mono/utils/strenc.c

   1 /*
   2  * strenc.c: string encoding conversions
   3  *
   4  * Author:
   5  *      Dick Porter (dick@ximian.com)
   6  *
   7  * (C) 2003 Ximian, Inc.
   8  */
   9
  10 #include <config.h>
  11 #include <glib.h>
  12 #include <string.h>
  13
  14 #include "strenc.h"
  15
  16 static const char trailingBytesForUTF8[256] = {
  17         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  18         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  19         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  20         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  21         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  22         0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
  23         1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
  24         2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,0,0
  25 };
  26
  27 /**
  28  * mono_unicode_from_external:
  29  * @in: pointers to the buffer.
  30  * @bytes: number of bytes in the string.
  31  *
  32  * Tries to turn a NULL-terminated string into UTF16.
  33  *
  34  * First, see if it's valid UTF8, in which case just turn it directly
  35  * into UTF16.  Next, run through the colon-separated encodings in
  36  * MONO_EXTERNAL_ENCODINGS and do an iconv conversion on each,
  37  * returning the first successful conversion to UTF16.  If no
  38  * conversion succeeds, return NULL.
  39  *
  40  * Callers must free the returned string if not NULL. bytes holds the number
  41  * of bytes in the returned string, not including the terminator.
  42  */
  43 gunichar2 *
  44 mono_unicode_from_external (const gchar *in, gsize *bytes)
  45 {
  46         gchar *res=NULL;
  47         gchar **encodings;
  48         const gchar *encoding_list;
  49         int i;
  50         glong lbytes;
  51
  52         if(in==NULL) {
  53                 return(NULL);
  54         }
  55
  56         encoding_list=g_getenv ("MONO_EXTERNAL_ENCODINGS");
  57         if(encoding_list==NULL) {
  58                 encoding_list = "";
  59         }
  60
  61         encodings=g_strsplit (encoding_list, ":", 0);
  62         for(i=0;encodings[i]!=NULL; i++) {
  63                 /* "default_locale" is a special case encoding */
  64                 if(!strcmp (encodings[i], "default_locale")) {
  65                         gchar *utf8=g_locale_to_utf8 (in, -1, NULL, NULL, NULL);
  66                         if(utf8!=NULL) {
  67                                 res=(gchar *) g_utf8_to_utf16 (utf8, -1, NULL, &lbytes, NULL);
  68                                 *bytes = (gsize) lbytes;
  69                         }
  70                         g_free (utf8);
  71                 } else {
  72                         /* Don't use UTF16 here. It returns the <FF FE> prepended to the string */
  73                         res = g_convert (in, strlen (in), "UTF8", encodings[i], NULL, bytes, NULL);
  74                         if (res != NULL) {
  75                                 gchar *ptr = res;
  76                                 res = (gchar *) g_utf8_to_utf16 (res, -1, NULL, &lbytes, NULL);
  77                                 *bytes = (gsize) lbytes;
  78                                 g_free (ptr);
  79                         }
  80                 }
  81
  82                 if(res!=NULL) {
  83                         g_strfreev (encodings);
  84                         *bytes *= 2;
  85                         return((gunichar2 *)res);
  86                 }
  87         }
  88
  89         g_strfreev (encodings);
  90
  91         if(g_utf8_validate (in, -1, NULL)) {
  92                 gunichar2 *unires=g_utf8_to_utf16 (in, -1, NULL, (glong *)bytes, NULL);
  93                 *bytes *= 2;
  94                 return(unires);
  95         }
  96
  97         return(NULL);
  98 }
  99
 100 /**
 101  * mono_utf8_from_external:
 102  * @in: pointer to the string buffer.
 103  *
 104  * Tries to turn a NULL-terminated string into UTF8.
 105  *
 106  * First, see if it's valid UTF8, in which case there's nothing more
 107  * to be done.  Next, run through the colon-separated encodings in
 108  * MONO_EXTERNAL_ENCODINGS and do an iconv conversion on each,
 109  * returning the first successful conversion to utf8.  If no
 110  * conversion succeeds, return NULL.
 111  *
 112  * Callers must free the returned string if not NULL.
 113  *
 114  * This function is identical to mono_unicode_from_external, apart
 115  * from returning utf8 not utf16; it's handy in a few places to work
 116  * in utf8.
