Implement per-port read options.

[bpt/guile.git] / libguile / read.c

/* Copyright (C) 1995, 1996, 1997, 1999, 2000, 2001, 2003, 2004, 2006,
 *   2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation, Inc.
 * 
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public License
 * as published by the Free Software Foundation; either version 3 of
 * the License, or (at your option) any later version.
 *
 * This library 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
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
 * 02110-1301 USA
 */


\f

#ifdef HAVE_CONFIG_H
# include <config.h>
#endif

#include <stdio.h>
#include <ctype.h>
#include <string.h>
#include <unistd.h>
#include <unicase.h>
#include <unictype.h>

#include "libguile/_scm.h"
#include "libguile/bytevectors.h"
#include "libguile/chars.h"
#include "libguile/eval.h"
#include "libguile/arrays.h"
#include "libguile/bitvectors.h"
#include "libguile/keywords.h"
#include "libguile/alist.h"
#include "libguile/srcprop.h"
#include "libguile/hashtab.h"
#include "libguile/hash.h"
#include "libguile/ports.h"
#include "libguile/fports.h"
#include "libguile/root.h"
#include "libguile/strings.h"
#include "libguile/strports.h"
#include "libguile/vectors.h"
#include "libguile/validate.h"
#include "libguile/srfi-4.h"
#include "libguile/srfi-13.h"

#include "libguile/read.h"
#include "libguile/private-options.h"


\f

SCM_GLOBAL_SYMBOL (scm_sym_dot, ".");
SCM_SYMBOL (scm_keyword_prefix, "prefix");
SCM_SYMBOL (scm_keyword_postfix, "postfix");
SCM_SYMBOL (sym_nil, "nil");

scm_t_option scm_read_opts[] = {
  { SCM_OPTION_BOOLEAN, "copy", 0,
    "Copy source code expressions." },
  { SCM_OPTION_BOOLEAN, "positions", 1,
    "Record positions of source code expressions." },
  { SCM_OPTION_BOOLEAN, "case-insensitive", 0,
    "Convert symbols to lower case."},
  { SCM_OPTION_SCM, "keywords", (scm_t_bits) SCM_BOOL_F_BITS,
    "Style of keyword recognition: #f, 'prefix or 'postfix."},
  { SCM_OPTION_BOOLEAN, "r6rs-hex-escapes", 0,
    "Use R6RS variable-length character and string hex escapes."},
  { SCM_OPTION_BOOLEAN, "square-brackets", 1,
    "Treat `[' and `]' as parentheses, for R6RS compatibility."},
  { SCM_OPTION_BOOLEAN, "hungry-eol-escapes", 0,
    "In strings, consume leading whitespace after an escaped end-of-line."},
  { 0, },
};
 
/* Internal read options structure.  This is initialized by 'scm_read'
   from the global and per-port read options, and a pointer is passed
   down to all helper functions. */

enum t_keyword_style
  {
    KEYWORD_STYLE_HASH_PREFIX,
    KEYWORD_STYLE_PREFIX,
    KEYWORD_STYLE_POSTFIX
  };

struct t_read_opts
{
  enum t_keyword_style keyword_style;
  unsigned int copy_source_p        : 1;
  unsigned int record_positions_p   : 1;
  unsigned int case_insensitive_p   : 1;
  unsigned int r6rs_escapes_p       : 1;
  unsigned int square_brackets_p    : 1;
  unsigned int hungry_eol_escapes_p : 1;
};

typedef struct t_read_opts scm_t_read_opts;


/*
  Give meaningful error messages for errors

  We use the format

  FILE:LINE:COL: MESSAGE
  This happened in ....

  This is not standard GNU format, but the test-suite likes the real
  message to be in front.

 */


void
scm_i_input_error (char const *function,
                   SCM port, const char *message, SCM arg)
{
  SCM fn = (scm_is_string (SCM_FILENAME(port))
            ? SCM_FILENAME(port)
            : scm_from_locale_string ("#<unknown port>"));

  SCM string_port = scm_open_output_string ();
  SCM string = SCM_EOL;
  scm_simple_format (string_port,
                     scm_from_locale_string ("~A:~S:~S: ~A"),
                     scm_list_4 (fn,
                                 scm_from_long (SCM_LINUM (port) + 1),
                                 scm_from_int (SCM_COL (port) + 1),
                                 scm_from_locale_string (message)));
    
  string = scm_get_output_string (string_port);
  scm_close_output_port (string_port);
  scm_error_scm (scm_from_latin1_symbol ("read-error"),
                 function? scm_from_locale_string (function) : SCM_BOOL_F,
                 string,
                 arg,
                 SCM_BOOL_F);
}


SCM_DEFINE (scm_read_options, "read-options-interface", 0, 1, 0, 
            (SCM setting),
            "Option interface for the read options. Instead of using\n"
            "this procedure directly, use the procedures @code{read-enable},\n"
            "@code{read-disable}, @code{read-set!} and @code{read-options}.")
#define FUNC_NAME s_scm_read_options
{
  SCM ans = scm_options (setting,
                         scm_read_opts,
                         FUNC_NAME);
  if (SCM_COPY_SOURCE_P)
    SCM_RECORD_POSITIONS_P = 1;
  return ans;
}
#undef FUNC_NAME

/* A fluid referring to an association list mapping extra hash
   characters to procedures.  */
static SCM *scm_i_read_hash_procedures;

static SCM
scm_i_read_hash_procedures_ref (void)
{
  return scm_fluid_ref (*scm_i_read_hash_procedures);
}

static void
scm_i_read_hash_procedures_set_x (SCM value)
{
  scm_fluid_set_x (*scm_i_read_hash_procedures, value);
}

\f
/* Token readers.  */


/* Size of the C buffer used to read symbols and numbers.  */
#define READER_BUFFER_SIZE            128

/* Number of 32-bit codepoints in the buffer used to read strings.  */
#define READER_STRING_BUFFER_SIZE     128

/* The maximum size of Scheme character names.  */
#define READER_CHAR_NAME_MAX_SIZE      50

/* The maximum size of reader directive names.  */
#define READER_DIRECTIVE_NAME_MAX_SIZE 50


/* `isblank' is only in C99.  */
#define CHAR_IS_BLANK_(_chr)                                    \
  (((_chr) == ' ') || ((_chr) == '\t') || ((_chr) == '\n')      \
   || ((_chr) == '\f') || ((_chr) == '\r'))

#ifdef MSDOS
# define CHAR_IS_BLANK(_chr)                    \
  ((CHAR_IS_BLANK_ (chr)) || ((_chr) == 26))
#else
# define CHAR_IS_BLANK CHAR_IS_BLANK_
#endif


/* R5RS one-character delimiters (see section 7.1.1, ``Lexical
   structure'').  */
#define CHAR_IS_R5RS_DELIMITER(c)                               \
  (CHAR_IS_BLANK (c)                                            \
   || (c) == ')' || (c) == '(' || (c) == ';' || (c) == '"')

#define CHAR_IS_DELIMITER(c)                                    \
  (CHAR_IS_R5RS_DELIMITER (c)                                   \
   || (((c) == ']' || (c) == '[') && opts->square_brackets_p))

/* Exponent markers, as defined in section 7.1.1 of R5RS, ``Lexical
   Structure''.  */
#define CHAR_IS_EXPONENT_MARKER(_chr)                           \
  (((_chr) == 'e') || ((_chr) == 's') || ((_chr) == 'f')        \
   || ((_chr) == 'd') || ((_chr) == 'l'))

/* Read an SCSH block comment.  */
static SCM scm_read_scsh_block_comment (scm_t_wchar, SCM);
static SCM scm_read_r6rs_block_comment (scm_t_wchar, SCM);
static SCM scm_read_commented_expression (scm_t_wchar, SCM, scm_t_read_opts *);
static SCM scm_read_shebang (scm_t_wchar, SCM, scm_t_read_opts *);
static SCM scm_get_hash_procedure (int);

