* config/alpha/alpha.c (alpha_emit_floatuns): Emit missing barrier.

[gcc.git] / gcc / cpplex.c

/* CPP Library - lexical analysis.
   Copyright (C) 2000 Free Software Foundation, Inc.
   Contributed by Per Bothner, 1994-95.
   Based on CCCP program by Paul Rubin, June 1986
   Adapted to ANSI C, Richard Stallman, Jan 1987
   Broken out to separate file, Zack Weinberg, Mar 2000

This program 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.

This program 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 this program; if not, write to the Free Software
Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */

#include "config.h"
#include "system.h"
#include "intl.h"
#include "hashtab.h"
#include "cpplib.h"
#include "cpphash.h"

#define PEEKBUF(BUFFER, N) \
  ((BUFFER)->rlimit - (BUFFER)->cur > (N) ? (BUFFER)->cur[N] : EOF)
#define GETBUF(BUFFER) \
  ((BUFFER)->cur < (BUFFER)->rlimit ? *(BUFFER)->cur++ : EOF)
#define FORWARDBUF(BUFFER, N) ((BUFFER)->cur += (N))

#define PEEKN(N) PEEKBUF (CPP_BUFFER (pfile), N)
#define FORWARD(N) FORWARDBUF (CPP_BUFFER (pfile), (N))
#define GETC() GETBUF (CPP_BUFFER (pfile))
#define PEEKC() PEEKBUF (CPP_BUFFER (pfile), 0)

static void skip_block_comment	PARAMS ((cpp_reader *));
static void skip_line_comment	PARAMS ((cpp_reader *));
static int maybe_macroexpand	PARAMS ((cpp_reader *, long));
static int skip_comment		PARAMS ((cpp_reader *, int));
static int copy_comment		PARAMS ((cpp_reader *, int));
static void skip_string		PARAMS ((cpp_reader *, int));
static void parse_string	PARAMS ((cpp_reader *, int));
static U_CHAR *find_position	PARAMS ((U_CHAR *, U_CHAR *, unsigned long *));
static int null_cleanup		PARAMS ((cpp_buffer *, cpp_reader *));
static void null_warning        PARAMS ((cpp_reader *, unsigned int));

static void safe_fwrite		PARAMS ((cpp_reader *, const U_CHAR *,
					 size_t, FILE *));
static void output_line_command	PARAMS ((cpp_reader *, cpp_printer *));

/* Re-allocates PFILE->token_buffer so it will hold at least N more chars.  */

void
_cpp_grow_token_buffer (pfile, n)
     cpp_reader *pfile;
     long n;
{
  long old_written = CPP_WRITTEN (pfile);
  pfile->token_buffer_size = n + 2 * pfile->token_buffer_size;
  pfile->token_buffer = (U_CHAR *)
    xrealloc(pfile->token_buffer, pfile->token_buffer_size);
  CPP_SET_WRITTEN (pfile, old_written);
}

static int
null_cleanup (pbuf, pfile)
     cpp_buffer *pbuf ATTRIBUTE_UNUSED;
     cpp_reader *pfile ATTRIBUTE_UNUSED;
{
  return 0;
}

/* Allocate a new cpp_buffer for PFILE, and push it on the input buffer stack.
   If BUFFER != NULL, then use the LENGTH characters in BUFFER
   as the new input buffer.
   Return the new buffer, or NULL on failure.  */

cpp_buffer *
cpp_push_buffer (pfile, buffer, length)
     cpp_reader *pfile;
     const U_CHAR *buffer;
     long length;
{
  cpp_buffer *buf = CPP_BUFFER (pfile);
  cpp_buffer *new;
  if (++pfile->buffer_stack_depth == CPP_STACK_MAX)
    {
      cpp_fatal (pfile, "macro or `#include' recursion too deep");
      return NULL;
    }

  new = (cpp_buffer *) xcalloc (1, sizeof (cpp_buffer));

  new->if_stack = pfile->if_stack;
  new->cleanup = null_cleanup;
  new->buf = new->cur = buffer;
  new->rlimit = buffer + length;
  new->prev = buf;
  new->mark = NULL;
  new->line_base = NULL;

  CPP_BUFFER (pfile) = new;
  return new;
}

cpp_buffer *
cpp_pop_buffer (pfile)
     cpp_reader *pfile;
{
  cpp_buffer *buf = CPP_BUFFER (pfile);
  if (ACTIVE_MARK_P (pfile))
    cpp_ice (pfile, "mark active in cpp_pop_buffer");
  (*buf->cleanup) (buf, pfile);
  CPP_BUFFER (pfile) = CPP_PREV_BUFFER (buf);
  free (buf);
  pfile->buffer_stack_depth--;
  return CPP_BUFFER (pfile);
}

/* Deal with the annoying semantics of fwrite.  */
static void
safe_fwrite (pfile, buf, len, fp)
     cpp_reader *pfile;
     const U_CHAR *buf;
     size_t len;
     FILE *fp;
{
  size_t count;

  while (len)
    {
      count = fwrite (buf, 1, len, fp);
      if (count == 0)
	goto error;
      len -= count;
      buf += count;
    }
  return;

 error:
  cpp_notice_from_errno (pfile, CPP_OPTION (pfile, out_fname));
}

/* Notify the compiler proper that the current line number has jumped,
   or the current file name has changed.  */

static void
output_line_command (pfile, print)
     cpp_reader *pfile;
     cpp_printer *print;
{
  unsigned int line;
  cpp_buffer *ip;
  enum { same = 0, enter, leave, rname } change;
  static const char * const codes[] = { "", " 1", " 2", "" };

  if (CPP_OPTION (pfile, no_line_commands))
    return;

  ip = cpp_file_buffer (pfile);
  if (ip == NULL)
    return;
  line = CPP_BUF_LINE (ip);

  /* Determine whether the current filename has changed, and if so,
     how.  'nominal_fname' values are unique, so they can be compared
     by comparing pointers.  */
  if (ip->nominal_fname == print->last_fname)
    change = same;
  else
    {
      if (pfile->buffer_stack_depth == print->last_bsd)
	change = rname;
      else
	{
	  if (pfile->buffer_stack_depth > print->last_bsd)
	    change = enter;
	  else
	    change = leave;
	  print->last_bsd = pfile->buffer_stack_depth;
	}
      print->last_fname = ip->nominal_fname;
    }
  /* If the current file has not changed, we can output a few newlines
     instead if we want to increase the line number by a small amount.
     We cannot do this if print->lineno is zero, because that means we
     haven't output any line commands yet.  (The very first line
     command output is a `same_file' command.)  */
  if (change == same && print->lineno != 0
      && line >= print->lineno && line < print->lineno + 8)
    {
      while (line > print->lineno)
	{
	  putc ('\n', print->outf);
	  print->lineno++;
	}
      return;
    }

