view libgfortran/io/read.c @ 158:494b0b89df80 default tip

...
author Shinji KONO <kono@ie.u-ryukyu.ac.jp>
date Mon, 25 May 2020 18:13:55 +0900
parents 1830386684a0
children
line wrap: on
line source

/* Copyright (C) 2002-2020 Free Software Foundation, Inc.
   Contributed by Andy Vaught
   F2003 I/O support contributed by Jerry DeLisle

This file is part of the GNU Fortran runtime library (libgfortran).

Libgfortran 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 3, or (at your option)
any later version.

Libgfortran 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.

Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.

You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
<http://www.gnu.org/licenses/>.  */

#include "io.h"
#include "fbuf.h"
#include "format.h"
#include "unix.h"
#include <string.h>
#include <ctype.h>
#include <assert.h>
#include "async.h"

typedef unsigned char uchar;

/* read.c -- Deal with formatted reads */


/* set_integer()-- All of the integer assignments come here to
   actually place the value into memory.  */

void
set_integer (void *dest, GFC_INTEGER_LARGEST value, int length)
{
  NOTE ("set_integer: %lld %p", (long long int) value, dest);
  switch (length)
    {
#ifdef HAVE_GFC_INTEGER_16
/* length=10 comes about for kind=10 real/complex BOZ, cf. PR41711. */
    case 10:
    case 16:
      {
	GFC_INTEGER_16 tmp = value;
	memcpy (dest, (void *) &tmp, length);
      }
      break;
#endif
    case 8:
      {
	GFC_INTEGER_8 tmp = value;
	memcpy (dest, (void *) &tmp, length);
      }
      break;
    case 4:
      {
	GFC_INTEGER_4 tmp = value;
	memcpy (dest, (void *) &tmp, length);
      }
      break;
    case 2:
      {
	GFC_INTEGER_2 tmp = value;
	memcpy (dest, (void *) &tmp, length);
      }
      break;
    case 1:
      {
	GFC_INTEGER_1 tmp = value;
	memcpy (dest, (void *) &tmp, length);
      }
      break;
    default:
      internal_error (NULL, "Bad integer kind");
    }
}


/* Max signed value of size give by length argument.  */

GFC_UINTEGER_LARGEST
si_max (int length)
{
#if defined HAVE_GFC_REAL_16 || defined HAVE_GFC_REAL_10
  GFC_UINTEGER_LARGEST value;
#endif

  switch (length)
      {
#if defined HAVE_GFC_REAL_16 || defined HAVE_GFC_REAL_10
    case 16:
    case 10:
      value = 1;
      for (int n = 1; n < 4 * length; n++)
        value = (value << 2) + 3;
      return value;
#endif
    case 8:
      return GFC_INTEGER_8_HUGE;
    case 4:
      return GFC_INTEGER_4_HUGE;
    case 2:
      return GFC_INTEGER_2_HUGE;
    case 1:
      return GFC_INTEGER_1_HUGE;
    default:
      internal_error (NULL, "Bad integer kind");
    }
}


/* convert_real()-- Convert a character representation of a floating
   point number to the machine number.  Returns nonzero if there is an
   invalid input.  Note: many architectures (e.g. IA-64, HP-PA)
   require that the storage pointed to by the dest argument is
   properly aligned for the type in question.  */

int
convert_real (st_parameter_dt *dtp, void *dest, const char *buffer, int length)
{
  char *endptr = NULL;
  int round_mode, old_round_mode;

  switch (dtp->u.p.current_unit->round_status)
    {
      case ROUND_COMPATIBLE:
	/* FIXME: As NEAREST but round away from zero for a tie.  */
      case ROUND_UNSPECIFIED:
	/* Should not occur.  */
      case ROUND_PROCDEFINED:
	round_mode = ROUND_NEAREST;
	break;
      default:
	round_mode = dtp->u.p.current_unit->round_status;
	break;
    }

  old_round_mode = get_fpu_rounding_mode();
  set_fpu_rounding_mode (round_mode);

  switch (length)
    {
    case 4:
      *((GFC_REAL_4*) dest) =
#if defined(HAVE_STRTOF)
	gfc_strtof (buffer, &endptr);
#else
	(GFC_REAL_4) gfc_strtod (buffer, &endptr);
#endif
      break;

    case 8:
      *((GFC_REAL_8*) dest) = gfc_strtod (buffer, &endptr);
      break;

