1/**
2 * \file unicode.c
3 *
4 * This file contains general Unicode string manipulation functions.
5 * It mainly consist of functions for converting between UCS-2 (used on
6 * the devices) and UTF-8 (used by several applications).
7 *
8 * For a deeper understanding of Unicode encoding formats see the
9 * Wikipedia entries for
10 * <a href="http://en.wikipedia.org/wiki/UTF-16/UCS-2">UTF-16/UCS-2</a>
11 * and <a href="http://en.wikipedia.org/wiki/UTF-8">UTF-8</a>.
12 *
13 * Copyright (C) 2005-2009 Linus Walleij <triad@df.lth.se>
14 *
15 * This library is free software; you can redistribute it and/or
16 * modify it under the terms of the GNU Lesser General Public
17 * License as published by the Free Software Foundation; either
18 * version 2 of the License, or (at your option) any later version.
19 *
20 * This library is distributed in the hope that it will be useful,
21 * but WITHOUT ANY WARRANTY; without even the implied warranty of
22 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
23 * Lesser General Public License for more details.
24 *
25 * You should have received a copy of the GNU Lesser General Public
26 * License along with this library; if not, write to the
27 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
28 * Boston, MA 02111-1307, USA.
29 *
30 */
31
32#include <config.h>
33#include <stdlib.h>
34#include <string.h>
35#ifdef HAVE_ICONV
36#include "iconv.h"
37#else
38#error "libmtp unicode.c needs fixing to work without iconv()!"
39#endif
40#include "libmtp.h"
41#include "unicode.h"
42#include "util.h"
43#include "ptp.h"
44
45/**
46 * The size of the buffer (in characters) used for creating string copies.
47 */
48#define STRING_BUFFER_LENGTH 1024
49
50/**
51 * Gets the length (in characters, not bytes) of a unicode
52 * UCS-2 string, eg a string which physically is 0x00 0x41 0x00 0x00
53 * will return a value of 1.
54 *
55 * @param unicstr a UCS-2 Unicode string
56 * @return the length of the string, in number of characters. If you
57 *         want to know the length in bytes, multiply this by two and
58 *         add two (for zero terminator).
59 */
60int ucs2_strlen(uint16_t const * const unicstr)
61{
62  int length;
63
64  /* Unicode strings are terminated with 2 * 0x00 */
65  for(length = 0; unicstr[length] != 0x0000U; length ++);
66  return length;
67}
68
69/**
70 * Converts a big-endian UTF-16 2-byte string
71 * to a UTF-8 string. Actually just a UCS-2 internal conversion
72 * routine that strips off the BOM if there is one.
73 *
74 * @param device a pointer to the current device.
75 * @param unicstr the UTF-16 unicode string to convert
76 * @return a UTF-8 string.
77 */
78char *utf16_to_utf8(LIBMTP_mtpdevice_t *device, const uint16_t *unicstr)
79{
80  PTPParams *params = (PTPParams *) device->params;
81  char *stringp = (char *) unicstr;
82  char loclstr[STRING_BUFFER_LENGTH*3+1]; // UTF-8 encoding is max 3 bytes per UCS2 char.
83  char *locp = loclstr;
84  size_t nconv;
85  size_t convlen = (ucs2_strlen(unicstr)+1) * sizeof(uint16_t); // UCS-2 is 16 bit wide, include terminator
86  size_t convmax = STRING_BUFFER_LENGTH*3;
87
88  loclstr[0]='\0';
89  /* Do the conversion.  */
90  nconv = iconv(params->cd_ucs2_to_locale, &stringp, &convlen, &locp, &convmax);
91  if (nconv == (size_t) -1) {
92    // Return partial string anyway.
93    *locp = '\0';
94  }
95  loclstr[STRING_BUFFER_LENGTH*3] = '\0';
96  // Strip off any BOM, it's totally useless...
97  if ((uint8_t) loclstr[0] == 0xEFU && (uint8_t) loclstr[1] == 0xBBU && (uint8_t) loclstr[2] == 0xBFU) {
98    return strdup(loclstr+3);
99  }
100  return strdup(loclstr);
101}
102
103/**
104 * Converts a UTF-8 string to a big-endian UTF-16 2-byte string
105 * Actually just a UCS-2 internal conversion.
106 *
107 * @param device a pointer to the current device.
108 * @param localstr the UTF-8 unicode string to convert
109 * @return a UTF-16 string.
110 */
111uint16_t *utf8_to_utf16(LIBMTP_mtpdevice_t *device, const char *localstr)
112{
113  PTPParams *params = (PTPParams *) device->params;
114  char *stringp = (char *) localstr; // cast away "const"
115  char unicstr[(STRING_BUFFER_LENGTH+1)*2]; // UCS2 encoding is 2 bytes per UTF-8 char.
116  char *unip = unicstr;
117  size_t nconv = 0;
118  size_t convlen = strlen(localstr)+1; // utf8 bytes, include terminator
119  size_t convmax = STRING_BUFFER_LENGTH*2;
120
121  unicstr[0]='\0';
122  unicstr[1]='\0';
123
124  /* Do the conversion.  */
125  nconv = iconv(params->cd_locale_to_ucs2, &stringp, &convlen, &unip, &convmax);
126
127  if (nconv == (size_t) -1) {
128    // Return partial string anyway.
129    unip[0] = '\0';
130    unip[1] = '\0';
131  }
132  // make sure the string is null terminated
133  unicstr[STRING_BUFFER_LENGTH*2] = '\0';
134  unicstr[STRING_BUFFER_LENGTH*2+1] = '\0';
135
136  // allocate the string to be returned
137  // Note: can't use strdup since every other byte is a null byte
138  int ret_len = ucs2_strlen((uint16_t*)unicstr)*sizeof(uint16_t)+2;
139  uint16_t* ret = malloc(ret_len);
140  memcpy(ret,unicstr,(size_t)ret_len);
141  return ret;
142}
143
144/**
145 * This helper function simply removes any consecutive chars
146 * > 0x7F and replace then with an underscore. In UTF-8
147 * consequtive chars > 0x7F represent one single character so
148 * it has to be done like this (and it's elegant). It will only
149 * shrink the string in size so no copying is needed.
150 */
151void strip_7bit_from_utf8(char *str)
152{
153  int i,j,k;
154  i = 0;
155  j = 0;
156  k = strlen(str);
157  while (i < k) {
158    if ((uint8_t) str[i] > 0x7FU) {
159      str[j] = '_';
160      i++;
161      // Skip over any consequtive > 0x7F chars.
162      while((uint8_t) str[i] > 0x7FU) {
163	i++;
164      }
165    } else {
166      str[j] = str[i];
167      i++;
168    }
169    j++;
170  }
171  // Terminate stripped string...
172  str[j] = '\0';
173}
174