#ifndef NO_IMPLICIT_EXTERN_C
  if (CPP_OPTION (pfile, cplusplus))
    fprintf (print->outf, "# %u \"%s\"%s%s%s\n", line, ip->nominal_fname,
	     codes[change],
	     ip->system_header_p ? " 3" : "",
	     (ip->system_header_p == 2) ? " 4" : "");
  else
#endif
    fprintf (print->outf, "# %u \"%s\"%s%s\n", line, ip->nominal_fname,
	     codes[change],
	     ip->system_header_p ? " 3" : "");
  print->lineno = line;
}

/* Write the contents of the token_buffer to the output stream, and
   clear the token_buffer.  Also handles generating line commands and
   keeping track of file transitions.  */

void
cpp_output_tokens (pfile, print)
     cpp_reader *pfile;
     cpp_printer *print;
{
  if (CPP_WRITTEN (pfile) - print->written)
    {
      if (CPP_PWRITTEN (pfile)[-1] == '\n' && print->lineno)
	print->lineno++;
      safe_fwrite (pfile, pfile->token_buffer,
		   CPP_WRITTEN (pfile) - print->written, print->outf);
    }
  output_line_command (pfile, print);
  CPP_SET_WRITTEN (pfile, print->written);
}

/* Scan a string (which may have escape marks), perform macro expansion,
   and write the result to the token_buffer.  */

void
_cpp_expand_to_buffer (pfile, buf, length)
     cpp_reader *pfile;
     const U_CHAR *buf;
     int length;
{
  cpp_buffer *ip;
  enum cpp_ttype token;
  U_CHAR *buf1;

  if (length < 0)
    {
      cpp_ice (pfile, "length < 0 in cpp_expand_to_buffer");
      return;
    }

  /* Copy the buffer, because it might be in an unsafe place - for
     example, a sequence on the token_buffer, where the pointers will
     be invalidated if we enlarge the token_buffer.  */
  buf1 = alloca (length);
  memcpy (buf1, buf, length);

  /* Set up the input on the input stack.  */
  ip = cpp_push_buffer (pfile, buf1, length);
  if (ip == NULL)
    return;
  ip->has_escapes = 1;

  /* Scan the input, create the output.  */
  for (;;)
    {
      token = cpp_get_token (pfile);
      if (token == CPP_EOF)
	break;
      if (token == CPP_POP && CPP_BUFFER (pfile) == ip)
	{
	  cpp_pop_buffer (pfile);
	  break;
	}
    }
}

/* Scan until CPP_BUFFER (PFILE) is exhausted, discarding output.
   Then pop the buffer.  */

void
cpp_scan_buffer_nooutput (pfile)
     cpp_reader *pfile;
{
  cpp_buffer *buffer = CPP_BUFFER (pfile);
  enum cpp_ttype token;
  unsigned int old_written = CPP_WRITTEN (pfile);
  /* In no-output mode, we can ignore everything but directives.  */
  for (;;)
    {
      if (! pfile->only_seen_white)
	_cpp_skip_rest_of_line (pfile);
      token = cpp_get_token (pfile);
      if (token == CPP_EOF)
	break;
      if (token == CPP_POP && CPP_BUFFER (pfile) == buffer)
	{
	  cpp_pop_buffer (pfile);
	  break;
	}
    }
  CPP_SET_WRITTEN (pfile, old_written);
}

/* Scan until CPP_BUFFER (pfile) is exhausted, writing output to PRINT.
   Then pop the buffer.  */

void
cpp_scan_buffer (pfile, print)
     cpp_reader *pfile;
     cpp_printer *print;
{
  cpp_buffer *buffer = CPP_BUFFER (pfile);
  enum cpp_ttype token;

  for (;;)
    {
      token = cpp_get_token (pfile);
      if ((token == CPP_POP && !CPP_IS_MACRO_BUFFER (CPP_BUFFER (pfile)))
	  || token == CPP_EOF || token == CPP_VSPACE
	  /* XXX Temporary kluge - force flush after #include only */
	  || (token == CPP_DIRECTIVE
	      && CPP_BUFFER (pfile)->nominal_fname != print->last_fname))
	{
	  cpp_output_tokens (pfile, print);
	  if (token == CPP_EOF)
	    return;
	  if (token == CPP_POP && CPP_BUFFER (pfile) == buffer)
	    {
	      cpp_pop_buffer (pfile);
	      return;
	    }
	}
    }
}

/* Return the topmost cpp_buffer that corresponds to a file (not a macro).  */

cpp_buffer *
cpp_file_buffer (pfile)
     cpp_reader *pfile;
{
  cpp_buffer *ip;

  for (ip = CPP_BUFFER (pfile); ip; ip = CPP_PREV_BUFFER (ip))
    if (ip->ihash != NULL)
      return ip;
  return NULL;
}

/* Skip a C-style block comment.  We know it's a comment, and point is
   at the second character of the starter.  */
static void
skip_block_comment (pfile)
     cpp_reader *pfile;
{
  unsigned int line, col;
  const U_CHAR *limit, *cur;

  FORWARD(1);
  line = CPP_BUF_LINE (CPP_BUFFER (pfile));
  col = CPP_BUF_COL (CPP_BUFFER (pfile));
  limit = CPP_BUFFER (pfile)->rlimit;
  cur = CPP_BUFFER (pfile)->cur;

  while (cur < limit)
    {
      char c = *cur++;
      if (c == '\n' || c == '\r')
	{
	  /* \r cannot be a macro escape marker here. */
	  if (!ACTIVE_MARK_P (pfile))
	    CPP_BUMP_LINE_CUR (pfile, cur);
	}
      else if (c == '*')
	{
	  /* Check for teminator.  */
	  if (cur < limit && *cur == '/')
	    goto out;

	  /* Warn about comment starter embedded in comment.  */
	  if (cur[-2] == '/' && CPP_OPTION (pfile, warn_comments))
	    cpp_warning_with_line (pfile, CPP_BUFFER (pfile)->lineno,
				   cur - CPP_BUFFER (pfile)->line_base,
				   "'/*' within comment");
	}
    }

  cpp_error_with_line (pfile, line, col, "unterminated comment");
  cur--;
 out:
  CPP_BUFFER (pfile)->cur = cur + 1;
}

/* Skip a C++/Chill line comment.  We know it's a comment, and point
   is at the second character of the initiator.  */
static void
skip_line_comment (pfile)
     cpp_reader *pfile;
{
  FORWARD(1);
  for (;;)
    {
      int c = GETC ();

      /* We don't have to worry about EOF in here.  */
      if (c == '\n')
	{
	  /* Don't consider final '\n' to be part of comment.  */
	  FORWARD(-1);
	  return;
	}
      else if (c == '\r')
	{
	  /* \r cannot be a macro escape marker here. */
	  if (!ACTIVE_MARK_P (pfile))
	    CPP_BUMP_LINE (pfile);
	  if (CPP_OPTION (pfile, warn_comments))
	    cpp_warning (pfile, "backslash-newline within line comment");
	}
    }
}