#if defined(HAVE_GFC_REAL_10) && defined (HAVE_STRTOLD)
    case 10:
      *((GFC_REAL_10*) dest) = gfc_strtold (buffer, &endptr);
      break;
#endif

#if defined(HAVE_GFC_REAL_16)
# if defined(GFC_REAL_16_IS_FLOAT128)
    case 16:
      *((GFC_REAL_16*) dest) = __qmath_(strtoflt128) (buffer, &endptr);
      break;
# elif defined(HAVE_STRTOLD)
    case 16:
      *((GFC_REAL_16*) dest) = gfc_strtold (buffer, &endptr);
      break;
# endif
#endif

    default:
      internal_error (&dtp->common, "Unsupported real kind during IO");
    }

  set_fpu_rounding_mode (old_round_mode);

  if (buffer == endptr)
    {
      generate_error (&dtp->common, LIBERROR_READ_VALUE,
  		      "Error during floating point read");
      next_record (dtp, 1);
      return 1;
    }

  return 0;
}

/* convert_infnan()-- Convert character INF/NAN representation to the
   machine number.  Note: many architectures (e.g. IA-64, HP-PA) require
   that the storage pointed to by the dest argument is properly aligned
   for the type in question.  */

int
convert_infnan (st_parameter_dt *dtp, void *dest, const char *buffer,
	        int length)
{
  const char *s = buffer;
  int is_inf, plus = 1;

  if (*s == '+')
    s++;
  else if (*s == '-')
    {
      s++;
      plus = 0;
    }

  is_inf = *s == 'i';

  switch (length)
    {
    case 4:
      if (is_inf)
	*((GFC_REAL_4*) dest) = plus ? __builtin_inff () : -__builtin_inff ();
      else
	*((GFC_REAL_4*) dest) = plus ? __builtin_nanf ("") : -__builtin_nanf ("");
      break;

    case 8:
      if (is_inf)
	*((GFC_REAL_8*) dest) = plus ? __builtin_inf () : -__builtin_inf ();
      else
	*((GFC_REAL_8*) dest) = plus ? __builtin_nan ("") : -__builtin_nan ("");
      break;

#if defined(HAVE_GFC_REAL_10)
    case 10:
      if (is_inf)
	*((GFC_REAL_10*) dest) = plus ? __builtin_infl () : -__builtin_infl ();
      else
	*((GFC_REAL_10*) dest) = plus ? __builtin_nanl ("") : -__builtin_nanl ("");
      break;
#endif

#if defined(HAVE_GFC_REAL_16)
# if defined(GFC_REAL_16_IS_FLOAT128)
    case 16:
      *((GFC_REAL_16*) dest) = __qmath_(strtoflt128) (buffer, NULL);
      break;
# else
    case 16:
      if (is_inf)
	*((GFC_REAL_16*) dest) = plus ? __builtin_infl () : -__builtin_infl ();
      else
	*((GFC_REAL_16*) dest) = plus ? __builtin_nanl ("") : -__builtin_nanl ("");
      break;
# endif
#endif

    default:
      internal_error (&dtp->common, "Unsupported real kind during IO");
    }

  return 0;
}


/* read_l()-- Read a logical value */

void
read_l (st_parameter_dt *dtp, const fnode *f, char *dest, int length)
{
  char *p;
  size_t w;

  w = f->u.w;

  p = read_block_form (dtp, &w);

  if (p == NULL)
    return;

  while (*p == ' ')
    {
      if (--w == 0)
	goto bad;
      p++;
    }

  if (*p == '.')
    {
      if (--w == 0)
	goto bad;
      p++;
    }

  switch (*p)
    {
    case 't':
    case 'T':
      set_integer (dest, (GFC_INTEGER_LARGEST) 1, length);
      break;
    case 'f':
    case 'F':
      set_integer (dest, (GFC_INTEGER_LARGEST) 0, length);
      break;
    default:
    bad:
      generate_error (&dtp->common, LIBERROR_READ_VALUE,
		      "Bad value on logical read");
      next_record (dtp, 1);
      break;
    }
}