/* Read from PORT until a delimiter (e.g., a whitespace) is read.  Put the
   result in the pre-allocated buffer BUF.  Return zero if the whole token has
   fewer than BUF_SIZE bytes, non-zero otherwise. READ will be set the number of
   bytes actually read.  */
static int
read_token (SCM port, scm_t_read_opts *opts,
            char *buf, size_t buf_size, size_t *read)
{
   *read = 0;

   while (*read < buf_size)
     {
       int chr;

       chr = scm_get_byte_or_eof (port);

       if (chr == EOF)
        return 0;
      else if (CHAR_IS_DELIMITER (chr))
        {
          scm_unget_byte (chr, port);
          return 0;
        }
      else
        {
          *buf = (char) chr;
          buf++, (*read)++;
        }
     }

   return 1;
 }

/* Like `read_token', but return either BUFFER, or a GC-allocated buffer
   if the token doesn't fit in BUFFER_SIZE bytes.  */
static char *
read_complete_token (SCM port, scm_t_read_opts *opts,
                     char *buffer, size_t buffer_size, size_t *read)
{
  int overflow = 0;
  size_t bytes_read, overflow_size = 0;
  char *overflow_buffer = NULL;

  do
    {
      overflow = read_token (port, opts, buffer, buffer_size, &bytes_read);
      if (bytes_read == 0)
        break;
      if (overflow || overflow_size != 0)
        {
          if (overflow_size == 0)
            {
              overflow_buffer = scm_gc_malloc_pointerless (bytes_read, "read");
              memcpy (overflow_buffer, buffer, bytes_read);
              overflow_size = bytes_read;
            }
          else
            {
              char *new_buf =
                scm_gc_malloc_pointerless (overflow_size + bytes_read, "read");

              memcpy (new_buf, overflow_buffer, overflow_size);
              memcpy (new_buf + overflow_size, buffer, bytes_read);

              overflow_buffer = new_buf;
              overflow_size += bytes_read;
            }
        }
    }
  while (overflow);

  if (overflow_size)
    *read = overflow_size;
  else
    *read = bytes_read;

  return (overflow_size > 0 ? overflow_buffer : buffer);
}

/* Skip whitespace from PORT and return the first non-whitespace character
   read.  Raise an error on end-of-file.  */
static int
flush_ws (SCM port, scm_t_read_opts *opts, const char *eoferr)
{
  scm_t_wchar c;
  while (1)
    switch (c = scm_getc (port))
      {
      case EOF:
      goteof:
        if (eoferr)
          {
            scm_i_input_error (eoferr,
                               port,
                               "end of file",
                               SCM_EOL);
          }
        return c;

      case ';':
      lp:
        switch (c = scm_getc (port))
          {
          case EOF:
            goto goteof;
          default:
            goto lp;
          case SCM_LINE_INCREMENTORS:
            break;
          }
        break;

      case '#':
        switch (c = scm_getc (port))
          {
          case EOF:
            eoferr = "read_sharp";
            goto goteof;
          case '!':
            scm_read_shebang (c, port, opts);
            break;
          case ';':
            scm_read_commented_expression (c, port, opts);
            break;
          case '|':
            if (scm_is_false (scm_get_hash_procedure (c)))
              {
                scm_read_r6rs_block_comment (c, port);
                break;
              }
            /* fall through */
          default:
            scm_ungetc (c, port);
            return '#';
          }
        break;

      case SCM_LINE_INCREMENTORS:
      case SCM_SINGLE_SPACES:
      case '\t':
        break;

      default:
        return c;
      }

  return 0;
}


\f
/* Token readers.  */

static SCM scm_read_expression (SCM port, scm_t_read_opts *opts);
static SCM scm_read_sharp (int chr, SCM port, scm_t_read_opts *opts,
                           long line, int column);


static SCM
maybe_annotate_source (SCM x, SCM port, scm_t_read_opts *opts,
                       long line, int column)
{
  if (opts->record_positions_p)
    scm_i_set_source_properties_x (x, line, column, SCM_FILENAME (port));
  return x;
}

static SCM
scm_read_sexp (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
#define FUNC_NAME "scm_i_lreadparen"
{
  int c;
  SCM tmp, tl, ans = SCM_EOL;
  const int terminating_char = ((chr == '[') ? ']' : ')');

  /* Need to capture line and column numbers here. */
  long line = SCM_LINUM (port);
  int column = SCM_COL (port) - 1;

  c = flush_ws (port, opts, FUNC_NAME);
  if (terminating_char == c)
    return SCM_EOL;

  scm_ungetc (c, port);
  tmp = scm_read_expression (port, opts);

  /* Note that it is possible for scm_read_expression to return
     scm_sym_dot, but not as part of a dotted pair: as in #{.}#.  So
     check that it's a real dot by checking `c'.  */
  if (c == '.' && scm_is_eq (scm_sym_dot, tmp))
    {
      ans = scm_read_expression (port, opts);
      if (terminating_char != (c = flush_ws (port, opts, FUNC_NAME)))
        scm_i_input_error (FUNC_NAME, port, "missing close paren",
                           SCM_EOL);
      return ans;
    }

  /* Build the head of the list structure. */
  ans = tl = scm_cons (tmp, SCM_EOL);

  while (terminating_char != (c = flush_ws (port, opts, FUNC_NAME)))
    {
      SCM new_tail;

      if (c == ')' || (c == ']' && opts->square_brackets_p))
        scm_i_input_error (FUNC_NAME, port,
                           "in pair: mismatched close paren: ~A",
                           scm_list_1 (SCM_MAKE_CHAR (c)));

      scm_ungetc (c, port);
      tmp = scm_read_expression (port, opts);

      /* See above note about scm_sym_dot.  */
      if (c == '.' && scm_is_eq (scm_sym_dot, tmp))
        {
          SCM_SETCDR (tl, scm_read_expression (port, opts));

          c = flush_ws (port, opts, FUNC_NAME);
          if (terminating_char != c)
            scm_i_input_error (FUNC_NAME, port,
                               "in pair: missing close paren", SCM_EOL);
          goto exit;
        }

      new_tail = scm_cons (tmp, SCM_EOL);
      SCM_SETCDR (tl, new_tail);
      tl = new_tail;
    }

 exit:
  return maybe_annotate_source (ans, port, opts, line, column);
}
#undef FUNC_NAME


/* Read a hexadecimal number NDIGITS in length.  Put its value into the variable
   C.  If TERMINATOR is non-null, terminate early if the TERMINATOR character is
   found.  */
#define SCM_READ_HEX_ESCAPE(ndigits, terminator)                   \
  do                                                               \
    {                                                              \
      scm_t_wchar a;                                               \
      size_t i = 0;                                                \
      c = 0;                                                       \
      while (i < ndigits)                                          \
        {                                                          \
          a = scm_getc (port);                                     \
          if (a == EOF)                                            \
            goto str_eof;                                          \
          if (terminator                                           \
              && (a == (scm_t_wchar) terminator)                   \
              && (i > 0))                                          \
            break;                                                 \
          if ('0' <= a && a <= '9')                                \
            a -= '0';                                              \
          else if ('A' <= a && a <= 'F')                           \
            a = a - 'A' + 10;                                      \
          else if ('a' <= a && a <= 'f')                           \
            a = a - 'a' + 10;                                      \
          else                                                     \
            {                                                      \
              c = a;                                               \
              goto bad_escaped;                                    \
            }                                                      \
          c = c * 16 + a;                                          \
          i ++;                                                    \
        }                                                          \
    } while (0)

static void
skip_intraline_whitespace (SCM port)
{
  scm_t_wchar c;
  
  do
    {
      c = scm_getc (port);
      if (c == EOF)
        return;
    }
  while (c == '\t' || uc_is_general_category (c, UC_SPACE_SEPARATOR));

  scm_ungetc (c, port);
}                                         

static SCM
scm_read_string (int chr, SCM port, scm_t_read_opts *opts)
#define FUNC_NAME "scm_lreadr"
{
  /* For strings smaller than C_STR, this function creates only one Scheme
     object (the string returned).  */

