[deliverable/binutils-gdb.git] / readline / mbutil.c

/* mbutil.c -- readline multibyte character utility functions */

/* Copyright (C) 2001-2009 Free Software Foundation, Inc.

   This file is part of the GNU Readline Library (Readline), a library
   for reading lines of text with interactive input and history editing.      

   Readline 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 of the License, or
   (at your option) any later version.

   Readline 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 Readline.  If not, see <http://www.gnu.org/licenses/>.
*/

#define READLINE_LIBRARY

#if defined (HAVE_CONFIG_H)
#  include <config.h>
#endif

#include <sys/types.h>
#include <fcntl.h>
#include "posixjmp.h"

#if defined (HAVE_UNISTD_H)
#  include <unistd.h>	   /* for _POSIX_VERSION */
#endif /* HAVE_UNISTD_H */

#if defined (HAVE_STDLIB_H)
#  include <stdlib.h>
#else
#  include "ansi_stdlib.h"
#endif /* HAVE_STDLIB_H */

#include <stdio.h>
#include <ctype.h>

/* System-specific feature definitions and include files. */
#include "rldefs.h"
#include "rlmbutil.h"

#if defined (TIOCSTAT_IN_SYS_IOCTL)
#  include <sys/ioctl.h>
#endif /* TIOCSTAT_IN_SYS_IOCTL */

/* Some standard library routines. */
#include "readline.h"

#include "rlprivate.h"
#include "xmalloc.h"

/* Declared here so it can be shared between the readline and history
   libraries. */
#if defined (HANDLE_MULTIBYTE)
int rl_byte_oriented = 0;
#else
int rl_byte_oriented = 1;
#endif

/* **************************************************************** */
/*								    */
/*		Multibyte Character Utility Functions		    */
/*								    */
/* **************************************************************** */

#if defined(HANDLE_MULTIBYTE)

static int
_rl_find_next_mbchar_internal (string, seed, count, find_non_zero)
     char *string;
     int seed, count, find_non_zero;
{
  size_t tmp, len;
  mbstate_t ps;
  int point;
  wchar_t wc;

  tmp = 0;

  memset(&ps, 0, sizeof (mbstate_t));
  if (seed < 0)
    seed = 0;
  if (count <= 0)
    return seed;

  point = seed + _rl_adjust_point (string, seed, &ps);
  /* if this is true, means that seed was not pointing to a byte indicating
     the beginning of a multibyte character.  Correct the point and consume
     one char. */
  if (seed < point)
    count--;

  while (count > 0)  
    {
      len = strlen (string + point);
      if (len == 0)
	break;
      tmp = mbrtowc (&wc, string+point, len, &ps);
      if (MB_INVALIDCH ((size_t)tmp))
	{
	  /* invalid bytes. assume a byte represents a character */
	  point++;
	  count--;
	  /* reset states. */
	  memset(&ps, 0, sizeof(mbstate_t));
	}
      else if (MB_NULLWCH (tmp))
	break;			/* found wide '\0' */
      else
	{
	  /* valid bytes */
	  point += tmp;
	  if (find_non_zero)
	    {
	      if (wcwidth (wc) == 0)
		continue;
	      else
		count--;
	    }
	  else
	    count--;
	}
    }

  if (find_non_zero)
    {
      tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
      while (MB_NULLWCH (tmp) == 0 && MB_INVALIDCH (tmp) == 0 && wcwidth (wc) == 0)
	{
	  point += tmp;
	  tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
	}
    }

  return point;
}

static int
_rl_find_prev_mbchar_internal (string, seed, find_non_zero)
     char *string;
     int seed, find_non_zero;
{
  mbstate_t ps;
  int prev, non_zero_prev, point, length;
  size_t tmp;
  wchar_t wc;

  memset(&ps, 0, sizeof(mbstate_t));
  length = strlen(string);
  
  if (seed < 0)
    return 0;
  else if (length < seed)
    return length;

  prev = non_zero_prev = point = 0;
  while (point < seed)
    {
      tmp = mbrtowc (&wc, string + point, length - point, &ps);
      if (MB_INVALIDCH ((size_t)tmp))
	{
	  /* in this case, bytes are invalid or shorted to compose
	     multibyte char, so assume that the first byte represents
	     a single character anyway. */
	  tmp = 1;
	  /* clear the state of the byte sequence, because
	     in this case effect of mbstate is undefined  */
	  memset(&ps, 0, sizeof (mbstate_t));