static gfc_char4_t
read_utf8 (st_parameter_dt *dtp, size_t *nbytes)
{
  static const uchar masks[6] = { 0x7F, 0x1F, 0x0F, 0x07, 0x02, 0x01 };
  static const uchar patns[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
  size_t nb, nread;
  gfc_char4_t c;
  char *s;

  *nbytes = 1;

  s = read_block_form (dtp, nbytes);
  if (s == NULL)
    return 0;

  /* If this is a short read, just return.  */
  if (*nbytes == 0)
    return 0;

  c = (uchar) s[0];
  if (c < 0x80)
    return c;

  /* The number of leading 1-bits in the first byte indicates how many
     bytes follow.  */
  for (nb = 2; nb < 7; nb++)
    if ((c & ~masks[nb-1]) == patns[nb-1])
      goto found;
  goto invalid;
	
 found:
  c = (c & masks[nb-1]);
  nread = nb - 1;

  s = read_block_form (dtp, &nread);
  if (s == NULL)
    return 0;
  /* Decode the bytes read.  */
  for (size_t i = 1; i < nb; i++)
    {
      gfc_char4_t n = *s++;

      if ((n & 0xC0) != 0x80)
	goto invalid;

      c = ((c << 6) + (n & 0x3F));
    }

  /* Make sure the shortest possible encoding was used.  */
  if (c <=      0x7F && nb > 1) goto invalid;
  if (c <=     0x7FF && nb > 2) goto invalid;
  if (c <=    0xFFFF && nb > 3) goto invalid;
  if (c <=  0x1FFFFF && nb > 4) goto invalid;
  if (c <= 0x3FFFFFF && nb > 5) goto invalid;

  /* Make sure the character is valid.  */
  if (c > 0x7FFFFFFF || (c >= 0xD800 && c <= 0xDFFF))
    goto invalid;

  return c;
      
 invalid:
  generate_error (&dtp->common, LIBERROR_READ_VALUE, "Invalid UTF-8 encoding");
  return (gfc_char4_t) '?';
}


static void
read_utf8_char1 (st_parameter_dt *dtp, char *p, size_t len, size_t width)
{
  gfc_char4_t c;
  char *dest;
  size_t nbytes, j;

  len = (width < len) ? len : width;

  dest = (char *) p;

  /* Proceed with decoding one character at a time.  */
  for (j = 0; j < len; j++, dest++)
    {
      c = read_utf8 (dtp, &nbytes);

      /* Check for a short read and if so, break out.  */
      if (nbytes == 0)
	break;

      *dest = c > 255 ? '?' : (uchar) c;
    }

  /* If there was a short read, pad the remaining characters.  */
  for (size_t i = j; i < len; i++)
    *dest++ = ' ';
  return;
}

static void
read_default_char1 (st_parameter_dt *dtp, char *p, size_t len, size_t width)
{
  char *s;
  size_t m;

  s = read_block_form (dtp, &width);
  
  if (s == NULL)
    return;
  if (width > len)
     s += (width - len);

  m = (width > len) ? len : width;
  memcpy (p, s, m);

  if (len > width)
    memset (p + m, ' ', len - width);
}


static void
read_utf8_char4 (st_parameter_dt *dtp, void *p, size_t len, size_t width)
{
  gfc_char4_t *dest;
  size_t nbytes, j;

  len = (width < len) ? len : width;

  dest = (gfc_char4_t *) p;

  /* Proceed with decoding one character at a time.  */
  for (j = 0; j < len; j++, dest++)
    {
      *dest = read_utf8 (dtp, &nbytes);

      /* Check for a short read and if so, break out.  */
      if (nbytes == 0)
	break;
    }

