input_char_description(int);

const int ESCAPE_QUESTION = 015;
const int BEGIN_TRAP = 016;
const int END_TRAP = 017;
const int PAGE_EJECTOR = 020;
const int ESCAPE_NEWLINE = 021;
const int ESCAPE_AMPERSAND = 022;
const int ESCAPE_UNDERSCORE = 023;
const int ESCAPE_BAR = 024;
const int ESCAPE_CIRCUMFLEX = 025;
const int ESCAPE_LEFT_BRACE = 026;
const int ESCAPE_RIGHT_BRACE = 027;
const int ESCAPE_LEFT_QUOTE = 030;
const int ESCAPE_RIGHT_QUOTE = 031;
const int ESCAPE_HYPHEN = 032;
const int ESCAPE_BANG = 033;
const int ESCAPE_c = 034;
const int ESCAPE_e = 035;
const int ESCAPE_PERCENT = 036;
const int ESCAPE_SPACE = 037;

const int TITLE_REQUEST = 0200;
const int COPY_FILE_REQUEST = 0201;
const int TRANSPARENT_FILE_REQUEST = 0202;
#ifdef COLUMN
const int VJUSTIFY_REQUEST = 0203;
#endif /* COLUMN */
const int ESCAPE_E = 0204;
const int LAST_PAGE_EJECTOR = 0205;
const int ESCAPE_RIGHT_PARENTHESIS = 0206;

void set_escape_char()
{
  if (has_arg()) {
    if (tok.ch() == 0) {
      error("bad escape character");
      escape_char = '\\';
    }
    else
      escape_char = tok.ch();
  }
  else
    escape_char = '\\';
  skip_line();
}

void escape_off()
{
  escape_char = 0;
  skip_line();
}

class input_iterator {
public:
  input_iterator();
  virtual ~input_iterator();
  int get(node **);
  friend class input_stack;
protected:
  const unsigned char *ptr;
  const unsigned char *eptr;
  input_iterator *next;
private:
  virtual int fill(node **);
  virtual int peek();
  virtual int has_args() { return 0; }
  virtual int nargs() { return 0; }
  virtual input_iterator *get_arg(int) { return NULL; }
  virtual int get_location(int, const char **,  int *)
    { return 0; }
  virtual void backtrace() {}
  virtual int set_location(const char *, int)
    { return 0; }
  virtual int next_file(FILE *, const char *) { return 0; }
  virtual void shift(int) {}
  virtual int is_boundary();
  virtual int internal_level() { return 0; }
  virtual int is_file() { return 0; }
};

input_iterator::input_iterator()
: ptr(0), eptr(0)
{
}

input_iterator::~input_iterator()
{
}

int input_iterator::fill(node **)
{
  return EOF;
}

int input_iterator::peek()
{
  return EOF;
}

int input_iterator::is_boundary()
{
  return 0;
}

inline int input_iterator::get(node **p)
{
  return ptr < eptr ? *ptr++ : fill(p);
}


class input_boundary : public input_iterator {
public:
  int is_boundary() { return 1; }
};

class file_iterator : public input_iterator {
  FILE *fp;
  int lineno;
  const char *filename;
  int newline_flag;
  enum { BUF_SIZE = 512 };
  unsigned char buf[BUF_SIZE];
public:
  file_iterator(FILE *, const char *);
  ~file_iterator();
  int fill(node **);
  int peek();
  int get_location(int, const char **, int *);
  void backtrace();
  int set_location(const char *, int);
  int next_file(FILE *, const char *);
  int is_file();
};

file_iterator::file_iterator(FILE *f, const char *fn)
: fp(f), filename(fn), lineno(1), newline_flag(0)
{
}

file_iterator::~file_iterator()
{
  if (fp != stdin)
    fclose(fp);
  else
    clearerr(stdin);
}

int file_iterator::is_file()
{
  return 1;
}

int file_iterator::next_file(FILE *f, const char *s)
{
  if (fp != stdin)
    fclose(fp);
  else
    clearerr(stdin);
  filename = s;
  fp = f;
  lineno = 1;
  newline_flag = 0;
  ptr = 0;
  eptr = 0;
  return 1;
}