  SCM str = SCM_EOL;
  size_t c_str_len = 0;
  scm_t_wchar c, c_str[READER_STRING_BUFFER_SIZE];

  /* Need to capture line and column numbers here. */
  long line = SCM_LINUM (port);
  int column = SCM_COL (port) - 1;

  while ('"' != (c = scm_getc (port)))
    {
      if (c == EOF)
        {
        str_eof:
          scm_i_input_error (FUNC_NAME, port,
                             "end of file in string constant", SCM_EOL);
        }

      if (c_str_len + 1 >= READER_STRING_BUFFER_SIZE)
        {
          str = scm_cons (scm_from_utf32_stringn (c_str, c_str_len), str);
          c_str_len = 0;
        }

      if (c == '\\')
        {
          switch (c = scm_getc (port))
            {
            case EOF:
              goto str_eof;
            case '"':
            case '\\':
              break;
            case '\n':
              if (opts->hungry_eol_escapes_p)
                skip_intraline_whitespace (port);
              continue;
            case '0':
              c = '\0';
              break;
            case 'f':
              c = '\f';
              break;
            case 'n':
              c = '\n';
              break;
            case 'r':
              c = '\r';
              break;
            case 't':
              c = '\t';
              break;
            case 'a':
              c = '\007';
              break;
            case 'v':
              c = '\v';
              break;
            case 'b':
              c = '\010';
              break;
            case 'x':
              if (opts->r6rs_escapes_p)
                SCM_READ_HEX_ESCAPE (10, ';');
              else
                SCM_READ_HEX_ESCAPE (2, '\0');
              break;
            case 'u':
              if (!opts->r6rs_escapes_p)
                {
                  SCM_READ_HEX_ESCAPE (4, '\0');
                  break;
                }
            case 'U':
              if (!opts->r6rs_escapes_p)
                {
                  SCM_READ_HEX_ESCAPE (6, '\0');
                  break;
                }
            default:
            bad_escaped:
              scm_i_input_error (FUNC_NAME, port,
                                 "illegal character in escape sequence: ~S",
                                 scm_list_1 (SCM_MAKE_CHAR (c)));
            }
        }

      c_str[c_str_len++] = c;
    }

  if (scm_is_null (str))
    /* Fast path: we got a string that fits in C_STR.  */
    str = scm_from_utf32_stringn (c_str, c_str_len);
  else
    {
      if (c_str_len > 0)
        str = scm_cons (scm_from_utf32_stringn (c_str, c_str_len), str);

      str = scm_string_concatenate_reverse (str, SCM_UNDEFINED, SCM_UNDEFINED);
    }

  return maybe_annotate_source (str, port, opts, line, column);
}
#undef FUNC_NAME


static SCM
scm_read_number (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
{
  SCM result, str = SCM_EOL;
  char local_buffer[READER_BUFFER_SIZE], *buffer;
  size_t bytes_read;
  scm_t_port *pt = SCM_PTAB_ENTRY (port);

  /* Need to capture line and column numbers here. */
  long line = SCM_LINUM (port);
  int column = SCM_COL (port) - 1;

  scm_ungetc (chr, port);
  buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
                                &bytes_read);

  str = scm_from_stringn (buffer, bytes_read, pt->encoding, pt->ilseq_handler);

  result = scm_string_to_number (str, SCM_UNDEFINED);
  if (scm_is_false (result))
    {
      /* Return a symbol instead of a number */
      if (opts->case_insensitive_p)
        str = scm_string_downcase_x (str);
      result = scm_string_to_symbol (str);
    }
  else if (SCM_NIMP (result))
    result = maybe_annotate_source (result, port, opts, line, column);

  SCM_COL (port) += scm_i_string_length (str);
  return result;
}

static SCM
scm_read_mixed_case_symbol (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
{
  SCM result;
  int ends_with_colon = 0;
  size_t bytes_read;
  int postfix = (opts->keyword_style == KEYWORD_STYLE_POSTFIX);
  char local_buffer[READER_BUFFER_SIZE], *buffer;
  scm_t_port *pt = SCM_PTAB_ENTRY (port);
  SCM str;

  scm_ungetc (chr, port);
  buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
                                &bytes_read);
  if (bytes_read > 0)
    ends_with_colon = buffer[bytes_read - 1] == ':';

  if (postfix && ends_with_colon && (bytes_read > 1))
    {
      str = scm_from_stringn (buffer, bytes_read - 1,
                              pt->encoding, pt->ilseq_handler);

      if (opts->case_insensitive_p)
        str = scm_string_downcase_x (str);
      result = scm_symbol_to_keyword (scm_string_to_symbol (str));
    }
  else
    {
      str = scm_from_stringn (buffer, bytes_read,
                              pt->encoding, pt->ilseq_handler);

      if (opts->case_insensitive_p)
        str = scm_string_downcase_x (str);
      result = scm_string_to_symbol (str);
    }

  SCM_COL (port) += scm_i_string_length (str);
  return result;
}

static SCM
scm_read_number_and_radix (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
#define FUNC_NAME "scm_lreadr"
{
  SCM result;
  size_t read;
  char local_buffer[READER_BUFFER_SIZE], *buffer;
  unsigned int radix;
  SCM str;
  scm_t_port *pt;

  switch (chr)
    {
    case 'B':
    case 'b':
      radix = 2;
      break;

    case 'o':
    case 'O':
      radix = 8;
      break;

    case 'd':
    case 'D':
      radix = 10;
      break;

    case 'x':
    case 'X':
      radix = 16;
      break;

    default:
      scm_ungetc (chr, port);
      scm_ungetc ('#', port);
      radix = 10;
    }

  buffer = read_complete_token (port, opts, local_buffer, sizeof local_buffer,
                                &read);

  pt = SCM_PTAB_ENTRY (port);
  str = scm_from_stringn (buffer, read, pt->encoding, pt->ilseq_handler);

  result = scm_string_to_number (str, scm_from_uint (radix));

  SCM_COL (port) += scm_i_string_length (str);

  if (scm_is_true (result))
    return result;

  scm_i_input_error (FUNC_NAME, port, "unknown # object", SCM_EOL);

  return SCM_BOOL_F;
}
#undef FUNC_NAME

static SCM
scm_read_quote (int chr, SCM port, scm_t_read_opts *opts)
{
  SCM p;
  long line = SCM_LINUM (port);
  int column = SCM_COL (port) - 1;

  switch (chr)
    {
    case '`':
      p = scm_sym_quasiquote;
      break;

    case '\'':
      p = scm_sym_quote;
      break;

    case ',':
      {
        scm_t_wchar c;

        c = scm_getc (port);
        if ('@' == c)
          p = scm_sym_uq_splicing;
        else
          {
            scm_ungetc (c, port);
            p = scm_sym_unquote;
          }
        break;
      }

    default:
      fprintf (stderr, "%s: unhandled quote character (%i)\n",
               "scm_read_quote", chr);
      abort ();
    }

  p = scm_cons2 (p, scm_read_expression (port, opts), SCM_EOL);
  return maybe_annotate_source (p, port, opts, line, column);
}