/* Skip a comment - C, C++, or Chill style.  M is the first character
   of the comment marker.  If this really is a comment, skip to its
   end and return ' '.  If this is not a comment, return M (which will
   be '/' or '-').  */

static int
skip_comment (pfile, m)
     cpp_reader *pfile;
     int m;
{
  if (m == '/' && PEEKC() == '*')
    {
      skip_block_comment (pfile);
      return ' ';
    }
  else if (m == '/' && PEEKC() == '/')
    {
      if (CPP_BUFFER (pfile)->system_header_p)
	{
	  /* We silently allow C++ comments in system headers, irrespective
	     of conformance mode, because lots of busted systems do that
	     and trying to clean it up in fixincludes is a nightmare.  */
	  skip_line_comment (pfile);
	  return ' ';
	}
      else if (CPP_OPTION (pfile, cplusplus_comments))
	{
	  if (CPP_OPTION (pfile, c89)
	      && CPP_PEDANTIC (pfile)
	      && ! CPP_BUFFER (pfile)->warned_cplusplus_comments)
	    {
	      cpp_pedwarn (pfile,
			   "C++ style comments are not allowed in ISO C89");
	      cpp_pedwarn (pfile,
			   "(this will be reported only once per input file)");
	      CPP_BUFFER (pfile)->warned_cplusplus_comments = 1;
	    }
	  skip_line_comment (pfile);
	  return ' ';
	}
      else
	return m;
    }
  else if (m == '-' && PEEKC() == '-'
	   && CPP_OPTION (pfile, chill))
    {
      skip_line_comment (pfile);
      return ' ';
    }
  else
    return m;
}

/* Identical to skip_comment except that it copies the comment into the
   token_buffer.  This is used if !discard_comments.  */
static int
copy_comment (pfile, m)
     cpp_reader *pfile;
     int m;
{
  const U_CHAR *start = CPP_BUFFER (pfile)->cur;  /* XXX Layering violation */
  const U_CHAR *limit;

  if (skip_comment (pfile, m) == m)
    return m;

  limit = CPP_BUFFER (pfile)->cur;
  CPP_RESERVE (pfile, limit - start + 2);
  CPP_PUTC_Q (pfile, m);
  for (; start <= limit; start++)
    if (*start != '\r')
      CPP_PUTC_Q (pfile, *start);

  return ' ';
}

static void
null_warning (pfile, count)
     cpp_reader *pfile;
     unsigned int count;
{
  if (count == 1)
    cpp_warning (pfile, "embedded null character ignored");
  else
    cpp_warning (pfile, "embedded null characters ignored");
}

/* Skip whitespace \-newline and comments.  Does not macro-expand.  */

void
_cpp_skip_hspace (pfile)
     cpp_reader *pfile;
{
  unsigned int null_count = 0;
  int c;

  while (1)
    {
      c = GETC();
      if (c == EOF)
	goto out;
      else if (is_hspace(c))
	{
	  if ((c == '\f' || c == '\v') && CPP_PEDANTIC (pfile))
	    cpp_pedwarn (pfile, "%s in preprocessing directive",
			 c == '\f' ? "formfeed" : "vertical tab");
	  else if (c == '\0')
	    null_count++;
	}
      else if (c == '\r')
	{
	  /* \r is a backslash-newline marker if !has_escapes, and
	     a deletable-whitespace or no-reexpansion marker otherwise. */
	  if (CPP_BUFFER (pfile)->has_escapes)
	    {
	      if (PEEKC() == ' ')
		FORWARD(1);
	      else
		break;
	    }
	  else
	    CPP_BUMP_LINE (pfile);
	}
      else if (c == '/' || c == '-')
	{
	  c = skip_comment (pfile, c);
	  if (c  != ' ')
	    break;
	}
      else
	break;
    }
  FORWARD(-1);
 out:
  if (null_count)
    null_warning (pfile, null_count);
}

/* Read and discard the rest of the current line.  */

void
_cpp_skip_rest_of_line (pfile)
     cpp_reader *pfile;
{
  for (;;)
    {
      int c = GETC();
      switch (c)
	{
	case '\n':
	  FORWARD(-1);
	case EOF:
	  return;

	case '\r':
	  if (! CPP_BUFFER (pfile)->has_escapes)
	    CPP_BUMP_LINE (pfile);
	  break;
	  
	case '\'':
	case '\"':
	  skip_string (pfile, c);
	  break;

	case '/':
	case '-':
	  skip_comment (pfile, c);
	  break;

	case '\f':
	case '\v':
	  if (CPP_PEDANTIC (pfile))
	    cpp_pedwarn (pfile, "%s in preprocessing directive",
			 c == '\f' ? "formfeed" : "vertical tab");
	  break;

	}
    }
}

/* Parse an identifier starting with C.  */

void
_cpp_parse_name (pfile, c)
     cpp_reader *pfile;
     int c;
{
  for (;;)
  {
      if (! is_idchar(c))
      {
	  FORWARD (-1);
	  break;
      }

      if (c == '$' && CPP_PEDANTIC (pfile))
	cpp_pedwarn (pfile, "`$' in identifier");

      CPP_RESERVE(pfile, 2); /* One more for final NUL.  */
      CPP_PUTC_Q (pfile, c);
      c = GETC();
      if (c == EOF)
	break;
  }
  return;
}

/* Parse and skip over a string starting with C.  A single quoted
   string is treated like a double -- some programs (e.g., troff) are
   perverse this way.  (However, a single quoted string is not allowed
   to extend over multiple lines.)  */
static void
skip_string (pfile, c)
     cpp_reader *pfile;
     int c;
{
  unsigned int start_line, start_column;
  unsigned int null_count = 0;

  start_line = CPP_BUF_LINE (CPP_BUFFER (pfile));
  start_column = CPP_BUF_COL (CPP_BUFFER (pfile));
  while (1)
    {
      int cc = GETC();
      switch (cc)
	{
	case EOF:
	  cpp_error_with_line (pfile, start_line, start_column,
			       "unterminated string or character constant");
	  if (pfile->multiline_string_line != start_line
	      && pfile->multiline_string_line != 0)
	    cpp_error_with_line (pfile,
				 pfile->multiline_string_line, -1,
			 "possible real start of unterminated constant");
	  pfile->multiline_string_line = 0;
	  goto out;

	case '\0':
	  null_count++;
	  break;
	  