  /* If there was a short read, pad the remaining characters.  */
  for (size_t i = j; i < len; i++)
    *dest++ = (gfc_char4_t) ' ';
  return;
}


static void
read_default_char4 (st_parameter_dt *dtp, char *p, size_t len, size_t width)
{
  size_t m, n;
  gfc_char4_t *dest;

  if (is_char4_unit(dtp))
    {
      gfc_char4_t *s4;

      s4 = (gfc_char4_t *) read_block_form4 (dtp, &width);

      if (s4 == NULL)
	return;
      if (width > len)
	 s4 += (width - len);

      m = (width > len) ? len : width;

      dest = (gfc_char4_t *) p;

      for (n = 0; n < m; n++)
	*dest++ = *s4++;

      if (len > width)
	{
	  for (n = 0; n < len - width; n++)
	    *dest++ = (gfc_char4_t) ' ';
	}
    }
  else
    {
      char *s;

      s = read_block_form (dtp, &width);

      if (s == NULL)
	return;
      if (width > len)
	 s += (width - len);

      m = (width > len) ? len : width;

      dest = (gfc_char4_t *) p;

      for (n = 0; n < m; n++, dest++, s++)
	*dest = (unsigned char ) *s;

      if (len > width)
	{
	  for (n = 0; n < len - width; n++, dest++)
	    *dest = (unsigned char) ' ';
	}
    }
}


/* read_a()-- Read a character record into a KIND=1 character destination,
   processing UTF-8 encoding if necessary.  */

void
read_a (st_parameter_dt *dtp, const fnode *f, char *p, size_t length)
{
  size_t w;

  if (f->u.w == -1) /* '(A)' edit descriptor  */
    w = length;
  else
    w = f->u.w;

  /* Read in w characters, treating comma as not a separator.  */
  dtp->u.p.sf_read_comma = 0;

  if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
    read_utf8_char1 (dtp, p, length, w);
  else
    read_default_char1 (dtp, p, length, w);

  dtp->u.p.sf_read_comma =
    dtp->u.p.current_unit->decimal_status == DECIMAL_COMMA ? 0 : 1;
}


/* read_a_char4()-- Read a character record into a KIND=4 character destination,
   processing UTF-8 encoding if necessary.  */

void
read_a_char4 (st_parameter_dt *dtp, const fnode *f, char *p, size_t length)
{
  size_t w;

  if (f->u.w == -1) /* '(A)' edit descriptor  */
    w = length;
  else
    w = f->u.w;

  /* Read in w characters, treating comma as not a separator.  */
  dtp->u.p.sf_read_comma = 0;

  if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
    read_utf8_char4 (dtp, p, length, w);
  else
    read_default_char4 (dtp, p, length, w);
  
  dtp->u.p.sf_read_comma =
    dtp->u.p.current_unit->decimal_status == DECIMAL_COMMA ? 0 : 1;
}

/* eat_leading_spaces()-- Given a character pointer and a width,
   ignore the leading spaces.  */

static char *
eat_leading_spaces (size_t *width, char *p)
{
  for (;;)
    {
      if (*width == 0 || *p != ' ')
	break;

      (*width)--;
      p++;
    }

  return p;
}


static char
next_char (st_parameter_dt *dtp, char **p, size_t *w)
{
  char c, *q;

  if (*w == 0)
    return '\0';

  q = *p;
  c = *q++;
  *p = q;

  (*w)--;

  if (c != ' ')
    return c;
  if (dtp->u.p.blank_status != BLANK_UNSPECIFIED)
    return ' ';  /* return a blank to signal a null */ 

  /* At this point, the rest of the field has to be trailing blanks */

  while (*w > 0)
    {
      if (*q++ != ' ')
	return '?';
      (*w)--;
    }

  *p = q;
  return '\0';
}


/* read_decimal()-- Read a decimal integer value.  The values here are
   signed values. */

void
read_decimal (st_parameter_dt *dtp, const fnode *f, char *dest, int length)
{
  GFC_UINTEGER_LARGEST value, maxv, maxv_10;
  GFC_INTEGER_LARGEST v;
  size_t w;
  int negative;
  char c, *p;

  w = f->u.w;