SCM_SYMBOL (sym_syntax, "syntax");
SCM_SYMBOL (sym_quasisyntax, "quasisyntax");
SCM_SYMBOL (sym_unsyntax, "unsyntax");
SCM_SYMBOL (sym_unsyntax_splicing, "unsyntax-splicing");

static SCM
scm_read_syntax (int chr, SCM port, scm_t_read_opts *opts)
{
  SCM p;
  long line = SCM_LINUM (port);
  int column = SCM_COL (port) - 1;

  switch (chr)
    {
    case '`':
      p = sym_quasisyntax;
      break;

    case '\'':
      p = sym_syntax;
      break;

    case ',':
      {
        int c;

        c = scm_getc (port);
        if ('@' == c)
          p = sym_unsyntax_splicing;
        else
          {
            scm_ungetc (c, port);
            p = sym_unsyntax;
          }
        break;
      }

    default:
      fprintf (stderr, "%s: unhandled syntax character (%i)\n",
               "scm_read_syntax", chr);
      abort ();
    }

  p = scm_cons2 (p, scm_read_expression (port, opts), SCM_EOL);
  return maybe_annotate_source (p, port, opts, line, column);
}

static SCM
scm_read_nil (int chr, SCM port, scm_t_read_opts *opts)
{
  SCM id = scm_read_mixed_case_symbol (chr, port, opts);

  if (!scm_is_eq (id, sym_nil))
    scm_i_input_error ("scm_read_nil", port,
                       "unexpected input while reading #nil: ~a",
                       scm_list_1 (id));

  return SCM_ELISP_NIL;
}
  
static SCM
scm_read_semicolon_comment (int chr, SCM port)
{
  int c;

  /* We use the get_byte here because there is no need to get the
     locale correct with comment input. This presumes that newline
     always represents itself no matter what the encoding is.  */
  for (c = scm_get_byte_or_eof (port);
       (c != EOF) && (c != '\n');
       c = scm_get_byte_or_eof (port));

  return SCM_UNSPECIFIED;
}

\f
/* Sharp readers, i.e. readers called after a `#' sign has been read.  */

static SCM
scm_read_boolean (int chr, SCM port)
{
  switch (chr)
    {
    case 't':
    case 'T':
      return SCM_BOOL_T;

    case 'f':
    case 'F':
      return SCM_BOOL_F;
    }

  return SCM_UNSPECIFIED;
}

static SCM
scm_read_character (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
#define FUNC_NAME "scm_lreadr"
{
  char buffer[READER_CHAR_NAME_MAX_SIZE];
  SCM charname;
  size_t charname_len, bytes_read;
  scm_t_wchar cp;
  int overflow;
  scm_t_port *pt;

  overflow = read_token (port, opts, buffer, READER_CHAR_NAME_MAX_SIZE,
                         &bytes_read);
  if (overflow)
    scm_i_input_error (FUNC_NAME, port, "character name too long", SCM_EOL);

  if (bytes_read == 0)
    {
      chr = scm_getc (port);
      if (chr == EOF)
        scm_i_input_error (FUNC_NAME, port, "unexpected end of file "
                           "while reading character", SCM_EOL);

      /* CHR must be a token delimiter, like a whitespace.  */
      return (SCM_MAKE_CHAR (chr));
    }

  pt = SCM_PTAB_ENTRY (port);

  /* Simple ASCII characters can be processed immediately.  Also, simple
     ISO-8859-1 characters can be processed immediately if the encoding for this
     port is ISO-8859-1.  */
  if (bytes_read == 1 && ((unsigned char) buffer[0] <= 127 || pt->encoding == NULL))
    {
      SCM_COL (port) += 1;
      return SCM_MAKE_CHAR (buffer[0]);
    }

  /* Otherwise, convert the buffer into a proper scheme string for
     processing.  */
  charname = scm_from_stringn (buffer, bytes_read, pt->encoding,
                               pt->ilseq_handler);
  charname_len = scm_i_string_length (charname);
  SCM_COL (port) += charname_len;
  cp = scm_i_string_ref (charname, 0);
  if (charname_len == 1)
    return SCM_MAKE_CHAR (cp);

  /* Ignore dotted circles, which may be used to keep combining characters from
     combining with the backslash in #\charname.  */
  if (cp == SCM_CODEPOINT_DOTTED_CIRCLE && charname_len == 2)
    return SCM_MAKE_CHAR (scm_i_string_ref (charname, 1));

  if (cp >= '0' && cp < '8')
    {
      /* Dirk:FIXME::  This type of character syntax is not R5RS
       * compliant.  Further, it should be verified that the constant
       * does only consist of octal digits.  */
      SCM p = scm_string_to_number (charname, scm_from_uint (8));
      if (SCM_I_INUMP (p))
        {
          scm_t_wchar c = scm_to_uint32 (p);
          if (SCM_IS_UNICODE_CHAR (c))
            return SCM_MAKE_CHAR (c);
          else
            scm_i_input_error (FUNC_NAME, port,
                               "out-of-range octal character escape: ~a",
                               scm_list_1 (charname));
        }
    }

  if (cp == 'x' && (charname_len > 1))
    {
      SCM p;

      /* Convert from hex, skipping the initial 'x' character in CHARNAME */
      p = scm_string_to_number (scm_c_substring (charname, 1, charname_len),
                                scm_from_uint (16));
      if (SCM_I_INUMP (p))
        {
          scm_t_wchar c = scm_to_uint32 (p);
          if (SCM_IS_UNICODE_CHAR (c))
            return SCM_MAKE_CHAR (c);
          else
            scm_i_input_error (FUNC_NAME, port,
                               "out-of-range hex character escape: ~a",
                               scm_list_1 (charname));
        }
    }

  /* The names of characters should never have non-Latin1
     characters.  */
  if (scm_i_is_narrow_string (charname)
      || scm_i_try_narrow_string (charname))
    { SCM ch = scm_i_charname_to_char (scm_i_string_chars (charname),
                                       charname_len);
      if (scm_is_true (ch))
        return ch;
    }

  scm_i_input_error (FUNC_NAME, port, "unknown character name ~a",
                     scm_list_1 (charname));

  return SCM_UNSPECIFIED;
}
#undef FUNC_NAME

static SCM
scm_read_keyword (int chr, SCM port, scm_t_read_opts *opts)
{
  SCM symbol;

  /* Read the symbol that comprises the keyword.  Doing this instead of
     invoking a specific symbol reader function allows `scm_read_keyword ()'
     to adapt to the delimiters currently valid of symbols.

     XXX: This implementation allows sloppy syntaxes like `#:  key'.  */
  symbol = scm_read_expression (port, opts);
  if (!scm_is_symbol (symbol))
    scm_i_input_error ("scm_read_keyword", port,
                       "keyword prefix `~a' not followed by a symbol: ~s",
                       scm_list_2 (SCM_MAKE_CHAR (chr), symbol));

  return (scm_symbol_to_keyword (symbol));
}

static SCM
scm_read_vector (int chr, SCM port, scm_t_read_opts *opts,
                 long line, int column)
{
  /* Note: We call `scm_read_sexp ()' rather than READER here in order to
     guarantee that it's going to do what we want.  After all, this is an
     implementation detail of `scm_read_vector ()', not a desirable
     property.  */
  return maybe_annotate_source (scm_vector (scm_read_sexp (chr, port, opts)),
                                port, opts, line, column);
}

/* Helper used by scm_read_array */
static int
read_decimal_integer (SCM port, int c, ssize_t *resp)
{
  ssize_t sign = 1;
  ssize_t res = 0;
  int got_it = 0;

  if (c == '-')
    {
      sign = -1;
      c = scm_getc (port);
    }

  while ('0' <= c && c <= '9')
    {
      res = 10*res + c-'0';
      got_it = 1;
      c = scm_getc (port);
    }

  if (got_it)
    *resp = sign * res;
  return c;
}

/* Read an array.  This function can also read vectors and uniform
   vectors.  Also, the conflict between '#f' and '#f32' and '#f64' is
   handled here.