	case '\n':
	  CPP_BUMP_LINE (pfile);
	  /* In Fortran and assembly language, silently terminate
	     strings of either variety at end of line.  This is a
	     kludge around not knowing where comments are in these
	     languages.  */
	  if (CPP_OPTION (pfile, lang_fortran)
	      || CPP_OPTION (pfile, lang_asm))
	    {
	      FORWARD(-1);
	      goto out;
	    }
	  /* Character constants may not extend over multiple lines.
	     In Standard C, neither may strings.  We accept multiline
	     strings as an extension.  */
	  if (c == '\'')
	    {
	      cpp_error_with_line (pfile, start_line, start_column,
				   "unterminated character constant");
	      FORWARD(-1);
	      goto out;
	    }
	  if (CPP_PEDANTIC (pfile) && pfile->multiline_string_line == 0)
	    cpp_pedwarn_with_line (pfile, start_line, start_column,
				   "string constant runs past end of line");
	  if (pfile->multiline_string_line == 0)
	    pfile->multiline_string_line = start_line;
	  break;

	case '\r':
	  if (CPP_BUFFER (pfile)->has_escapes)
	    {
	      cpp_ice (pfile, "\\r escape inside string constant");
	      FORWARD(1);
	    }
	  else
	    /* Backslash newline is replaced by nothing at all.  */
	    CPP_BUMP_LINE (pfile);
	  break;

	case '\\':
	  FORWARD(1);
	  break;

	case '\"':
	case '\'':
	  if (cc == c)
	    goto out;
	  break;
	}
    }

 out:
  if (null_count == 1)
    cpp_warning (pfile, "null character in string or character constant");
  else if (null_count > 1)
    cpp_warning (pfile, "null characters in string or character constant");
}

/* Parse a string and copy it to the output.  */

static void
parse_string (pfile, c)
     cpp_reader *pfile;
     int c;
{
  const U_CHAR *start = CPP_BUFFER (pfile)->cur;  /* XXX Layering violation */
  const U_CHAR *limit;

  skip_string (pfile, c);

  limit = CPP_BUFFER (pfile)->cur;
  CPP_RESERVE (pfile, limit - start + 2);
  CPP_PUTC_Q (pfile, c);
  for (; start < limit; start++)
    if (*start != '\r')
      CPP_PUTC_Q (pfile, *start);
}

/* Read an assertion into the token buffer, converting to
   canonical form: `#predicate(a n swe r)'  The next non-whitespace
   character to read should be the first letter of the predicate.
   Returns 0 for syntax error, 1 for bare predicate, 2 for predicate
   with answer (see callers for why). In case of 0, an error has been
   printed. */
int
_cpp_parse_assertion (pfile)
     cpp_reader *pfile;
{
  int c, dropwhite;
  _cpp_skip_hspace (pfile);
  c = PEEKC();
  if (c == '\n')
    {
      cpp_error (pfile, "assertion without predicate");
      return 0;
    }
  else if (! is_idstart(c))
    {
      cpp_error (pfile, "assertion predicate is not an identifier");
      return 0;
    }
  CPP_PUTC(pfile, '#');
  FORWARD(1);
  _cpp_parse_name (pfile, c);

  c = PEEKC();
  if (c != '(')
    {
      if (is_hspace(c) || c == '\r')
	_cpp_skip_hspace (pfile);
      c = PEEKC();
    }
  if (c != '(')
    return 1;

  CPP_PUTC(pfile, '(');
  FORWARD(1);
  dropwhite = 1;
  while ((c = GETC()) != ')')
    {
      if (is_space(c))
	{
	  if (! dropwhite)
	    {
	      CPP_PUTC(pfile, ' ');
	      dropwhite = 1;
	    }
	}
      else if (c == '\n' || c == EOF)
	{
	  if (c == '\n') FORWARD(-1);
	  cpp_error (pfile, "un-terminated assertion answer");
	  return 0;
	}
      else if (c == '\r')
	/* \r cannot be a macro escape here. */
	CPP_BUMP_LINE (pfile);
      else
	{
	  CPP_PUTC (pfile, c);
	  dropwhite = 0;
	}
    }

  if (pfile->limit[-1] == ' ')
    pfile->limit[-1] = ')';
  else if (pfile->limit[-1] == '(')
    {
      cpp_error (pfile, "empty token sequence in assertion");
      return 0;
    }
  else
    CPP_PUTC (pfile, ')');

  return 2;
}

/* Get the next token, and add it to the text in pfile->token_buffer.
   Return the kind of token we got.  */

enum cpp_ttype
_cpp_lex_token (pfile)
     cpp_reader *pfile;
{
  register int c, c2;
  enum cpp_ttype token;

  if (CPP_BUFFER (pfile) == NULL)
    return CPP_EOF;

 get_next:
  c = GETC();
  switch (c)
    {
    case EOF:
      return CPP_EOF;

    case '/':
      if (PEEKC () == '=')
	goto op2;

    comment:
      if (CPP_OPTION (pfile, discard_comments))
	c = skip_comment (pfile, c);
      else
	c = copy_comment (pfile, c);
      if (c != ' ')
	goto randomchar;
	  
      /* Comments are equivalent to spaces.
	 For -traditional, a comment is equivalent to nothing.  */
      if (!CPP_OPTION (pfile, discard_comments))
	return CPP_COMMENT;
      else if (CPP_TRADITIONAL (pfile)
	       && ! is_space (PEEKC ()))
	{
	  if (pfile->parsing_define_directive)
	    return CPP_COMMENT;
	  else
	    goto get_next;
	}
      else
	{
	  CPP_PUTC (pfile, c);
	  return CPP_HSPACE;
	}

    case '#':
      CPP_PUTC (pfile, c);

    hash:
      if (pfile->parsing_if_directive)
	{
	  CPP_ADJUST_WRITTEN (pfile, -1);
	  if (_cpp_parse_assertion (pfile))
	    return CPP_ASSERTION;
	  return CPP_OTHER;
	}

      if (pfile->parsing_define_directive && ! CPP_TRADITIONAL (pfile))
	{
	  c2 = PEEKC ();
	  if (c2 == '#')
	    {
	      FORWARD (1);
	      CPP_PUTC (pfile, c2);
	    }
	  else if (c2 == '%' && PEEKN (1) == ':')
	    {
	      /* Digraph: "%:" == "#".  */
	      FORWARD (1);
	      CPP_RESERVE (pfile, 2);
	      CPP_PUTC_Q (pfile, c2);
	      CPP_PUTC_Q (pfile, GETC ());
	    }
	  else
	    return CPP_STRINGIZE;

	  return CPP_TOKPASTE;
	}

      if (!pfile->only_seen_white)
	return CPP_OTHER;