  /* This is a legacy extension, and the frontend will only allow such cases
   * through when -fdec-format-defaults is passed.
   */
  if (w == (size_t) DEFAULT_WIDTH)
    w = default_width_for_integer (length);

  p = read_block_form (dtp, &w);

  if (p == NULL)
    return;

  p = eat_leading_spaces (&w, p);
  if (w == 0)
    {
      set_integer (dest, (GFC_INTEGER_LARGEST) 0, length);
      return;
    }

  negative = 0;

  switch (*p)
    {
    case '-':
      negative = 1;
      /* Fall through */

    case '+':
      p++;
      if (--w == 0)
	goto bad;
      /* Fall through */

    default:
      break;
    }

  maxv = si_max (length);
  if (negative)
    maxv++;
  maxv_10 = maxv / 10;

  /* At this point we have a digit-string */
  value = 0;

  for (;;)
    {
      c = next_char (dtp, &p, &w);
      if (c == '\0')
	break;
	
      if (c == ' ')
        {
	  if (dtp->u.p.blank_status == BLANK_NULL)
	    {
	      /* Skip spaces.  */
	      for ( ; w > 0; p++, w--)
		if (*p != ' ') break; 
	      continue;
	    }
	  if (dtp->u.p.blank_status == BLANK_ZERO) c = '0';
        }
        
      if (c < '0' || c > '9')
	goto bad;

      if (value > maxv_10)
	goto overflow;

      c -= '0';
      value = 10 * value;

      if (value > maxv - c)
	goto overflow;
      value += c;
    }

  if (negative)
    v = -value;
  else
    v = value;

  set_integer (dest, v, length);
  return;

 bad:
  generate_error (&dtp->common, LIBERROR_READ_VALUE,
		  "Bad value during integer read");
  next_record (dtp, 1);
  return;

 overflow:
  generate_error (&dtp->common, LIBERROR_READ_OVERFLOW,
		  "Value overflowed during integer read");
  next_record (dtp, 1);

}


/* read_radix()-- This function reads values for non-decimal radixes.
   The difference here is that we treat the values here as unsigned
   values for the purposes of overflow.  If minus sign is present and
   the top bit is set, the value will be incorrect. */

void
read_radix (st_parameter_dt *dtp, const fnode *f, char *dest, int length,
	    int radix)
{
  GFC_UINTEGER_LARGEST value, maxv, maxv_r;
  GFC_INTEGER_LARGEST v;
  size_t w;
  int negative;
  char c, *p;

  w = f->u.w;

  p = read_block_form (dtp, &w);

  if (p == NULL)
    return;

  p = eat_leading_spaces (&w, p);
  if (w == 0)
    {
      set_integer (dest, (GFC_INTEGER_LARGEST) 0, length);
      return;
    }

  /* Maximum unsigned value, assuming two's complement.  */
  maxv = 2 * si_max (length) + 1;
  maxv_r = maxv / radix;

  negative = 0;
  value = 0;

  switch (*p)
    {
    case '-':
      negative = 1;
      /* Fall through */

    case '+':
      p++;
      if (--w == 0)
	goto bad;
      /* Fall through */

    default:
      break;
    }

  /* At this point we have a digit-string */
  value = 0;

  for (;;)
    {
      c = next_char (dtp, &p, &w);
      if (c == '\0')
	break;
      if (c == ' ')
        {
	  if (dtp->u.p.blank_status == BLANK_NULL) continue;
	  if (dtp->u.p.blank_status == BLANK_ZERO) c = '0';
        }

      switch (radix)
	{
	case 2:
	  if (c < '0' || c > '1')
	    goto bad;
	  break;

	case 8:
	  if (c < '0' || c > '7')
	    goto bad;
	  break;

	case 16:
	  switch (c)
	    {
	    case '0':
	    case '1':
	    case '2':
	    case '3':
	    case '4':
	    case '5':
	    case '6':
	    case '7':
	    case '8':
	    case '9':
	      break;

	    case 'a':
	    case 'b':
	    case 'c':
	    case 'd':
	    case 'e':
	    case 'f':
	      c = c - 'a' + '9' + 1;
	      break;