   C is the first character read after the '#'.
*/
static SCM
scm_read_array (int c, SCM port, scm_t_read_opts *opts, long line, int column)
{
  ssize_t rank;
  scm_t_wchar tag_buf[8];
  int tag_len;

  SCM tag, shape = SCM_BOOL_F, elements, array;

  /* XXX - shortcut for ordinary vectors.  Shouldn't be necessary but
     the array code can not deal with zero-length dimensions yet, and
     we want to allow zero-length vectors, of course.
  */
  if (c == '(')
    return scm_read_vector (c, port, opts, line, column);

  /* Disambiguate between '#f' and uniform floating point vectors.
   */
  if (c == 'f')
    {
      c = scm_getc (port);
      if (c != '3' && c != '6')
        {
          if (c != EOF)
            scm_ungetc (c, port);
          return SCM_BOOL_F;
        }
      rank = 1;
      tag_buf[0] = 'f';
      tag_len = 1;
      goto continue_reading_tag;
    }

  /* Read rank. */
  rank = 1;
  c = read_decimal_integer (port, c, &rank);
  if (rank < 0)
    scm_i_input_error (NULL, port, "array rank must be non-negative",
                       SCM_EOL);

  /* Read tag. */
  tag_len = 0;
 continue_reading_tag:
  while (c != EOF && c != '(' && c != '@' && c != ':'
         && tag_len < sizeof tag_buf / sizeof tag_buf[0])
    {
      tag_buf[tag_len++] = c;
      c = scm_getc (port);
    }
  if (tag_len == 0)
    tag = SCM_BOOL_T;
  else
    {
      tag = scm_string_to_symbol (scm_from_utf32_stringn (tag_buf, tag_len));
      if (tag_len == sizeof tag_buf / sizeof tag_buf[0])
        scm_i_input_error (NULL, port, "invalid array tag, starting with: ~a",
                           scm_list_1 (tag));
    }

  /* Read shape. */
  if (c == '@' || c == ':')
    {
      shape = SCM_EOL;

      do
        {
          ssize_t lbnd = 0, len = 0;
          SCM s;

          if (c == '@')
            {
              c = scm_getc (port);
              c = read_decimal_integer (port, c, &lbnd);
            }

          s = scm_from_ssize_t (lbnd);

          if (c == ':')
            {
              c = scm_getc (port);
              c = read_decimal_integer (port, c, &len);
              if (len < 0)
                scm_i_input_error (NULL, port,
                                   "array length must be non-negative",
                                   SCM_EOL);

              s = scm_list_2 (s, scm_from_ssize_t (lbnd+len-1));
            }

          shape = scm_cons (s, shape);
        } while (c == '@' || c == ':');

      shape = scm_reverse_x (shape, SCM_EOL);
    }

  /* Read nested lists of elements. */
  if (c != '(')
    scm_i_input_error (NULL, port,
                       "missing '(' in vector or array literal",
                       SCM_EOL);
  elements = scm_read_sexp (c, port, opts);

  if (scm_is_false (shape))
    shape = scm_from_ssize_t (rank);
  else if (scm_ilength (shape) != rank)
    scm_i_input_error
      (NULL, port,
       "the number of shape specifications must match the array rank",
       SCM_EOL);

  /* Handle special print syntax of rank zero arrays; see
     scm_i_print_array for a rationale. */
  if (rank == 0)
    {
      if (!scm_is_pair (elements))
        scm_i_input_error (NULL, port,
                           "too few elements in array literal, need 1",
                           SCM_EOL);
      if (!scm_is_null (SCM_CDR (elements)))
        scm_i_input_error (NULL, port,
                           "too many elements in array literal, want 1",
                           SCM_EOL);
      elements = SCM_CAR (elements);
    }

  /* Construct array, annotate with source location, and return. */
  array = scm_list_to_typed_array (tag, shape, elements);
  return maybe_annotate_source (array, port, opts, line, column);
}

static SCM
scm_read_srfi4_vector (int chr, SCM port, scm_t_read_opts *opts,
                       long line, int column)
{
  return scm_read_array (chr, port, opts, line, column);
}

static SCM
scm_read_bytevector (scm_t_wchar chr, SCM port, scm_t_read_opts *opts,
                     long line, int column)
{
  chr = scm_getc (port);
  if (chr != 'u')
    goto syntax;

  chr = scm_getc (port);
  if (chr != '8')
    goto syntax;

  chr = scm_getc (port);
  if (chr != '(')
    goto syntax;

  return maybe_annotate_source
    (scm_u8_list_to_bytevector (scm_read_sexp (chr, port, opts)),
     port, opts, line, column);

 syntax:
  scm_i_input_error ("read_bytevector", port,
                     "invalid bytevector prefix",
                     SCM_MAKE_CHAR (chr));
  return SCM_UNSPECIFIED;
}

static SCM
scm_read_guile_bit_vector (scm_t_wchar chr, SCM port, scm_t_read_opts *opts,
                           long line, int column)
{
  /* Read the `#*10101'-style read syntax for bit vectors in Guile.  This is
     terribly inefficient but who cares?  */
  SCM s_bits = SCM_EOL;

  for (chr = scm_getc (port);
       (chr != EOF) && ((chr == '0') || (chr == '1'));
       chr = scm_getc (port))
    {
      s_bits = scm_cons ((chr == '0') ? SCM_BOOL_F : SCM_BOOL_T, s_bits);
    }

  if (chr != EOF)
    scm_ungetc (chr, port);

  return maybe_annotate_source
    (scm_bitvector (scm_reverse_x (s_bits, SCM_EOL)),
     port, opts, line, column);
}

static SCM
scm_read_scsh_block_comment (scm_t_wchar chr, SCM port)
{
  int bang_seen = 0;

  for (;;)
    {
      int c = scm_getc (port);

      if (c == EOF)
        scm_i_input_error ("skip_block_comment", port,
                           "unterminated `#! ... !#' comment", SCM_EOL);

      if (c == '!')
        bang_seen = 1;
      else if (c == '#' && bang_seen)
        break;
      else
        bang_seen = 0;
    }

  return SCM_UNSPECIFIED;
}

static SCM
scm_read_shebang (scm_t_wchar chr, SCM port, scm_t_read_opts *opts)
{
  char name[READER_DIRECTIVE_NAME_MAX_SIZE + 1];
  int c;
  int i = 0;

  while (i <= READER_DIRECTIVE_NAME_MAX_SIZE)
    {
      c = scm_getc (port);
      if (c == EOF)
        scm_i_input_error ("skip_block_comment", port,
                           "unterminated `#! ... !#' comment", SCM_EOL);
      else if (('a' <= c && c <= 'z') || ('0' <= c && c <= '9') || c == '-')
        name[i++] = c;
      else if (CHAR_IS_DELIMITER (c))
        {
          scm_ungetc (c, port);
          name[i] = '\0';
          if (0 == strcmp ("r6rs", name))
            ;  /* Silently ignore */
          else
            break;

          return SCM_UNSPECIFIED;
        }
    }
  while (i > 0)
    scm_ungetc (name[--i], port);
  return scm_read_scsh_block_comment (chr, port);
}

static SCM
scm_read_r6rs_block_comment (scm_t_wchar chr, SCM port)
{
  /* Unlike SCSH-style block comments, SRFI-30/R6RS block comments may be
     nested.  So care must be taken.  */
  int nesting_level = 1;

  int a = scm_getc (port);

  if (a == EOF)
    scm_i_input_error ("scm_read_r6rs_block_comment", port,
                       "unterminated `#| ... |#' comment", SCM_EOL);

  while (nesting_level > 0)
    {
      int b = scm_getc (port);

      if (b == EOF)
        scm_i_input_error ("scm_read_r6rs_block_comment", port,
                           "unterminated `#| ... |#' comment", SCM_EOL);

      if (a == '|' && b == '#')
        {
          nesting_level--;
          b = EOF;
        }
      else if (a == '#' && b == '|')
        {
          nesting_level++;
          b = EOF;
        }

      a = b;
    }

  return SCM_UNSPECIFIED;
}

static SCM
scm_read_commented_expression (scm_t_wchar chr, SCM port,
                               scm_t_read_opts *opts)
{
  scm_t_wchar c;
  
  c = flush_ws (port, opts, (char *) NULL);
  if (EOF == c)
    scm_i_input_error ("read_commented_expression", port,
                       "no expression after #; comment", SCM_EOL);
  scm_ungetc (c, port);
  scm_read_expression (port, opts);
  return SCM_UNSPECIFIED;
}

static SCM
scm_read_extended_symbol (scm_t_wchar chr, SCM port)
{
  /* Guile's extended symbol read syntax looks like this:

       #{This is all a symbol name}#

     So here, CHR is expected to be `{'.  */
  int saw_brace = 0;
  size_t len = 0;
  SCM buf = scm_i_make_string (1024, NULL, 0);

  buf = scm_i_string_start_writing (buf);

  while ((chr = scm_getc (port)) != EOF)
    {
      if (saw_brace)
        {
          if (chr == '#')
            {
              break;
            }
          else
            {
              saw_brace = 0;
              scm_i_string_set_x (buf, len++, '}');
            }
        }

      if (chr == '}')
        saw_brace = 1;
      else if (chr == '\\')
        {
          /* It used to be that print.c would print extended-read-syntax
             symbols with backslashes before "non-standard" chars, but
             this routine wouldn't do anything with those escapes.
             Bummer.  What we've done is to change print.c to output
             R6RS hex escapes for those characters, relying on the fact
             that the extended read syntax would never put a `\' before
             an `x'.  For now, we just ignore other instances of
             backslash in the string.  */
          switch ((chr = scm_getc (port)))
            {
            case EOF:
              goto done;
            case 'x':
              {
                scm_t_wchar c;
                
                SCM_READ_HEX_ESCAPE (10, ';');
                scm_i_string_set_x (buf, len++, c);
                break;

              str_eof:
                chr = EOF;
                goto done;

              bad_escaped:
                scm_i_string_stop_writing ();
                scm_i_input_error ("scm_read_extended_symbol", port,
                                   "illegal character in escape sequence: ~S",
                                   scm_list_1 (SCM_MAKE_CHAR (c)));
                break;
              }
            default:
              scm_i_string_set_x (buf, len++, chr);
              break;
            }
        }
      else
        scm_i_string_set_x (buf, len++, chr);

      if (len >= scm_i_string_length (buf) - 2)
        {
          SCM addy;

          scm_i_string_stop_writing ();
          addy = scm_i_make_string (1024, NULL, 0);
          buf = scm_string_append (scm_list_2 (buf, addy));
          len = 0;
          buf = scm_i_string_start_writing (buf);
        }
    }

 done:
  scm_i_string_stop_writing ();
  if (chr == EOF)
    scm_i_input_error ("scm_read_extended_symbol", port,
                       "end of file while reading symbol", SCM_EOL);

  return (scm_string_to_symbol (scm_c_substring (buf, 0, len)));
}


\f
/* Top-level token readers, i.e., dispatchers.  */

static SCM
scm_read_sharp_extension (int chr, SCM port, scm_t_read_opts *opts)
{
  SCM proc;

  proc = scm_get_hash_procedure (chr);
  if (scm_is_true (scm_procedure_p (proc)))
    {
      long line = SCM_LINUM (port);
      int column = SCM_COL (port) - 2;
      SCM got;

      got = scm_call_2 (proc, SCM_MAKE_CHAR (chr), port);

      if (opts->record_positions_p && SCM_NIMP (got)
          && !scm_i_has_source_properties (got))
        scm_i_set_source_properties_x (got, line, column, SCM_FILENAME (port));
      
      return got;
    }

  return SCM_UNSPECIFIED;
}

/* The reader for the sharp `#' character.  It basically dispatches reads
   among the above token readers.   */
static SCM
scm_read_sharp (scm_t_wchar chr, SCM port, scm_t_read_opts *opts,
                long line, int column)
#define FUNC_NAME "scm_lreadr"
{
  SCM result;

  chr = scm_getc (port);

  result = scm_read_sharp_extension (chr, port, opts);
  if (!scm_is_eq (result, SCM_UNSPECIFIED))
    return result;

  switch (chr)
    {
    case '\\':
      return (scm_read_character (chr, port, opts));
    case '(':
      return (scm_read_vector (chr, port, opts, line, column));
    case 's':
    case 'u':
    case 'f':
    case 'c':
      /* This one may return either a boolean or an SRFI-4 vector.  */
      return (scm_read_srfi4_vector (chr, port, opts, line, column));
    case 'v':
      return (scm_read_bytevector (chr, port, opts, line, column));
    case '*':
      return (scm_read_guile_bit_vector (chr, port, opts, line, column));
    case 't':
    case 'T':
    case 'F':
      return (scm_read_boolean (chr, port));
    case ':':
      return (scm_read_keyword (chr, port, opts));
    case '0': case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8': case '9':
    case '@':
#if SCM_ENABLE_DEPRECATED
      /* See below for 'i' and 'e'. */
    case 'a':
    case 'y':
    case 'h':
    case 'l':
#endif
      return (scm_read_array (chr, port, opts, line, column));

    case 'i':
    case 'e':
#if SCM_ENABLE_DEPRECATED
      {
        /* When next char is '(', it really is an old-style
           uniform array. */
        scm_t_wchar next_c = scm_getc (port);
        if (next_c != EOF)
          scm_ungetc (next_c, port);
        if (next_c == '(')
          return scm_read_array (chr, port, opts, line, column);
        /* Fall through. */
      }
#endif
    case 'b':
    case 'B':
    case 'o':
    case 'O':
    case 'd':
    case 'D':
    case 'x':
    case 'X':
    case 'I':
    case 'E':
      return (scm_read_number_and_radix (chr, port, opts));
    case '{':
      return (scm_read_extended_symbol (chr, port));
    case '!':
      return (scm_read_shebang (chr, port, opts));
    case ';':
      return (scm_read_commented_expression (chr, port, opts));
    case '`':
    case '\'':
    case ',':
      return (scm_read_syntax (chr, port, opts));
    case 'n':
      return (scm_read_nil (chr, port, opts));
    default:
      result = scm_read_sharp_extension (chr, port, opts);
      if (scm_is_eq (result, SCM_UNSPECIFIED))
        {
          /* To remain compatible with 1.8 and earlier, the following
             characters have lower precedence than `read-hash-extend'
             characters.  */
          switch (chr)
            {
            case '|':
              return scm_read_r6rs_block_comment (chr, port);
            default:
              scm_i_input_error (FUNC_NAME, port, "Unknown # object: ~S",
                                 scm_list_1 (SCM_MAKE_CHAR (chr)));
            }
        }
      else
        return result;
    }

  return SCM_UNSPECIFIED;
}
#undef FUNC_NAME

static SCM
scm_read_expression (SCM port, scm_t_read_opts *opts)
#define FUNC_NAME "scm_read_expression"
{
  while (1)
    {
      scm_t_wchar chr;

      chr = scm_getc (port);