      /* Remove the "#" or "%:" from the token buffer.  */
      CPP_ADJUST_WRITTEN (pfile, (c == '#' ? -1 : -2));
      return CPP_DIRECTIVE;

    case '\"':
    case '\'':
      parse_string (pfile, c);
      return c == '\'' ? CPP_CHAR : CPP_STRING;

    case '$':
      if (!CPP_OPTION (pfile, dollars_in_ident))
	goto randomchar;
      goto letter;

    case ':':
      c2 = PEEKC ();
      /* Digraph: ":>" == "]".  */
      if (c2 == '>'
	  || (c2 == ':' && CPP_OPTION (pfile, cplusplus)))
	goto op2;
      goto randomchar;

    case '&':
    case '+':
    case '|':
      c2 = PEEKC ();
      if (c2 == c || c2 == '=')
	goto op2;
      goto randomchar;

    case '%':
      /* Digraphs: "%:" == "#", "%>" == "}".  */
      c2 = PEEKC ();
      if (c2 == ':')
	{
	  FORWARD (1);
	  CPP_RESERVE (pfile, 2);
	  CPP_PUTC_Q (pfile, c);
	  CPP_PUTC_Q (pfile, c2);
	  goto hash;
	}
      else if (c2 == '>')
	{
	  FORWARD (1);
	  CPP_RESERVE (pfile, 2);
	  CPP_PUTC_Q (pfile, c);
	  CPP_PUTC_Q (pfile, c2);
	  return CPP_RBRACE;
	}
      /* else fall through */

    case '*':
    case '!':
    case '=':
    case '^':
      if (PEEKC () == '=')
	goto op2;
      goto randomchar;

    case '-':
      c2 = PEEKC ();
      if (c2 == '-')
	{
	  if (CPP_OPTION (pfile, chill))
	    goto comment;  /* Chill style comment */
	  else
	    goto op2;
	}
      else if (c2 == '=')
	goto op2;
      else if (c2 == '>')
	{
	  if (CPP_OPTION (pfile, cplusplus) && PEEKN (1) == '*')
	    {
	      /* In C++, there's a ->* operator.  */
	      token = CPP_OTHER;
	      CPP_RESERVE (pfile, 4);
	      CPP_PUTC_Q (pfile, c);
	      CPP_PUTC_Q (pfile, GETC ());
	      CPP_PUTC_Q (pfile, GETC ());
	      return token;
	    }
	  goto op2;
	}
      goto randomchar;

    case '<':
      if (pfile->parsing_include_directive)
	{
	  for (;;)
	    {
	      CPP_PUTC (pfile, c);
	      if (c == '>')
		break;
	      c = GETC ();
	      if (c == '\n' || c == EOF)
		{
		  cpp_error (pfile,
			     "missing '>' in `#include <FILENAME>'");
		  break;
		}
	      else if (c == '\r')
		{
		  if (!CPP_BUFFER (pfile)->has_escapes)
		    {
		      /* Backslash newline is replaced by nothing. */
		      CPP_ADJUST_WRITTEN (pfile, -1);
		      CPP_BUMP_LINE (pfile);
		    }
		  else
		    {
		      /* We might conceivably get \r- or \r<space> in
			 here.  Just delete 'em. */
		      int d = GETC();
		      if (d != '-' && d != ' ')
			cpp_ice (pfile, "unrecognized escape \\r%c", d);
		      CPP_ADJUST_WRITTEN (pfile, -1);
		    }			  
		}
	    }
	  return CPP_STRING;
	}
      /* Digraphs: "<%" == "{", "<:" == "[".  */
      c2 = PEEKC ();
      if (c2 == '%')
	{
	  FORWARD (1);
	  CPP_RESERVE (pfile, 2);
	  CPP_PUTC_Q (pfile, c);
	  CPP_PUTC_Q (pfile, c2);
	  return CPP_LBRACE;
	}
      else if (c2 == ':')
	goto op2;
      /* else fall through */
    case '>':
      c2 = PEEKC ();
      if (c2 == '=')
	goto op2;
      /* GNU C++ supports MIN and MAX operators <? and >?.  */
      if (c2 != c && (!CPP_OPTION (pfile, cplusplus) || c2 != '?'))
	goto randomchar;
      FORWARD(1);
      CPP_RESERVE (pfile, 3);
      CPP_PUTC_Q (pfile, c);
      CPP_PUTC_Q (pfile, c2);
      if (PEEKC () == '=')
	CPP_PUTC_Q (pfile, GETC ());
      return CPP_OTHER;

    case '.':
      c2 = PEEKC ();
      if (ISDIGIT (c2))
	{
	  CPP_PUTC (pfile, c);
	  c = GETC ();
	  goto number;
	}

      /* In C++ there's a .* operator.  */
      if (CPP_OPTION (pfile, cplusplus) && c2 == '*')
	goto op2;

      if (c2 == '.' && PEEKN(1) == '.')
	{
	  CPP_RESERVE (pfile, 3);
	  CPP_PUTC_Q (pfile, '.');
	  CPP_PUTC_Q (pfile, '.');
	  CPP_PUTC_Q (pfile, '.');
	  FORWARD (2);
	  return CPP_3DOTS;
	}
      goto randomchar;

    op2:
      CPP_RESERVE (pfile, 2);
      CPP_PUTC_Q (pfile, c);
      CPP_PUTC_Q (pfile, GETC ());
      return CPP_OTHER;

    case 'L':
      c2 = PEEKC ();
      if ((c2 == '\'' || c2 == '\"') && !CPP_TRADITIONAL (pfile))
	{
	  CPP_PUTC (pfile, c);
	  c = GETC ();
	  parse_string (pfile, c);
	  return c == '\'' ? CPP_WCHAR : CPP_WSTRING;
	}
      goto letter;

    case '0': case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8': case '9':
    number:
    c2  = '.';
    for (;;)
      {
	CPP_RESERVE (pfile, 2);
	CPP_PUTC_Q (pfile, c);
	c = PEEKC ();
	if (c == EOF)
	  break;
	if (!is_numchar(c) && c != '.'
	    && ((c2 != 'e' && c2 != 'E'
		 && ((c2 != 'p' && c2 != 'P')
		     || CPP_OPTION (pfile, c89)))
		|| (c != '+' && c != '-')))
	  break;
	FORWARD(1);
	c2= c;
      }
    return CPP_NUMBER;
    case 'b': case 'c': case 'd': case 'h': case 'o':
    case 'B': case 'C': case 'D': case 'H': case 'O':
      if (CPP_OPTION (pfile, chill) && PEEKC () == '\'')
	{
	  CPP_RESERVE (pfile, 2);
	  CPP_PUTC_Q (pfile, c);
	  CPP_PUTC_Q (pfile, '\'');
	  FORWARD(1);
	  for (;;)
	    {
	      c = GETC();
	      if (c == EOF)
		goto chill_number_eof;
	      if (!is_numchar(c))
		break;
	      CPP_PUTC (pfile, c);
	    }
	  if (c == '\'')
	    {
	      CPP_RESERVE (pfile, 2);
	      CPP_PUTC_Q (pfile, c);
	      return CPP_STRING;
	    }
	  else
	    {
	      FORWARD(-1);
	    chill_number_eof:
	      return CPP_NUMBER;
	    }
	}
      else
	goto letter;
    case '_':
    case 'a': case 'e': case 'f': case 'g': case 'i': case 'j':
    case 'k': case 'l': case 'm': case 'n': case 'p': case 'q':
    case 'r': case 's': case 't': case 'u': case 'v': case 'w':
    case 'x': case 'y': case 'z':
    case 'A': case 'E': case 'F': case 'G': case 'I': case 'J':
    case 'K': case 'M': case 'N': case 'P': case 'Q': case 'R':
    case 'S': case 'T': case 'U': case 'V': case 'W': case 'X':
    case 'Y': case 'Z':
    letter:
    _cpp_parse_name (pfile, c);
    return CPP_MACRO;