	  /* Since we're assuming that this byte represents a single
	     non-zero-width character, don't forget about it. */
	  prev = point;
	}
      else if (MB_NULLWCH (tmp))
	break;			/* Found '\0' char.  Can this happen? */
      else
	{
	  if (find_non_zero)
	    {
	      if (wcwidth (wc) != 0)
		prev = point;
	    }
	  else
	    prev = point;  
	}

      point += tmp;
    }

  return prev;
}

/* return the number of bytes parsed from the multibyte sequence starting
   at src, if a non-L'\0' wide character was recognized. It returns 0, 
   if a L'\0' wide character was recognized. It  returns (size_t)(-1), 
   if an invalid multibyte sequence was encountered. It returns (size_t)(-2) 
   if it couldn't parse a complete  multibyte character.  */
int
_rl_get_char_len (src, ps)
     char *src;
     mbstate_t *ps;
{
  size_t tmp;

  tmp = mbrlen((const char *)src, (size_t)strlen (src), ps);
  if (tmp == (size_t)(-2))
    {
      /* shorted to compose multibyte char */
      if (ps)
	memset (ps, 0, sizeof(mbstate_t));
      return -2;
    }
  else if (tmp == (size_t)(-1))
    {
      /* invalid to compose multibyte char */
      /* initialize the conversion state */
      if (ps)
	memset (ps, 0, sizeof(mbstate_t));
      return -1;
    }
  else if (tmp == (size_t)0)
    return 0;
  else
    return (int)tmp;
}

/* compare the specified two characters. If the characters matched,
   return 1. Otherwise return 0. */
int
_rl_compare_chars (buf1, pos1, ps1, buf2, pos2, ps2)
     char *buf1;
     int pos1;
     mbstate_t *ps1;
     char *buf2;
     int pos2;
     mbstate_t *ps2;
{
  int i, w1, w2;

  if ((w1 = _rl_get_char_len (&buf1[pos1], ps1)) <= 0 || 
	(w2 = _rl_get_char_len (&buf2[pos2], ps2)) <= 0 ||
	(w1 != w2) ||
	(buf1[pos1] != buf2[pos2]))
    return 0;

  for (i = 1; i < w1; i++)
    if (buf1[pos1+i] != buf2[pos2+i])
      return 0;

  return 1;
}

/* adjust pointed byte and find mbstate of the point of string.
   adjusted point will be point <= adjusted_point, and returns
   differences of the byte(adjusted_point - point).
   if point is invalied (point < 0 || more than string length),
   it returns -1 */
int
_rl_adjust_point(string, point, ps)
     char *string;
     int point;
     mbstate_t *ps;
{
  size_t tmp = 0;
  int length;
  int pos = 0;

  length = strlen(string);
  if (point < 0)
    return -1;
  if (length < point)
    return -1;
  
  while (pos < point)
    {
      tmp = mbrlen (string + pos, length - pos, ps);
      if (MB_INVALIDCH ((size_t)tmp))
	{
	  /* in this case, bytes are invalid or shorted to compose
	     multibyte char, so assume that the first byte represents
	     a single character anyway. */
	  pos++;
	  /* clear the state of the byte sequence, because
	     in this case effect of mbstate is undefined  */
	  if (ps)
	    memset (ps, 0, sizeof (mbstate_t));
	}
      else if (MB_NULLWCH (tmp))
	pos++;
      else
	pos += tmp;
    }

  return (pos - point);
}

int
_rl_is_mbchar_matched (string, seed, end, mbchar, length)
     char *string;
     int seed, end;
     char *mbchar;
     int length;
{
  int i;

  if ((end - seed) < length)
    return 0;

  for (i = 0; i < length; i++)
    if (string[seed + i] != mbchar[i])
      return 0;
  return 1;
}

wchar_t
_rl_char_value (buf, ind)
     char *buf;
     int ind;
{
  size_t tmp;
  wchar_t wc;
  mbstate_t ps;
  int l;

  if (MB_LEN_MAX == 1 || rl_byte_oriented)
    return ((wchar_t) buf[ind]);
  l = strlen (buf);
  if (ind >= l - 1)
    return ((wchar_t) buf[ind]);
  memset (&ps, 0, sizeof (mbstate_t));
  tmp = mbrtowc (&wc, buf + ind, l - ind, &ps);
  if (MB_INVALIDCH (tmp) || MB_NULLWCH (tmp))  
    return ((wchar_t) buf[ind]);
  return wc;
}
#endif /* HANDLE_MULTIBYTE */