      switch (chr)
        {
        case SCM_WHITE_SPACES:
        case SCM_LINE_INCREMENTORS:
          break;
        case ';':
          (void) scm_read_semicolon_comment (chr, port);
          break;
        case '[':
          if (!opts->square_brackets_p)
            return (scm_read_mixed_case_symbol (chr, port, opts));
          /* otherwise fall through */
        case '(':
          return (scm_read_sexp (chr, port, opts));
        case '"':
          return (scm_read_string (chr, port, opts));
        case '\'':
        case '`':
        case ',':
          return (scm_read_quote (chr, port, opts));
        case '#':
          {
            long line  = SCM_LINUM (port);
            int column = SCM_COL (port) - 1;
            SCM result = scm_read_sharp (chr, port, opts, line, column);
            if (scm_is_eq (result, SCM_UNSPECIFIED))
              /* We read a comment or some such.  */
              break;
            else
              return result;
          }
        case ')':
          scm_i_input_error (FUNC_NAME, port, "unexpected \")\"", SCM_EOL);
          break;
        case ']':
          if (opts->square_brackets_p)
            scm_i_input_error (FUNC_NAME, port, "unexpected \"]\"", SCM_EOL);
          /* otherwise fall through */
        case EOF:
          return SCM_EOF_VAL;
        case ':':
          if (opts->keyword_style == KEYWORD_STYLE_PREFIX)
            return scm_symbol_to_keyword (scm_read_expression (port, opts));
          /* Fall through.  */

        default:
          {
            if (((chr >= '0') && (chr <= '9'))
                || (strchr ("+-.", chr)))
              return (scm_read_number (chr, port, opts));
            else
              return (scm_read_mixed_case_symbol (chr, port, opts));
          }
        }
    }
}
#undef FUNC_NAME

\f
/* Actual reader.  */

static void init_read_options (SCM port, scm_t_read_opts *opts);

SCM_DEFINE (scm_read, "read", 0, 1, 0, 
            (SCM port),
            "Read an s-expression from the input port @var{port}, or from\n"
            "the current input port if @var{port} is not specified.\n"
            "Any whitespace before the next token is discarded.")
#define FUNC_NAME s_scm_read
{
  scm_t_read_opts opts;
  int c;

  if (SCM_UNBNDP (port))
    port = scm_current_input_port ();
  SCM_VALIDATE_OPINPORT (1, port);

  init_read_options (port, &opts);

  c = flush_ws (port, &opts, (char *) NULL);
  if (EOF == c)
    return SCM_EOF_VAL;
  scm_ungetc (c, port);

  return (scm_read_expression (port, &opts));
}
#undef FUNC_NAME


\f

/* Manipulate the read-hash-procedures alist.  This could be written in
   Scheme, but maybe it will also be used by C code during initialisation.  */
SCM_DEFINE (scm_read_hash_extend, "read-hash-extend", 2, 0, 0,
            (SCM chr, SCM proc),
            "Install the procedure @var{proc} for reading expressions\n"
            "starting with the character sequence @code{#} and @var{chr}.\n"
            "@var{proc} will be called with two arguments:  the character\n"
            "@var{chr} and the port to read further data from. The object\n"
            "returned will be the return value of @code{read}. \n"
            "Passing @code{#f} for @var{proc} will remove a previous setting. \n"
            )
#define FUNC_NAME s_scm_read_hash_extend
{
  SCM this;
  SCM prev;

  SCM_VALIDATE_CHAR (1, chr);
  SCM_ASSERT (scm_is_false (proc)
              || scm_is_eq (scm_procedure_p (proc), SCM_BOOL_T),
              proc, SCM_ARG2, FUNC_NAME);

  /* Check if chr is already in the alist.  */
  this = scm_i_read_hash_procedures_ref ();
  prev = SCM_BOOL_F;
  while (1)
    {
      if (scm_is_null (this))
        {
          /* not found, so add it to the beginning.  */
          if (scm_is_true (proc))
            {
              SCM new = scm_cons (scm_cons (chr, proc),
                                  scm_i_read_hash_procedures_ref ());
              scm_i_read_hash_procedures_set_x (new);
            }
          break;
        }
      if (scm_is_eq (chr, SCM_CAAR (this)))
        {
          /* already in the alist.  */
          if (scm_is_false (proc))
            {
              /* remove it.  */
              if (scm_is_false (prev))
                {
                  SCM rest = SCM_CDR (scm_i_read_hash_procedures_ref ());
                  scm_i_read_hash_procedures_set_x (rest);
                }
              else
                scm_set_cdr_x (prev, SCM_CDR (this));
            }
          else
            {
              /* replace it.  */
              scm_set_cdr_x (SCM_CAR (this), proc);
            }
          break;
        }
      prev = this;
      this = SCM_CDR (this);
    }

  return SCM_UNSPECIFIED;
}
#undef FUNC_NAME

/* Recover the read-hash procedure corresponding to char c.  */
static SCM
scm_get_hash_procedure (int c)
{
  SCM rest = scm_i_read_hash_procedures_ref ();

  while (1)
    {
      if (scm_is_null (rest))
        return SCM_BOOL_F;
  
      if (SCM_CHAR (SCM_CAAR (rest)) == c)
        return SCM_CDAR (rest);
     
      rest = SCM_CDR (rest);
    }
}

#define SCM_ENCODING_SEARCH_SIZE (500)

/* Search the first few hundred characters of a file for an Emacs-like coding
   declaration.  Returns either NULL or a string whose storage has been
   allocated with `scm_gc_malloc ()'.  */
char *
scm_i_scan_for_encoding (SCM port)
{
  scm_t_port *pt;
  char header[SCM_ENCODING_SEARCH_SIZE+1];
  size_t bytes_read, encoding_length, i;
  char *encoding = NULL;
  int utf8_bom = 0;
  char *pos, *encoding_start;
  int in_comment;

  pt = SCM_PTAB_ENTRY (port);

  if (pt->rw_active == SCM_PORT_WRITE)
    scm_flush (port);

  if (pt->rw_random)
    pt->rw_active = SCM_PORT_READ;

  if (pt->read_pos == pt->read_end)
    {
      /* We can use the read buffer, and thus avoid a seek. */
      if (scm_fill_input (port) == EOF)
        return NULL;

      bytes_read = pt->read_end - pt->read_pos;
      if (bytes_read > SCM_ENCODING_SEARCH_SIZE)
        bytes_read = SCM_ENCODING_SEARCH_SIZE;

      if (bytes_read <= 1)
        /* An unbuffered port -- don't scan.  */
        return NULL;

      memcpy (header, pt->read_pos, bytes_read);
      header[bytes_read] = '\0';
    }
  else
    {
      /* Try to read some bytes and then seek back.  Not all ports
         support seeking back; and indeed some file ports (like
         /dev/urandom) will succeed on an lseek (fd, 0, SEEK_CUR)---the
         check performed by SCM_FPORT_FDES---but fail to seek
         backwards.  Hence this block comes second.  We prefer to use
         the read buffer in-place.  */
      if (SCM_FPORTP (port) && !SCM_FDES_RANDOM_P (SCM_FPORT_FDES (port)))
        return NULL;

      bytes_read = scm_c_read (port, header, SCM_ENCODING_SEARCH_SIZE);
      header[bytes_read] = '\0';
      scm_seek (port, scm_from_int (0), scm_from_int (SEEK_SET));
    }

  if (bytes_read > 3 
      && header[0] == '\xef' && header[1] == '\xbb' && header[2] == '\xbf')
    utf8_bom = 1;

  /* search past "coding[:=]" */
  pos = header;
  while (1)
    {
      if ((pos = strstr(pos, "coding")) == NULL)
        return NULL;

      pos += strlen("coding");
      if (pos - header >= SCM_ENCODING_SEARCH_SIZE || 
          (*pos == ':' || *pos == '='))
        {
          pos ++;
          break;
        }
    }

  /* skip spaces */
  while (pos - header <= SCM_ENCODING_SEARCH_SIZE && 
         (*pos == ' ' || *pos == '\t'))
    pos ++;