    case ' ':  case '\t':  case '\v': case '\f': case '\0':
      {
	int null_count = 0;

	for (;;)
	  {
	    if (c == '\0')
	      null_count++;
	    else
	      CPP_PUTC (pfile, c);
	    c = PEEKC ();
	    if (c == EOF || !is_hspace(c))
	      break;
	    FORWARD(1);
	  }
	if (null_count)
	  null_warning (pfile, null_count);
	return CPP_HSPACE;
      }

    case '\r':
      if (CPP_BUFFER (pfile)->has_escapes)
	{
	  c = GETC ();
	  if (c == '-')
	    {
	      if (pfile->output_escapes)
		CPP_PUTS (pfile, "\r-", 2);
	      _cpp_parse_name (pfile, GETC ());
	      return CPP_NAME;
	    }
	  else if (c == ' ')
	    {
	      /* "\r " means a space, but only if necessary to prevent
		 accidental token concatenation.  */
	      CPP_RESERVE (pfile, 2);
	      if (pfile->output_escapes)
		CPP_PUTC_Q (pfile, '\r');
	      CPP_PUTC_Q (pfile, c);
	      return CPP_HSPACE;
	    }
	  else
	    {
	      cpp_ice (pfile, "unrecognized escape \\r%c", c);
	      goto get_next;
	    }
	}
      else
	{
	  /* Backslash newline is ignored. */
	  if (!ACTIVE_MARK_P (pfile))
	    CPP_BUMP_LINE (pfile);
	  goto get_next;
	}

    case '\n':
      CPP_PUTC (pfile, c);
      return CPP_VSPACE;

    case '(': token = CPP_LPAREN;    goto char1;
    case ')': token = CPP_RPAREN;    goto char1;
    case '{': token = CPP_LBRACE;    goto char1;
    case '}': token = CPP_RBRACE;    goto char1;
    case ',': token = CPP_COMMA;     goto char1;
    case ';': token = CPP_SEMICOLON; goto char1;

    randomchar:
    default:
      token = CPP_OTHER;
    char1:
      CPP_PUTC (pfile, c);
      return token;
    }
}

/* Check for and expand a macro, which is from WRITTEN to CPP_WRITTEN (pfile).
   Caller is expected to have checked no_macro_expand.  */
static int
maybe_macroexpand (pfile, written)
     cpp_reader *pfile;
     long written;
{
  U_CHAR *macro = pfile->token_buffer + written;
  size_t len = CPP_WRITTEN (pfile) - written;
  HASHNODE *hp = _cpp_lookup (pfile, macro, len);

  if (!hp)
    return 0;
  if (hp->type == T_DISABLED)
    {
      if (pfile->output_escapes)
	{
	  /* Insert a no-reexpand marker before IDENT.  */
	  CPP_RESERVE (pfile, 2);
	  CPP_ADJUST_WRITTEN (pfile, 2);
	  macro = pfile->token_buffer + written;

	  memmove (macro + 2, macro, len);
	  macro[0] = '\r';
	  macro[1] = '-';
	}
      return 0;
    }
  if (hp->type == T_EMPTY)
    {
      /* Special case optimization: macro expands to nothing.  */
      CPP_SET_WRITTEN (pfile, written);
      CPP_PUTC_Q (pfile, ' ');
      return 1;
    }

  /* If macro wants an arglist, verify that a '(' follows.  */
  if (hp->type == T_MACRO && hp->value.defn->nargs >= 0)
    {
      int macbuf_whitespace = 0;
      int c;

      while (CPP_IS_MACRO_BUFFER (CPP_BUFFER (pfile)))
	{
	  const U_CHAR *point = CPP_BUFFER (pfile)->cur;
	  for (;;)
	    {
	      _cpp_skip_hspace (pfile);
	      c = PEEKC ();
	      if (c == '\n')
		FORWARD(1);
	      else
		break;
	    }
	  if (point != CPP_BUFFER (pfile)->cur)
	    macbuf_whitespace = 1;
	  if (c == '(')
	    goto is_macro_call;
	  else if (c != EOF)
	    goto not_macro_call;
	  cpp_pop_buffer (pfile);
	}

      CPP_SET_MARK (pfile);
      for (;;)
	{
	  _cpp_skip_hspace (pfile);
	  c = PEEKC ();
	  if (c == '\n')
	    FORWARD(1);
	  else
	    break;
	}
      CPP_GOTO_MARK (pfile);

      if (c != '(')
	{
	not_macro_call:
	  if (macbuf_whitespace)
	    CPP_PUTC (pfile, ' ');
	  return 0;
	}
    }

 is_macro_call:
  /* This is now known to be a macro call.
     Expand the macro, reading arguments as needed,
     and push the expansion on the input stack.  */
  _cpp_macroexpand (pfile, hp);
  CPP_SET_WRITTEN (pfile, written);
  return 1;
}

enum cpp_ttype
cpp_get_token (pfile)
     cpp_reader *pfile;
{
  enum cpp_ttype token;
  long written = CPP_WRITTEN (pfile);

 get_next:
  token = _cpp_lex_token (pfile);

  switch (token)
    {
    default:
      pfile->potential_control_macro = 0;
      pfile->only_seen_white = 0;
      return token;

    case CPP_VSPACE:
      if (pfile->only_seen_white == 0)
	pfile->only_seen_white = 1;
      CPP_BUMP_LINE (pfile);
      return token;

    case CPP_HSPACE:
    case CPP_COMMENT:
      return token;

    case CPP_DIRECTIVE:
      pfile->potential_control_macro = 0;
      if (_cpp_handle_directive (pfile))
	return CPP_DIRECTIVE;
      pfile->only_seen_white = 0;
      CPP_PUTC (pfile, '#');
      return CPP_OTHER;

    case CPP_MACRO:
      pfile->potential_control_macro = 0;
      pfile->only_seen_white = 0;
      if (! pfile->no_macro_expand
	  && maybe_macroexpand (pfile, written))
	goto get_next;
      return CPP_NAME;

    case CPP_EOF:
      if (CPP_BUFFER (pfile) == NULL)
	return CPP_EOF;
      if (CPP_BUFFER (pfile)->manual_pop)
	/* If we've been reading from redirected input, the
	   frontend will pop the buffer.  */
	return CPP_EOF;

      if (CPP_BUFFER (pfile)->seen_eof)
	{
	  cpp_pop_buffer (pfile);
	  goto get_next;
	}
      else
	{
	  _cpp_handle_eof (pfile);
	  return CPP_POP;
	}
    }
}

/* Like cpp_get_token, but skip spaces and comments.  */

enum cpp_ttype
cpp_get_non_space_token (pfile)
     cpp_reader *pfile;
{
  int old_written = CPP_WRITTEN (pfile);
  for (;;)
    {
      enum cpp_ttype token = cpp_get_token (pfile);
      if (token != CPP_COMMENT && token != CPP_HSPACE && token != CPP_VSPACE)
	return token;
      CPP_SET_WRITTEN (pfile, old_written);
    }
}

/* Like cpp_get_token, except that it does not execute directives,
   does not consume vertical space, and automatically pops off macro
   buffers.