/* Find next `count' characters started byte point of the specified seed.
   If flags is MB_FIND_NONZERO, we look for non-zero-width multibyte
   characters. */
#undef _rl_find_next_mbchar
int
_rl_find_next_mbchar (string, seed, count, flags)
     char *string;
     int seed, count, flags;
{
#if defined (HANDLE_MULTIBYTE)
  return _rl_find_next_mbchar_internal (string, seed, count, flags);
#else
  return (seed + count);
#endif
}

/* Find previous character started byte point of the specified seed.
   Returned point will be point <= seed.  If flags is MB_FIND_NONZERO,
   we look for non-zero-width multibyte characters. */
#undef _rl_find_prev_mbchar
int
_rl_find_prev_mbchar (string, seed, flags)
     char *string;
     int seed, flags;
{
#if defined (HANDLE_MULTIBYTE)
  return _rl_find_prev_mbchar_internal (string, seed, flags);
#else
  return ((seed == 0) ? seed : seed - 1);
#endif
}
Commit	Line	Data
	1	/* mbutil.c -- readline multibyte character utility functions */
	2
	3	/* Copyright (C) 2001-2009 Free Software Foundation, Inc.
	4
	5	This file is part of the GNU Readline Library (Readline), a library
	6	for reading lines of text with interactive input and history editing.
	7
	8	Readline is free software: you can redistribute it and/or modify
	9	it under the terms of the GNU General Public License as published by
	10	the Free Software Foundation, either version 3 of the License, or
	11	(at your option) any later version.
	12
	13	Readline is distributed in the hope that it will be useful,
	14	but WITHOUT ANY WARRANTY; without even the implied warranty of
	15	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	16	GNU General Public License for more details.
	17
	18	You should have received a copy of the GNU General Public License
	19	along with Readline. If not, see <http://www.gnu.org/licenses/>.
	20	*/
	21
	22	#define READLINE_LIBRARY
	23
	24	#if defined (HAVE_CONFIG_H)
	25	# include <config.h>
	26	#endif
	27
	28	#include <sys/types.h>
	29	#include <fcntl.h>
	30	#include "posixjmp.h"
	31
	32	#if defined (HAVE_UNISTD_H)
	33	# include <unistd.h> /* for _POSIX_VERSION */
	34	#endif /* HAVE_UNISTD_H */
	35
	36	#if defined (HAVE_STDLIB_H)
	37	# include <stdlib.h>
	38	#else
	39	# include "ansi_stdlib.h"
	40	#endif /* HAVE_STDLIB_H */
	41
	42	#include <stdio.h>
	43	#include <ctype.h>
	44
	45	/* System-specific feature definitions and include files. */
	46	#include "rldefs.h"
	47	#include "rlmbutil.h"
	48
	49	#if defined (TIOCSTAT_IN_SYS_IOCTL)
	50	# include <sys/ioctl.h>
	51	#endif /* TIOCSTAT_IN_SYS_IOCTL */
	52
	53	/* Some standard library routines. */
	54	#include "readline.h"
	55
	56	#include "rlprivate.h"
	57	#include "xmalloc.h"
	58
	59	/* Declared here so it can be shared between the readline and history
	60	libraries. */
	61	#if defined (HANDLE_MULTIBYTE)
	62	int rl_byte_oriented = 0;
	63	#else
	64	int rl_byte_oriented = 1;
	65	#endif
	66
	67	/* **************************************************************** */
	68	/* */
	69	/* Multibyte Character Utility Functions */
	70	/* */
	71	/* **************************************************************** */
	72
	73	#if defined(HANDLE_MULTIBYTE)
	74
	75	static int
	76	_rl_find_next_mbchar_internal (string, seed, count, find_non_zero)
	77	char *string;
	78	int seed, count, find_non_zero;
	79	{
	80	size_t tmp, len;
	81	mbstate_t ps;
	82	int point;
	83	wchar_t wc;
	84
	85	tmp = 0;
	86
	87	memset(&ps, 0, sizeof (mbstate_t));
	88	if (seed < 0)
	89	seed = 0;
	90	if (count <= 0)
	91	return seed;
	92
	93	point = seed + _rl_adjust_point (string, seed, &ps);