int file_iterator::fill(node **)
{
  if (newline_flag)
    lineno++;
  newline_flag = 0;
  unsigned char *p = buf;
  ptr = p;
  unsigned char *e = p + BUF_SIZE;
  while (p < e) {
    int c = getc(fp);
    if (c == EOF)
      break;
    if (illegal_input_char(c))
      warning(WARN_INPUT, "illegal input character code %1", int(c));
    else {
      *p++ = c;
      if (c == '\n') {
	newline_flag = 1;
	break;
      }
    }
  }
  if (p > buf) {
    eptr = p;
    return *ptr++;
  }
  else {
    eptr = p;
    return EOF;
  }
}

int file_iterator::peek()
{
  int c = getc(fp);
  while (illegal_input_char(c)) {
    warning(WARN_INPUT, "illegal input character code %1", int(c));
    c = getc(fp);
  }
  if (c != EOF)
    ungetc(c, fp);
  return c;
}

int file_iterator::get_location(int /*allow_macro*/,
				const char **filenamep, int *linenop)
{
  *linenop = lineno;
  if (filename != 0 && strcmp(filename, "-") == 0)
    *filenamep = "<standard input>";
  else
    *filenamep = filename;
  return 1;
}

void file_iterator::backtrace()
{
  errprint("%1:%2: backtrace: file `%1'\n", filename, lineno);
}

int file_iterator::set_location(const char *f, int ln)
{
  if (f)
    filename = f;
  lineno = ln;
  return 1;
}

input_iterator nil_iterator;  

class input_stack {
public:
  static int get(node **);
  static int peek();
  static void push(input_iterator *);
  static input_iterator *get_arg(int);
  static int nargs();
  static int get_location(int, const char **, int *);
  static int set_location(const char *, int);
  static void backtrace();
  static void backtrace_all();
  static void next_file(FILE *, const char *);
  static void end_file();
  static void shift(int n);
  static void add_boundary();
  static void remove_boundary();
  static int get_level();
  static void clear();

  static int limit;
private:
  static input_iterator *top;
  static int level;

  static int finish_get(node **);
  static int finish_peek();
};

input_iterator *input_stack::top = &nil_iterator;
int input_stack::level = 0;
int input_stack::limit = DEFAULT_INPUT_STACK_LIMIT;

inline int input_stack::get_level()
{
  return level + top->internal_level();
}

inline int input_stack::get(node **np)
{
  return (top->ptr < top->eptr) ? *top->ptr++ : finish_get(np);
}

int input_stack::finish_get(node **np)
{
  for (;;) {
    int c = top->fill(np);
    if (c != EOF || top->is_boundary())
      return c;
    if (top == &nil_iterator)
      break;
    input_iterator *tem = top;
    top = top->next;
    level--;
    delete tem;
    if (top->ptr < top->eptr)
      return *top->ptr++;
  }
  assert(level == 0);
  return EOF;
}

inline int input_stack::peek()
{
  return (top->ptr < top->eptr) ? *top->ptr : finish_peek();
}

int input_stack::finish_peek()
{
  for (;;) {
    int c = top->peek();
    if (c != EOF || top->is_boundary())
      return c;
    if (top == &nil_iterator)
      break;
    input_iterator *tem = top;
    top = top->next;
    level--;
    delete tem;
    if (top->ptr < top->eptr)
      return *top->ptr;
  }
  assert(level == 0);
  return EOF;
}

void input_stack::add_boundary()
{
  push(new input_boundary);
}

void input_stack::remove_boundary()
{
  assert(top->is_boundary());
  input_iterator *temp = top->next;
  delete top;
  top = temp;
  level--;
}

void input_stack::push(input_iterator *in)
{
  if (in == 0)
    return;
  if (++level > limit && limit > 0)
    fatal("input stack limit exceeded (probable infinite loop)");
  in->next = top;
  top = in;
}

input_iterator *input_stack::get_arg(int i)
{
  input_iterator *p;
  for (p = top; p != NULL; p = p->next)
    if (p->has_args())
      return p->get_arg(i);
  return 0;
}

void input_stack::shift(int n)
{
  for (input_iterator *p = top; p; p = p->next)
    if (p->has_args()) {
      p->shift(n);
      return;
    }
}

int input_stack::nargs()
{
  for (input_iterator *p =top; p != 0; p = p->next)
    if (p->has_args())
      return p->nargs();
  return 0;
}

int input_stack::get_location(int allow_macro, const char **filenamep, int *linenop)
{
  for (input_iterator *p = top; p; p = p->next)
    if (p->get_location(allow_macro, filenamep, linenop))
      return 1;
  return 0;
}