  /* grab the next token */
  encoding_start = pos;
  i = 0;
  while (encoding_start + i - header <= SCM_ENCODING_SEARCH_SIZE
         && encoding_start + i - header < bytes_read
         && (isalnum ((int) encoding_start[i])
             || strchr ("_-.:/,+=()", encoding_start[i]) != NULL))
    i++;

  encoding_length = i;
  if (encoding_length == 0)
    return NULL;

  encoding = scm_gc_strndup (encoding_start, encoding_length, "encoding");
  for (i = 0; i < encoding_length; i++)
    encoding[i] = toupper ((int) encoding[i]);

  /* push backwards to make sure we were in a comment */
  in_comment = 0;
  pos = encoding_start;
  while (pos >= header)
    {
      if (*pos == ';')
        {
          in_comment = 1;
          break;
        }
      else if (*pos == '\n' || pos == header)
        {
          /* This wasn't in a semicolon comment. Check for a
           hash-bang comment. */
          char *beg = strstr (header, "#!");
          char *end = strstr (header, "!#");
          if (beg < encoding_start && encoding_start + encoding_length <= end)
            in_comment = 1;
          break;
        }
      else
        {
          pos --;
          continue;
        }
    }
  if (!in_comment)
    /* This wasn't in a comment */
    return NULL;

  if (utf8_bom && strcmp(encoding, "UTF-8"))
    scm_misc_error (NULL,
                    "the port input declares the encoding ~s but is encoded as UTF-8",
                    scm_list_1 (scm_from_locale_string (encoding)));

  return encoding;
}

SCM_DEFINE (scm_file_encoding, "file-encoding", 1, 0, 0,
            (SCM port),
            "Scans the port for an Emacs-like character coding declaration\n"
            "near the top of the contents of a port with random-accessible contents.\n"
            "The coding declaration is of the form\n"
            "@code{coding: XXXXX} and must appear in a scheme comment.\n"
            "\n"
            "Returns a string containing the character encoding of the file\n"
            "if a declaration was found, or @code{#f} otherwise.\n")
#define FUNC_NAME s_scm_file_encoding
{
  char *enc;
  SCM s_enc;

  SCM_VALIDATE_OPINPORT (SCM_ARG1, port);

  enc = scm_i_scan_for_encoding (port);
  if (enc == NULL)
    return SCM_BOOL_F;
  else
    {
      s_enc = scm_from_locale_string (enc);
      return s_enc;
    }

  return SCM_BOOL_F;
}
#undef FUNC_NAME

\f
/* Per-port read options.

   We store per-port read options in the 'port-read-options' key of the
   port's alist, which is stored in 'scm_i_port_weak_hash'.  The value
   stored in the alist is a single integer that contains a two-bit field
   for each read option.

   If a bit field contains READ_OPTION_INHERIT (3), that indicates that
   the applicable value should be inherited from the corresponding
   global read option.  Otherwise, the bit field contains the value of
   the read option.  For boolean read options that have been set
   per-port, the possible values are 0 or 1.  If the 'keyword_style'
   read option has been set per-port, its possible values are those in
   'enum t_keyword_style'. */

/* Key to read options in per-port alists. */
SCM_SYMBOL (sym_port_read_options, "port-read-options");

/* Offsets of bit fields for each per-port override */
#define READ_OPTION_COPY_SOURCE_P          0
#define READ_OPTION_RECORD_POSITIONS_P     2
#define READ_OPTION_CASE_INSENSITIVE_P     4
#define READ_OPTION_KEYWORD_STYLE          6
#define READ_OPTION_R6RS_ESCAPES_P         8
#define READ_OPTION_SQUARE_BRACKETS_P     10
#define READ_OPTION_HUNGRY_EOL_ESCAPES_P  12

#define READ_OPTIONS_NUM_BITS             14

#define READ_OPTIONS_INHERIT_ALL  ((1UL << READ_OPTIONS_NUM_BITS) - 1)
#define READ_OPTIONS_MAX_VALUE    READ_OPTIONS_INHERIT_ALL

#define READ_OPTION_MASK     3
#define READ_OPTION_INHERIT  3

static void
set_port_read_option (SCM port, int option, int new_value)
{
  SCM alist, scm_read_options;
  unsigned int read_options;

  new_value &= READ_OPTION_MASK;
  scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex);
  alist = scm_hashq_ref (scm_i_port_weak_hash, port, SCM_BOOL_F);
  scm_read_options = scm_assq_ref (alist, sym_port_read_options);
  if (scm_is_unsigned_integer (scm_read_options, 0, READ_OPTIONS_MAX_VALUE))
    read_options = scm_to_uint (scm_read_options);
  else
    read_options = READ_OPTIONS_INHERIT_ALL;
  read_options &= ~(READ_OPTION_MASK << option);
  read_options |= new_value << option;
  scm_read_options = scm_from_uint (read_options);
  alist = scm_assq_set_x (alist, sym_port_read_options, scm_read_options);
  scm_hashq_set_x (scm_i_port_weak_hash, port, alist);
  scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);
}

/* Initialize OPTS based on PORT's read options and the global read
   options. */
static void
init_read_options (SCM port, scm_t_read_opts *opts)
{
  SCM alist, val, scm_read_options;
  unsigned int read_options, x;

  scm_i_scm_pthread_mutex_lock (&scm_i_port_table_mutex);
  alist = scm_hashq_ref (scm_i_port_weak_hash, port, SCM_BOOL_F);
  scm_read_options = scm_assq_ref (alist, sym_port_read_options);
  scm_i_pthread_mutex_unlock (&scm_i_port_table_mutex);

  if (scm_is_unsigned_integer (scm_read_options, 0, READ_OPTIONS_MAX_VALUE))
    read_options = scm_to_uint (scm_read_options);
  else
    read_options = READ_OPTIONS_INHERIT_ALL;

  x = READ_OPTION_MASK & (read_options >> READ_OPTION_KEYWORD_STYLE);
  if (x == READ_OPTION_INHERIT)
    {
      val = SCM_PACK (SCM_KEYWORD_STYLE);
      if (scm_is_eq (val, scm_keyword_prefix))
        x = KEYWORD_STYLE_PREFIX;
      else if (scm_is_eq (val, scm_keyword_postfix))
        x = KEYWORD_STYLE_POSTFIX;
      else
        x = KEYWORD_STYLE_HASH_PREFIX;
    }
  opts->keyword_style = x;

#define RESOLVE_BOOLEAN_OPTION(NAME, name)                              \
  do                                                                    \
    {                                                                   \
      x = READ_OPTION_MASK & (read_options >> READ_OPTION_ ## NAME);    \
      if (x == READ_OPTION_INHERIT)                                     \
        x = !!SCM_ ## NAME;                                             \
          opts->name = x;                                               \
    }                                                                   \
  while (0)

  RESOLVE_BOOLEAN_OPTION (COPY_SOURCE_P,        copy_source_p);
  RESOLVE_BOOLEAN_OPTION (RECORD_POSITIONS_P,   record_positions_p);
  RESOLVE_BOOLEAN_OPTION (CASE_INSENSITIVE_P,   case_insensitive_p);
  RESOLVE_BOOLEAN_OPTION (R6RS_ESCAPES_P,       r6rs_escapes_p);
  RESOLVE_BOOLEAN_OPTION (SQUARE_BRACKETS_P,    square_brackets_p);
  RESOLVE_BOOLEAN_OPTION (HUNGRY_EOL_ESCAPES_P, hungry_eol_escapes_p);

#undef RESOLVE_BOOLEAN_OPTION
}

void
scm_init_read ()
{
  SCM read_hash_procs;

  read_hash_procs = scm_make_fluid_with_default (SCM_EOL);
  
  scm_i_read_hash_procedures =
    SCM_VARIABLE_LOC (scm_c_define ("%read-hash-procedures", read_hash_procs));

  scm_init_opts (scm_read_options, scm_read_opts);
#include "libguile/read.x"
}

/*
  Local Variables:
  c-file-style: "gnu"
  End:
*/