	    case 'A':
	    case 'B':
	    case 'C':
	    case 'D':
	    case 'E':
	    case 'F':
	      c = c - 'A' + '9' + 1;
	      break;

	    default:
	      goto bad;
	    }

	  break;
	}

      if (value > maxv_r)
	goto overflow;

      c -= '0';
      value = radix * value;

      if (maxv - c < value)
	goto overflow;
      value += c;
    }

  v = value;
  if (negative)
    v = -v;

  set_integer (dest, v, length);
  return;

 bad:
  generate_error (&dtp->common, LIBERROR_READ_VALUE,
		  "Bad value during integer read");
  next_record (dtp, 1);
  return;

 overflow:
  generate_error (&dtp->common, LIBERROR_READ_OVERFLOW,
		  "Value overflowed during integer read");
  next_record (dtp, 1);

}


/* read_f()-- Read a floating point number with F-style editing, which
   is what all of the other floating point descriptors behave as.  The
   tricky part is that optional spaces are allowed after an E or D,
   and the implicit decimal point if a decimal point is not present in
   the input.  */

void
read_f (st_parameter_dt *dtp, const fnode *f, char *dest, int length)
{
#define READF_TMP 50
  char tmp[READF_TMP];
  size_t buf_size = 0;
  size_t w;
  int seen_dp, exponent;
  int exponent_sign;
  const char *p;
  char *buffer;
  char *out;
  int seen_int_digit; /* Seen a digit before the decimal point?  */
  int seen_dec_digit; /* Seen a digit after the decimal point?  */

  seen_dp = 0;
  seen_int_digit = 0;
  seen_dec_digit = 0;
  exponent_sign = 1;
  exponent = 0;
  w = f->u.w;
  buffer = tmp;

  /* Read in the next block.  */
  p = read_block_form (dtp, &w);
  if (p == NULL)
    return;
  p = eat_leading_spaces (&w, (char*) p);
  if (w == 0)
    goto zero;

  /* In this buffer we're going to re-format the number cleanly to be parsed
     by convert_real in the end; this assures we're using strtod from the
     C library for parsing and thus probably get the best accuracy possible.
     This process may add a '+0.0' in front of the number as well as change the
     exponent because of an implicit decimal point or the like.  Thus allocating
     strlen ("+0.0e-1000") == 10 characters plus one for NUL more than the
     original buffer had should be enough.  */
  buf_size = w + 11;
  if (buf_size > READF_TMP)
    buffer = xmalloc (buf_size);

  out = buffer;

  /* Optional sign */
  if (*p == '-' || *p == '+')
    {
      if (*p == '-')
	*(out++) = '-';
      ++p;
      --w;
    }

  p = eat_leading_spaces (&w, (char*) p);
  if (w == 0)
    goto zero;

  /* Check for Infinity or NaN.  */    
  if (unlikely ((w >= 3 && (*p == 'i' || *p == 'I' || *p == 'n' || *p == 'N'))))
    {
      int seen_paren = 0;
      char *save = out;

      /* Scan through the buffer keeping track of spaces and parenthesis. We
	 null terminate the string as soon as we see a left paren or if we are
	 BLANK_NULL mode.  Leading spaces have already been skipped above,
	 trailing spaces are ignored by converting to '\0'. A space
	 between "NaN" and the optional perenthesis is not permitted.  */
      while (w > 0)
	{
	  *out = tolower (*p);
	  switch (*p)
	    {
	    case ' ':
	      if (dtp->u.p.blank_status == BLANK_ZERO)
		{
		  *out = '0';
		  break;
		}
	      *out = '\0';
	      if (seen_paren == 1)
	        goto bad_float;
	      break;
	    case '(':
	      seen_paren++;
	      *out = '\0';
	      break;
	    case ')':
	      if (seen_paren++ != 1)
		goto bad_float;
	      break;
	    default:
	      if (!isalnum (*out))
		goto bad_float;
	    }
	  --w;
	  ++p;
	  ++out;
	}
	 