void input_stack::backtrace()
{
  const char *f;
  int n;
  // only backtrace down to (not including) the topmost file
  for (input_iterator *p = top;
       p && !p->get_location(0, &f, &n);
       p = p->next)
    p->backtrace();
}

void input_stack::backtrace_all()
{
  for (input_iterator *p = top; p; p = p->next)
    p->backtrace();
}

int input_stack::set_location(const char *filename, int lineno)
{
  for (input_iterator *p = top; p; p = p->next)
    if (p->set_location(filename, lineno))
      return 1;
  return 0;
}

void input_stack::next_file(FILE *fp, const char *s)
{
  for (input_iterator **pp = &top; *pp != &nil_iterator; pp = &(*pp)->next)
    if ((*pp)->next_file(fp, s))
      return;
  if (++level > limit && limit > 0)
    fatal("input stack limit exceeded");
  *pp = new file_iterator(fp, s);
  (*pp)->next = &nil_iterator;
}

void input_stack::end_file()
{
  for (input_iterator **pp = &top; *pp != &nil_iterator; pp = &(*pp)->next)
    if ((*pp)->is_file()) {
      input_iterator *tem = *pp;
      *pp = (*pp)->next;
      delete tem;
      level--;
      return;
    }
}

void input_stack::clear()
{
  int nboundaries = 0;
  while (top != &nil_iterator) {
    if (top->is_boundary())
      nboundaries++;
    input_iterator *tem = top;
    top = top->next;
    level--;
    delete tem;
  }
  // Keep while_request happy.
  for (; nboundaries > 0; --nboundaries)
    add_boundary();
}

void backtrace_request()
{
  input_stack::backtrace_all();
  fflush(stderr);
  skip_line();
}

void next_file()
{
  symbol nm = get_long_name(0);
  while (!tok.newline() && !tok.eof())
    tok.next();
  if (nm.is_null())
    input_stack::end_file();
  else {
    errno = 0;
    FILE *fp = fopen(nm.contents(), "r");
    if (!fp)
      error("can't open `%1': %2", nm.contents(), strerror(errno));
    else
      input_stack::next_file(fp, nm.contents());
  }
  tok.next();
}

void shift()
{
  int n;
  if (!has_arg() || !get_integer(&n))
    n = 1;
  input_stack::shift(n);
  skip_line();
}

static int get_char_for_escape_name()
{
  int c = get_copy(NULL);
  switch (c) {
  case EOF:
    error("end of input in escape name");
    return '\0';
  default:
    if (!illegal_input_char(c))
      break;
    // fall through
  case ' ':
  case '\n':
  case '\t':
  case '\001':
  case '\b':
    error("%1 is not allowed in an escape name", input_char_description(c));
    return '\0';
  }
  return c;
}

static symbol read_two_char_escape_name()
{
  char buf[3];
  buf[0] = get_char_for_escape_name();
  if (buf[0] != '\0') {
    buf[1] = get_char_for_escape_name();
    if (buf[1] == '\0')
      buf[0] = 0;
    else
      buf[2] = 0;
  }
  return symbol(buf);
}

static symbol read_long_escape_name()
{
  int start_level = input_stack::get_level();
  char abuf[ABUF_SIZE];
  char *buf = abuf;
  int buf_size = ABUF_SIZE;
  int i = 0;
  for (;;) {
    int c = get_char_for_escape_name();
    if (c == 0) {
      if (buf != abuf)
	a_delete buf;
      return NULL_SYMBOL;
    }
    if (i + 2 > buf_size) {
      if (buf == abuf) {
	buf = new char [ABUF_SIZE*2];
	memcpy(buf, abuf, buf_size);
	buf_size = ABUF_SIZE*2;
      }
      else {
	char *old_buf = buf;
	buf = new char[buf_size*2];
	memcpy(buf, old_buf, buf_size);
	buf_size *= 2;
	a_delete old_buf;
      }
    }
    if (c == ']' && input_stack::get_level() == start_level)
      break;
    buf[i++] = c;
  }
  buf[i] = 0;
  if (buf == abuf) {
    if (i == 0) {
      error("empty escape name");
      return NULL_SYMBOL;
    }
    return symbol(abuf);
  }
  else {
    symbol s(buf);
    a_delete buf;
    return s;
  }
}