	94	/* if this is true, means that seed was not pointing to a byte indicating
	95	the beginning of a multibyte character. Correct the point and consume
	96	one char. */
	97	if (seed < point)
	98	count--;
	99
	100	while (count > 0)
	101	{
	102	len = strlen (string + point);
	103	if (len == 0)
	104	break;
	105	tmp = mbrtowc (&wc, string+point, len, &ps);
	106	if (MB_INVALIDCH ((size_t)tmp))
	107	{
	108	/* invalid bytes. assume a byte represents a character */
	109	point++;
	110	count--;
	111	/* reset states. */
	112	memset(&ps, 0, sizeof(mbstate_t));
	113	}
	114	else if (MB_NULLWCH (tmp))
	115	break; /* found wide '\0' */
	116	else
	117	{
	118	/* valid bytes */
	119	point += tmp;
	120	if (find_non_zero)
	121	{
	122	if (wcwidth (wc) == 0)
	123	continue;
	124	else
	125	count--;
	126	}
	127	else
	128	count--;
	129	}
	130	}
	131
	132	if (find_non_zero)
	133	{
	134	tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
	135	while (MB_NULLWCH (tmp) == 0 && MB_INVALIDCH (tmp) == 0 && wcwidth (wc) == 0)
	136	{
	137	point += tmp;
	138	tmp = mbrtowc (&wc, string + point, strlen (string + point), &ps);
	139	}
	140	}
	141
	142	return point;
	143	}
	144
	145	static int
	146	_rl_find_prev_mbchar_internal (string, seed, find_non_zero)
	147	char *string;
	148	int seed, find_non_zero;
	149	{
	150	mbstate_t ps;
	151	int prev, non_zero_prev, point, length;
	152	size_t tmp;
	153	wchar_t wc;
	154
	155	memset(&ps, 0, sizeof(mbstate_t));
	156	length = strlen(string);
	157
	158	if (seed < 0)
	159	return 0;
	160	else if (length < seed)
	161	return length;
	162
	163	prev = non_zero_prev = point = 0;
	164	while (point < seed)
	165	{
	166	tmp = mbrtowc (&wc, string + point, length - point, &ps);
	167	if (MB_INVALIDCH ((size_t)tmp))
	168	{
	169	/* in this case, bytes are invalid or shorted to compose
	170	multibyte char, so assume that the first byte represents
	171	a single character anyway. */
	172	tmp = 1;
	173	/* clear the state of the byte sequence, because
	174	in this case effect of mbstate is undefined */
	175	memset(&ps, 0, sizeof (mbstate_t));
	176
	177	/* Since we're assuming that this byte represents a single
	178	non-zero-width character, don't forget about it. */
	179	prev = point;
	180	}
	181	else if (MB_NULLWCH (tmp))
	182	break; /* Found '\0' char. Can this happen? */
	183	else
	184	{
	185	if (find_non_zero)
	186	{
	187	if (wcwidth (wc) != 0)
	188	prev = point;
	189	}
	190	else
	191	prev = point;
	192	}
	193
	194	point += tmp;
	195	}
	196
	197	return prev;
	198	}
	199
	200	/* return the number of bytes parsed from the multibyte sequence starting
	201	at src, if a non-L'\0' wide character was recognized. It returns 0,
	202	if a L'\0' wide character was recognized. It returns (size_t)(-1),
	203	if an invalid multibyte sequence was encountered. It returns (size_t)(-2)
	204	if it couldn't parse a complete multibyte character. */
	205	int
	206	_rl_get_char_len (src, ps)
	207	char *src;
	208	mbstate_t *ps;
	209	{
	210	size_t tmp;
	211
	212	tmp = mbrlen((const char *)src, (size_t)strlen (src), ps);
	213	if (tmp == (size_t)(-2))
	214	{
	215	/* shorted to compose multibyte char */
	216	if (ps)
	217	memset (ps, 0, sizeof(mbstate_t));
	218	return -2;
	219	}
	220	else if (tmp == (size_t)(-1))
	221	{
	222	/* invalid to compose multibyte char */
	223	/* initialize the conversion state */
	224	if (ps)
	225	memset (ps, 0, sizeof(mbstate_t));
	226	return -1;
	227	}
	228	else if (tmp == (size_t)0)
	229	return 0;
	230	else
	231	return (int)tmp;
	232	}
	233
	234	/* compare the specified two characters. If the characters matched,
	235	return 1. Otherwise return 0. */