 117  */
 118 gchar *mono_utf8_from_external (const gchar *in)
 119 {
 120         gchar *res=NULL;
 121         gchar **encodings;
 122         const gchar *encoding_list;
 123         int i;
 124
 125         if(in==NULL) {
 126                 return(NULL);
 127         }
 128
 129         encoding_list=g_getenv ("MONO_EXTERNAL_ENCODINGS");
 130         if(encoding_list==NULL) {
 131                 encoding_list = "";
 132         }
 133
 134         encodings=g_strsplit (encoding_list, ":", 0);
 135         for(i=0;encodings[i]!=NULL; i++) {
 136
 137                 /* "default_locale" is a special case encoding */
 138                 if(!strcmp (encodings[i], "default_locale")) {
 139                         res=g_locale_to_utf8 (in, -1, NULL, NULL, NULL);
 140                         if(res!=NULL && !g_utf8_validate (res, -1, NULL)) {
 141                                 g_free (res);
 142                                 res=NULL;
 143                         }
 144                 } else {
 145                         res=g_convert (in, -1, "UTF8", encodings[i], NULL,
 146                                        NULL, NULL);
 147                 }
 148
 149                 if(res!=NULL) {
 150                         g_strfreev (encodings);
 151                         return(res);
 152                 }
 153         }
 154
 155         g_strfreev (encodings);
 156
 157         if(g_utf8_validate (in, -1, NULL)) {
 158                 return(g_strdup (in));
 159         }
 160
 161         return(NULL);
 162 }
 163
 164 /**
 165  * mono_unicode_to_external:
 166  * @uni: an UTF16 string to conver to an external representation.
 167  *
 168  * Turns NULL-terminated UTF16 into either UTF8, or the first
 169  * working item in MONO_EXTERNAL_ENCODINGS if set.  If no conversions
 170  * work, then UTF8 is returned.
 171  *
 172  * Callers must free the returned string.
 173  */
 174 gchar *mono_unicode_to_external (const gunichar2 *uni)
 175 {
 176         gchar *utf8;
 177         const gchar *encoding_list;
 178
 179         /* Turn the unicode into utf8 to start with, because its
 180          * easier to work with gchar * than gunichar2 *
 181          */
 182         utf8=g_utf16_to_utf8 (uni, -1, NULL, NULL, NULL);
 183         g_assert (utf8!=NULL);
 184
 185         encoding_list=g_getenv ("MONO_EXTERNAL_ENCODINGS");
 186         if(encoding_list==NULL) {
 187                 /* Do UTF8 */
 188                 return(utf8);
 189         } else {
 190                 gchar *res, **encodings;
 191                 int i;
 192
 193                 encodings=g_strsplit (encoding_list, ":", 0);
 194                 for(i=0; encodings[i]!=NULL; i++) {
 195                         if(!strcmp (encodings[i], "default_locale")) {
 196                                 res=g_locale_from_utf8 (utf8, -1, NULL, NULL,
 197                                                         NULL);
 198                         } else {
 199                                 res=g_convert (utf8, -1, encodings[i], "UTF8",
 200                                                NULL, NULL, NULL);
 201                         }
 202
 203                         if(res!=NULL) {
 204                                 g_free (utf8);
 205                                 g_strfreev (encodings);
 206
 207                                 return(res);
 208                         }
 209                 }
 210
 211                 g_strfreev (encodings);
 212         }
 213
 214         /* Nothing else worked, so just return the utf8 */
 215         return(utf8);
 216 }
 217
 218 /**
 219  * mono_utf8_validate_and_len
 220  * @source: Pointer to putative UTF-8 encoded string.
 221  *
 222  * Checks @source for being valid UTF-8. @utf is assumed to be
 223  * null-terminated.
 224  *
 225  * Return value: true if @source is valid.
 226  * oEnd : will equal the null terminator at the end of the string if valid.
 227  *            if not valid, it will equal the first charater of the invalid sequence.
 228  * oLengh : will equal the length to @oEnd
 229  **/
 230 gboolean
 231 mono_utf8_validate_and_len (const gchar *source, glong* oLength, const gchar** oEnd)
 232 {
 233         gboolean retVal = TRUE;
 234         gboolean lastRet = TRUE;
 235         guchar* ptr = (guchar*) source;
 236         guchar* srcPtr;
 237         guint length;
 238         guchar a;
 239         *oLength = 0;
 240         while (*ptr != 0) {
 241                 length = trailingBytesForUTF8 [*ptr] + 1;
 242                 srcPtr = (guchar*) ptr + length;
 243                 switch (length) {
 244                 default: retVal = FALSE;
 245                 /* Everything else falls through when "TRUE"... */
 246                 case 4: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 247                                 if ((a == (guchar) 0xBF || a == (guchar) 0xBE) && *(srcPtr-1) == (guchar) 0xBF) {
 248                                 if (*(srcPtr-2) == (guchar) 0x8F || *(srcPtr-2) == (guchar) 0x9F ||
 249                                         *(srcPtr-2) == (guchar) 0xAF || *(srcPtr-2) == (guchar) 0xBF)
 250                                         retVal = FALSE;
 251                                 }
 252                 case 3: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 253                 case 2: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 254
 255                 switch (*ptr) {
 256                 /* no fall-through in this inner switch */
 257                 case 0xE0: if (a < (guchar) 0xA0) retVal = FALSE; break;
 258                 case 0xED: if (a > (guchar) 0x9F) retVal = FALSE; break;
 259                 case 0xEF: if (a == (guchar)0xB7 && (*(srcPtr+1) > (guchar) 0x8F && *(srcPtr+1) < 0xB0)) retVal = FALSE;
 260                                    if (a == (guchar)0xBF && (*(srcPtr+1) == (guchar) 0xBE || *(srcPtr+1) == 0xBF)) retVal = FALSE; break;
 261                 case 0xF0: if (a < (guchar) 0x90) retVal = FALSE; break;
 262                 case 0xF4: if (a > (guchar) 0x8F) retVal = FALSE; break;
 263                 default:   if (a < (guchar) 0x80) retVal = FALSE;
 264                 }
 265
 266                 case 1: if (*ptr >= (guchar ) 0x80 && *ptr < (guchar) 0xC2) retVal = FALSE;
 267                 }
 268                 if (*ptr > (guchar) 0xF4)
 269                         retVal = FALSE;
 270                 //If the string is invalid, set the end to the invalid byte.