   XXX This function will exist only till collect_expansion doesn't
   need to see whitespace anymore, then it'll be merged with
   _cpp_get_directive_token (below).  */
enum cpp_ttype
_cpp_get_define_token (pfile)
     cpp_reader *pfile;
{
  long old_written;
  enum cpp_ttype token;

 get_next:
  old_written = CPP_WRITTEN (pfile);
  token = _cpp_lex_token (pfile);
  switch (token)
    {
    default:
      return token;

    case CPP_VSPACE:
      /* Put it back and return VSPACE.  */
      FORWARD(-1);
      CPP_ADJUST_WRITTEN (pfile, -1);
      return CPP_VSPACE;

    case CPP_HSPACE:
      if (CPP_PEDANTIC (pfile))
	{
	  U_CHAR *p, *limit;
	  p = pfile->token_buffer + old_written;
	  limit = CPP_PWRITTEN (pfile);
	  while (p < limit)
	    {
	      if (*p == '\v' || *p == '\f')
		cpp_pedwarn (pfile, "%s in preprocessing directive",
			     *p == '\f' ? "formfeed" : "vertical tab");
	      p++;
	    }
	}
      return CPP_HSPACE;

    case CPP_DIRECTIVE:
      /* Don't execute the directive, but don't smash it to OTHER either.  */
      CPP_PUTC (pfile, '#');
      return CPP_DIRECTIVE;

    case CPP_MACRO:
      if (! pfile->no_macro_expand
	  && maybe_macroexpand (pfile, old_written))
	goto get_next;
      return CPP_NAME;

    case CPP_EOF:
      if (CPP_IS_MACRO_BUFFER (CPP_BUFFER (pfile)))
	{
	  cpp_pop_buffer (pfile);
	  goto get_next;
	}
      else
	/* This can happen for files that don't end with a newline,
	   and for cpp_define and friends.  Pretend they do, so
	   callers don't have to deal.  A warning will be issued by
	   someone else, if necessary.  */
	return CPP_VSPACE;
    }
}

/* Just like _cpp_get_define_token except that it discards horizontal
   whitespace.  */

enum cpp_ttype
_cpp_get_directive_token (pfile)
     cpp_reader *pfile;
{
  int old_written = CPP_WRITTEN (pfile);
  for (;;)
    {
      enum cpp_ttype token = _cpp_get_define_token (pfile);
      if (token != CPP_COMMENT && token != CPP_HSPACE)
	return token;
      CPP_SET_WRITTEN (pfile, old_written);
    }
}

/* Determine the current line and column.  Used only by read_and_prescan. */
static U_CHAR *
find_position (start, limit, linep)
     U_CHAR *start;
     U_CHAR *limit;
     unsigned long *linep;
{
  unsigned long line = *linep;
  U_CHAR *lbase = start;
  while (start < limit)
    {
      U_CHAR ch = *start++;
      if (ch == '\n' || ch == '\r')
	{
	  line++;
	  lbase = start;
	}
    }
  *linep = line;
  return lbase;
}

/* The following table is used by _cpp_read_and_prescan.  If we have
   designated initializers, it can be constant data; otherwise, it is
   set up at runtime by _cpp_init_input_buffer.  */

#ifndef UCHAR_MAX
#define UCHAR_MAX 255	/* assume 8-bit bytes */
#endif

#if (GCC_VERSION >= 2007) || (__STDC_VERSION__ >= 199901L)
#define init_chartab()  /* nothing */
#define CHARTAB static const unsigned char chartab[UCHAR_MAX + 1] = {
#define END };
#define s(p, v) [p] = v,
#else
#define CHARTAB static unsigned char chartab[UCHAR_MAX + 1] = { 0 }; \
 static void init_chartab PARAMS ((void)) { \
 unsigned char *x = chartab;
#define END }
#define s(p, v) x[p] = v;
#endif

/* Table of characters that can't be handled in the inner loop.
   Also contains the mapping between trigraph third characters and their
   replacements.  */
#define SPECCASE_CR        1
#define SPECCASE_BACKSLASH 2
#define SPECCASE_QUESTION  3
 
CHARTAB
  s('\r', SPECCASE_CR)
  s('\\', SPECCASE_BACKSLASH)
  s('?',  SPECCASE_QUESTION)

  s('=', '#')	s(')', ']')	s('!', '|')
  s('(', '[')	s('\'', '^')	s('>', '}')
  s('/', '\\')	s('<', '{')	s('-', '~')
END

#undef CHARTAB
#undef END
#undef s

#define NORMAL(c) ((chartab[c]) == 0 || (chartab[c]) > SPECCASE_QUESTION)
#define NONTRI(c) ((c) <= SPECCASE_QUESTION)

/* Read the entire contents of file DESC into buffer BUF.  LEN is how
   much memory to allocate initially; more will be allocated if
   necessary.  Convert end-of-line markers (\n, \r, \r\n, \n\r) to
   canonical form (\n).  If enabled, convert and/or warn about
   trigraphs.  Convert backslash-newline to a one-character escape
   (\r) and remove it from "embarrassing" places (i.e. the middle of a
   token).  If there is no newline at the end of the file, add one and
   warn.  Returns -1 on failure, or the actual length of the data to
   be scanned.

   This function does a lot of work, and can be a serious performance
   bottleneck.  It has been tuned heavily; make sure you understand it
   before hacking.  The common case - no trigraphs, Unix style line
   breaks, backslash-newline set off by whitespace, newline at EOF -
   has been optimized at the expense of the others.  The performance
   penalty for DOS style line breaks (\r\n) is about 15%.
   
   Warnings lose particularly heavily since we have to determine the
   line number, which involves scanning from the beginning of the file
   or from the last warning.  The penalty for the absence of a newline
   at the end of reload1.c is about 60%.  (reload1.c is 329k.)

   If your file has more than one kind of end-of-line marker, you
   will get messed-up line numbering.
   
   So that the cases of the switch statement do not have to concern
   themselves with the complications of reading beyond the end of the
   buffer, the buffer is guaranteed to have at least 3 characters in
   it (or however many are left in the file, if less) on entry to the
   switch.  This is enough to handle trigraphs and the "\\\n\r" and
   "\\\r\n" cases.
   