      *out = '\0';
      
      if (seen_paren != 0 && seen_paren != 2)
	goto bad_float;

      if ((strcmp (save, "inf") == 0) || (strcmp (save, "infinity") == 0))
	{
	   if (seen_paren)
	     goto bad_float;
	}
      else if (strcmp (save, "nan") != 0)
	goto bad_float;

      convert_infnan (dtp, dest, buffer, length);
      if (buf_size > READF_TMP)
	free (buffer);
      return;
    }

  /* Process the mantissa string.  */
  while (w > 0)
    {
      switch (*p)
	{
	case ',':
	  if (dtp->u.p.current_unit->decimal_status != DECIMAL_COMMA)
	    goto bad_float;
	  /* Fall through.  */
	case '.':
	  if (seen_dp)
	    goto bad_float;
	  if (!seen_int_digit)
	    *(out++) = '0';
	  *(out++) = '.';
	  seen_dp = 1;
	  break;

	case ' ':
	  if (dtp->u.p.blank_status == BLANK_ZERO)
	    {
	      *(out++) = '0';
	      goto found_digit;
	    }
	  else if (dtp->u.p.blank_status == BLANK_NULL)
	    break;
	  else
	    /* TODO: Should we check instead that there are only trailing
	       blanks here, as is done below for exponents?  */
	    goto done;
	  /* Fall through.  */
	case '0':
	case '1':
	case '2':
	case '3':
	case '4':
	case '5':
	case '6':
	case '7':
	case '8':
	case '9':
	  *(out++) = *p;
found_digit:
	  if (!seen_dp)
	    seen_int_digit = 1;
	  else
	    seen_dec_digit = 1;
	  break;

	case '-':
	case '+':
	  goto exponent;

	case 'e':
	case 'E':
	case 'd':
	case 'D':
	case 'q':
	case 'Q':
	  ++p;
	  --w;
	  goto exponent;

	default:
	  goto bad_float;
	}

      ++p;
      --w;
    }
  
  /* No exponent has been seen, so we use the current scale factor.  */
  exponent = - dtp->u.p.scale_factor;
  goto done;

  /* At this point the start of an exponent has been found.  */
exponent:
  p = eat_leading_spaces (&w, (char*) p);
  if (*p == '-' || *p == '+')
    {
      if (*p == '-')
	exponent_sign = -1;
      ++p;
      --w;
    }

  /* At this point a digit string is required.  We calculate the value
     of the exponent in order to take account of the scale factor and
     the d parameter before explict conversion takes place.  */

  if (w == 0)
    {
      /* Extension: allow default exponent of 0 when omitted.  */
      if (dtp->common.flags & IOPARM_DT_DEC_EXT)
	goto done;
      else
	goto bad_float;
    }

  if (dtp->u.p.blank_status == BLANK_UNSPECIFIED)
    {
      while (w > 0 && isdigit (*p))
	{
	  exponent *= 10;
	  exponent += *p - '0';
	  ++p;
	  --w;
	}
	
      /* Only allow trailing blanks.  */
      while (w > 0)
	{
	  if (*p != ' ')
	    goto bad_float;
	  ++p;
	  --w;
	}
    }    
  else  /* BZ or BN status is enabled.  */
    {
      while (w > 0)
	{
	  if (*p == ' ')
	    {
	      if (dtp->u.p.blank_status == BLANK_ZERO)
		exponent *= 10;
	      else
		assert (dtp->u.p.blank_status == BLANK_NULL);
	    }
	  else if (!isdigit (*p))
	    goto bad_float;
	  else
	    {
	      exponent *= 10;
	      exponent += *p - '0';
	    }

	  ++p;
	  --w;
	}
    }

  exponent *= exponent_sign;

done:
  /* Use the precision specified in the format if no decimal point has been
     seen.  */
  if (!seen_dp)
    exponent -= f->u.real.d;

  /* Output a trailing '0' after decimal point if not yet found.  */
  if (seen_dp && !seen_dec_digit)
    *(out++) = '0';
  /* Handle input of style "E+NN" by inserting a 0 for the
     significand.  */
  else if (!seen_int_digit && !seen_dec_digit)
    {
      notify_std (&dtp->common, GFC_STD_LEGACY, 
		  "REAL input of style 'E+NN'");
      *(out++) = '0';
    }