 271                 if (!retVal && lastRet) {
 272                         if (oEnd != NULL)
 273                                 *oEnd = (gchar*) ptr;
 274                         lastRet = FALSE;
 275                 }
 276                 ptr += length;
 277                 (*oLength)++;
 278         }
 279         if (retVal && oEnd != NULL)
 280                 *oEnd = (gchar*) ptr;
 281         return retVal;
 282 }
 283
 284
 285 /**
 286  * mono_utf8_validate_and_len_with_bounds
 287  * @source: Pointer to putative UTF-8 encoded string.
 288  * @max_bytes: Max number of bytes that can be decoded. This function returns FALSE if
 289  * it needs to decode characters beyond that.
 290  *
 291  * Checks @source for being valid UTF-8. @utf is assumed to be
 292  * null-terminated.
 293  *
 294  * Return value: true if @source is valid.
 295  * oEnd : will equal the null terminator at the end of the string if valid.
 296  *            if not valid, it will equal the first charater of the invalid sequence.
 297  * oLengh : will equal the length to @oEnd
 298  **/
 299 gboolean
 300 mono_utf8_validate_and_len_with_bounds (const gchar *source, glong max_bytes, glong* oLength, const gchar** oEnd)
 301 {
 302         gboolean retVal = TRUE;
 303         gboolean lastRet = TRUE;
 304         guchar* ptr = (guchar*) source;
 305         guchar *end = ptr + max_bytes;
 306         guchar* srcPtr;
 307         guint length;
 308         guchar a;
 309         *oLength = 0;
 310
 311         if (max_bytes < 1) {
 312                 if (oEnd)
 313                         *oEnd = (gchar*) ptr;
 314                 return FALSE;
 315         }
 316
 317         while (*ptr != 0) {
 318                 length = trailingBytesForUTF8 [*ptr] + 1;
 319                 srcPtr = (guchar*) ptr + length;
 320
 321                 /* since *ptr is not zero we must ensure that we can decode the current char + the byte after
 322                    srcPtr points to the first byte after the current char.*/
 323                 if (srcPtr >= end) {
 324                         retVal = FALSE;
 325                         break;
 326                 }
 327                 switch (length) {
 328                 default: retVal = FALSE;
 329                 /* Everything else falls through when "TRUE"... */
 330                 case 4: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 331                                 if ((a == (guchar) 0xBF || a == (guchar) 0xBE) && *(srcPtr-1) == (guchar) 0xBF) {
 332                                 if (*(srcPtr-2) == (guchar) 0x8F || *(srcPtr-2) == (guchar) 0x9F ||
 333                                         *(srcPtr-2) == (guchar) 0xAF || *(srcPtr-2) == (guchar) 0xBF)
 334                                         retVal = FALSE;
 335                                 }
 336                 case 3: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 337                 case 2: if ((a = (*--srcPtr)) < (guchar) 0x80 || a > (guchar) 0xBF) retVal = FALSE;
 338
 339                 switch (*ptr) {
 340                 /* no fall-through in this inner switch */
 341                 case 0xE0: if (a < (guchar) 0xA0) retVal = FALSE; break;
 342                 case 0xED: if (a > (guchar) 0x9F) retVal = FALSE; break;
 343                 case 0xEF: if (a == (guchar)0xB7 && (*(srcPtr+1) > (guchar) 0x8F && *(srcPtr+1) < 0xB0)) retVal = FALSE;
 344                                    if (a == (guchar)0xBF && (*(srcPtr+1) == (guchar) 0xBE || *(srcPtr+1) == 0xBF)) retVal = FALSE; break;
 345                 case 0xF0: if (a < (guchar) 0x90) retVal = FALSE; break;
 346                 case 0xF4: if (a > (guchar) 0x8F) retVal = FALSE; break;
 347                 default:   if (a < (guchar) 0x80) retVal = FALSE;
 348                 }
 349
 350                 case 1: if (*ptr >= (guchar ) 0x80 && *ptr < (guchar) 0xC2) retVal = FALSE;
 351                 }
 352                 if (*ptr > (guchar) 0xF4)
 353                         retVal = FALSE;
 354                 //If the string is invalid, set the end to the invalid byte.
 355                 if (!retVal && lastRet) {
 356                         if (oEnd != NULL)
 357                                 *oEnd = (gchar*) ptr;
 358                         lastRet = FALSE;
 359                 }
 360                 ptr += length;
 361                 (*oLength)++;
 362         }
 363         if (retVal && oEnd != NULL)
 364                 *oEnd = (gchar*) ptr;
 365         return retVal;
 366 }
 367