	236	int
	237	_rl_compare_chars (buf1, pos1, ps1, buf2, pos2, ps2)
	238	char *buf1;
	239	int pos1;
	240	mbstate_t *ps1;
	241	char *buf2;
	242	int pos2;
	243	mbstate_t *ps2;
	244	{
	245	int i, w1, w2;
	246
	247	if ((w1 = _rl_get_char_len (&buf1[pos1], ps1)) <= 0 \|\|
	248	(w2 = _rl_get_char_len (&buf2[pos2], ps2)) <= 0 \|\|
	249	(w1 != w2) \|\|
	250	(buf1[pos1] != buf2[pos2]))
	251	return 0;
	252
	253	for (i = 1; i < w1; i++)
	254	if (buf1[pos1+i] != buf2[pos2+i])
	255	return 0;
	256
	257	return 1;
	258	}
	259
	260	/* adjust pointed byte and find mbstate of the point of string.
	261	adjusted point will be point <= adjusted_point, and returns
	262	differences of the byte(adjusted_point - point).
	263	if point is invalied (point < 0 \|\| more than string length),
	264	it returns -1 */
	265	int
	266	_rl_adjust_point(string, point, ps)
	267	char *string;
	268	int point;
	269	mbstate_t *ps;
	270	{
	271	size_t tmp = 0;
	272	int length;
	273	int pos = 0;
	274
	275	length = strlen(string);
	276	if (point < 0)
	277	return -1;
	278	if (length < point)
	279	return -1;
	280
	281	while (pos < point)
	282	{
	283	tmp = mbrlen (string + pos, length - pos, ps);
	284	if (MB_INVALIDCH ((size_t)tmp))
	285	{
	286	/* in this case, bytes are invalid or shorted to compose
	287	multibyte char, so assume that the first byte represents
	288	a single character anyway. */
	289	pos++;
	290	/* clear the state of the byte sequence, because
	291	in this case effect of mbstate is undefined */
	292	if (ps)
	293	memset (ps, 0, sizeof (mbstate_t));
	294	}
	295	else if (MB_NULLWCH (tmp))
	296	pos++;
	297	else
	298	pos += tmp;
	299	}
	300
	301	return (pos - point);
	302	}
	303
	304	int
	305	_rl_is_mbchar_matched (string, seed, end, mbchar, length)
	306	char *string;
	307	int seed, end;
	308	char *mbchar;
	309	int length;
	310	{
	311	int i;
	312
	313	if ((end - seed) < length)
	314	return 0;
	315
	316	for (i = 0; i < length; i++)
	317	if (string[seed + i] != mbchar[i])
	318	return 0;
	319	return 1;
	320	}
	321
	322	wchar_t
	323	_rl_char_value (buf, ind)
	324	char *buf;
	325	int ind;
	326	{
	327	size_t tmp;
	328	wchar_t wc;
	329	mbstate_t ps;
	330	int l;
	331
	332	if (MB_LEN_MAX == 1 \|\| rl_byte_oriented)
	333	return ((wchar_t) buf[ind]);
	334	l = strlen (buf);
	335	if (ind >= l - 1)
	336	return ((wchar_t) buf[ind]);
	337	memset (&ps, 0, sizeof (mbstate_t));
	338	tmp = mbrtowc (&wc, buf + ind, l - ind, &ps);
	339	if (MB_INVALIDCH (tmp) \|\| MB_NULLWCH (tmp))
	340	return ((wchar_t) buf[ind]);
	341	return wc;
	342	}
	343	#endif /* HANDLE_MULTIBYTE */
	344
	345	/* Find next `count' characters started byte point of the specified seed.
	346	If flags is MB_FIND_NONZERO, we look for non-zero-width multibyte
	347	characters. */
	348	#undef _rl_find_next_mbchar
	349	int
	350	_rl_find_next_mbchar (string, seed, count, flags)
	351	char *string;
	352	int seed, count, flags;
	353	{
	354	#if defined (HANDLE_MULTIBYTE)
	355	return _rl_find_next_mbchar_internal (string, seed, count, flags);
	356	#else
	357	return (seed + count);
	358	#endif
	359	}
	360
	361	/* Find previous character started byte point of the specified seed.
	362	Returned point will be point <= seed. If flags is MB_FIND_NONZERO,
	363	we look for non-zero-width multibyte characters. */
	364	#undef _rl_find_prev_mbchar
	365	int
	366	_rl_find_prev_mbchar (string, seed, flags)
	367	char *string;
	368	int seed, flags;
	369	{
	370	#if defined (HANDLE_MULTIBYTE)
	371	return _rl_find_prev_mbchar_internal (string, seed, flags);
	372	#else
	373	return ((seed == 0) ? seed : seed - 1);
	374	#endif
	375	}