   The end of the buffer is marked by a '\\', which, being a special
   character, guarantees we will exit the fast-scan loops and perform
   a refill. */
 
long
_cpp_read_and_prescan (pfile, fp, desc, len)
     cpp_reader *pfile;
     cpp_buffer *fp;
     int desc;
     size_t len;
{
  U_CHAR *buf = (U_CHAR *) xmalloc (len);
  U_CHAR *ip, *op, *line_base;
  U_CHAR *ibase;
  unsigned long line;
  unsigned int deferred_newlines;
  size_t offset;
  int count = 0;

  offset = 0;
  deferred_newlines = 0;
  op = buf;
  line_base = buf;
  line = 1;
  ibase = pfile->input_buffer + 3;
  ip = ibase;
  ip[-1] = '\0';  /* Guarantee no match with \n for SPECCASE_CR */

  for (;;)
    {
      U_CHAR *near_buff_end;

      count = read (desc, ibase, pfile->input_buffer_len);
      if (count < 0)
	goto error;
      
      ibase[count] = '\\';  /* Marks end of buffer */
      if (count)
	{
	  near_buff_end = pfile->input_buffer + count;
	  offset += count;
	  if (offset > len)
	    {
	      size_t delta_op;
	      size_t delta_line_base;
	      len = offset * 2;
	      if (offset > len)
		/* len overflowed.
		   This could happen if the file is larger than half the
		   maximum address space of the machine. */
		goto too_big;

	      delta_op = op - buf;
	      delta_line_base = line_base - buf;
	      buf = (U_CHAR *) xrealloc (buf, len);
	      op = buf + delta_op;
	      line_base = buf + delta_line_base;
	    }
	}
      else
	{
	  if (ip == ibase)
	    break;
	  /* Allow normal processing of the (at most 2) remaining
	     characters.  The end-of-buffer marker is still present
	     and prevents false matches within the switch. */
	  near_buff_end = ibase - 1;
	}

      for (;;)
	{
	  unsigned int span;

	  /* Deal with \-newline, potentially in the middle of a token. */
	  if (deferred_newlines)
	    {
	      if (op != buf && ! is_space (op[-1]) && op[-1] != '\r')
		{
		  /* Previous was not white space.  Skip to white
		     space, if we can, before outputting the \r's */
		  span = 0;
		  while (ip[span] != ' '
			 && ip[span] != '\t'
			 && ip[span] != '\n'
			 && NORMAL(ip[span]))
		    span++;
		  memcpy (op, ip, span);
		  op += span;
		  ip += span;
		  if (! NORMAL(ip[0]))
		    goto do_speccase;
		}
	      while (deferred_newlines)
		deferred_newlines--, *op++ = '\r';
	    }

	  /* Copy as much as we can without special treatment. */
	  span = 0;
	  while (NORMAL (ip[span])) span++;
	  memcpy (op, ip, span);
	  op += span;
	  ip += span;

	do_speccase:
	  if (ip > near_buff_end) /* Do we have enough chars? */
	    break;
	  switch (chartab[*ip++])
	    {
	    case SPECCASE_CR:  /* \r */
	      if (ip[-2] != '\n')
		{
		  if (*ip == '\n')
		    ip++;
		  *op++ = '\n';
		}
	      break;

	    case SPECCASE_BACKSLASH:  /* \ */
	      if (*ip == '\n')
		{
		  deferred_newlines++;
		  ip++;
		  if (*ip == '\r') ip++;
		}
	      else if (*ip == '\r')
		{
		  deferred_newlines++;
		  ip++;
		  if (*ip == '\n') ip++;
		}
	      else
		*op++ = '\\';
	      break;

	    case SPECCASE_QUESTION: /* ? */
	      {
		unsigned int d, t;

		*op++ = '?'; /* Normal non-trigraph case */
		if (ip[0] != '?')
		  break;
		    
		d = ip[1];
		t = chartab[d];
		if (NONTRI (t))
		  break;

		if (CPP_OPTION (pfile, warn_trigraphs))
		  {
		    unsigned long col;
		    line_base = find_position (line_base, op, &line);
		    col = op - line_base + 1;
		    if (CPP_OPTION (pfile, trigraphs))
		      cpp_warning_with_line (pfile, line, col,
					     "trigraph ??%c converted to %c", d, t);
		    else
		      cpp_warning_with_line (pfile, line, col,
					     "trigraph ??%c ignored", d);
		  }

		ip += 2;
		if (CPP_OPTION (pfile, trigraphs))
		  {
		    op[-1] = t;	    /* Overwrite '?' */
		    if (t == '\\')
		      {
			op--;
			*--ip = '\\';
			goto do_speccase; /* May need buffer refill */
		      }
		  }
		else
		  {
		    *op++ = '?';
		    *op++ = d;
		  }
	      }
	      break;
	    }
	}
      /* Copy previous char plus unprocessed (at most 2) chars
	 to beginning of buffer, refill it with another
	 read(), and continue processing */
      memmove (ip - count - 1, ip - 1, 4 - (ip - near_buff_end));
      ip -= count;
    }

  if (offset == 0)
    return 0;

  if (op[-1] != '\n')
    {
      unsigned long col;
      line_base = find_position (line_base, op, &line);
      col = op - line_base + 1;
      cpp_warning_with_line (pfile, line, col, "no newline at end of file");
      if (offset + 1 > len)
	{
	  len += 1;
	  if (offset + 1 > len)
	    goto too_big;
	  buf = (U_CHAR *) xrealloc (buf, len);
	  op = buf + offset;
	}
      *op++ = '\n';
    }

  fp->buf = ((len - offset < 20) ? buf : (U_CHAR *)xrealloc (buf, op - buf));
  return op - buf;

 too_big:
  cpp_notice (pfile, "%s is too large (>%lu bytes)", fp->ihash->name,
	      (unsigned long)offset);
  free (buf);
  return -1;

 error:
  cpp_error_from_errno (pfile, fp->ihash->name);
  free (buf);
  return -1;
}

/* Allocate pfile->input_buffer, and initialize chartab[]
   if it hasn't happened already.  */
 
void
_cpp_init_input_buffer (pfile)
     cpp_reader *pfile;
{
  U_CHAR *tmp;

  init_chartab ();

  /* Determine the appropriate size for the input buffer.  Normal C
     source files are smaller than eight K.  */
  /* 8Kbytes of buffer proper, 1 to detect running off the end without
     address arithmetic all the time, and 3 for pushback during buffer
     refill, in case there's a potential trigraph or end-of-line
     digraph at the end of a block. */

  tmp = (U_CHAR *) xmalloc (8192 + 1 + 3);
  pfile->input_buffer = tmp;
  pfile->input_buffer_len = 8192;
}