  /* Print out the exponent to finish the reformatted number.  Maximum 4
     digits for the exponent.  */
  if (exponent != 0)
    {
      int dig;

      *(out++) = 'e';
      if (exponent < 0)
	{
	  *(out++) = '-';
	  exponent = - exponent;
	}

      if (exponent >= 10000)
	goto bad_float;

      for (dig = 3; dig >= 0; --dig)
	{
	  out[dig] = (char) ('0' + exponent % 10);
	  exponent /= 10;
	}
      out += 4;
    }
  *(out++) = '\0';

  /* Do the actual conversion.  */
  convert_real (dtp, dest, buffer, length);
  if (buf_size > READF_TMP)
    free (buffer);
  return;

  /* The value read is zero.  */
zero:
  switch (length)
    {
      case 4:
	*((GFC_REAL_4 *) dest) = 0.0;
	break;

      case 8:
	*((GFC_REAL_8 *) dest) = 0.0;
	break;

#ifdef HAVE_GFC_REAL_10
      case 10:
	*((GFC_REAL_10 *) dest) = 0.0;
	break;
#endif

#ifdef HAVE_GFC_REAL_16
      case 16:
	*((GFC_REAL_16 *) dest) = 0.0;
	break;
#endif

      default:
	internal_error (&dtp->common, "Unsupported real kind during IO");
    }
  return;

bad_float:
  if (buf_size > READF_TMP)
    free (buffer);
  generate_error (&dtp->common, LIBERROR_READ_VALUE,
		  "Bad value during floating point read");
  next_record (dtp, 1);
  return;
}


/* read_x()-- Deal with the X/TR descriptor.  We just read some data
   and never look at it. */

void
read_x (st_parameter_dt *dtp, size_t n)
{
  size_t length;
  int q, q2;

  if ((dtp->u.p.current_unit->pad_status == PAD_NO || is_internal_unit (dtp))
      && dtp->u.p.current_unit->bytes_left < (gfc_offset) n)
    n = dtp->u.p.current_unit->bytes_left;
    
  if (n == 0)
    return;

  length = n;

  if (is_internal_unit (dtp))
    {
      mem_alloc_r (dtp->u.p.current_unit->s, &length);
      if (unlikely (length < n))
	n = length;
      goto done;
    }

  if (dtp->u.p.sf_seen_eor)
    return;

  n = 0;
  while (n < length)
    {
      q = fbuf_getc (dtp->u.p.current_unit);
      if (q == EOF)
	break;
      else if (dtp->u.p.current_unit->flags.cc != CC_NONE
	       && (q == '\n' || q == '\r'))
	{
	  /* Unexpected end of line. Set the position.  */
	  dtp->u.p.sf_seen_eor = 1;

	  /* If we see an EOR during non-advancing I/O, we need to skip
	     the rest of the I/O statement.  Set the corresponding flag.  */
	  if (dtp->u.p.advance_status == ADVANCE_NO || dtp->u.p.seen_dollar)
	    dtp->u.p.eor_condition = 1;
	    
	  /* If we encounter a CR, it might be a CRLF.  */
	  if (q == '\r') /* Probably a CRLF */
	    {
	      /* See if there is an LF.  */
	      q2 = fbuf_getc (dtp->u.p.current_unit);
	      if (q2 == '\n')
		dtp->u.p.sf_seen_eor = 2;
	      else if (q2 != EOF) /* Oops, seek back.  */
		fbuf_seek (dtp->u.p.current_unit, -1, SEEK_CUR);
	    }
	  goto done;
	}
      n++;
    } 

 done:
  if (((dtp->common.flags & IOPARM_DT_HAS_SIZE) != 0) ||
      dtp->u.p.current_unit->has_size)
    dtp->u.p.current_unit->size_used += (GFC_IO_INT) n;
  dtp->u.p.current_unit->bytes_left -= n;
  dtp->u.p.current_unit->strm_pos += (gfc_offset) n;
}