| 1 |
/* |
| 2 |
* Copyright (C) 1999-2001, 2008, 2016 Free Software Foundation, Inc. |
| 3 |
* This file is part of the GNU LIBICONV Library. |
| 4 |
* |
| 5 |
* The GNU LIBICONV Library is free software; you can redistribute it |
| 6 |
* and/or modify it under the terms of the GNU Library General Public |
| 7 |
* License as published by the Free Software Foundation; either version 2 |
| 8 |
* of the License, or (at your option) any later version. |
| 9 |
* |
| 10 |
* The GNU LIBICONV Library is distributed in the hope that it will be |
| 11 |
* useful, but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 13 |
* Library General Public License for more details. |
| 14 |
* |
| 15 |
* You should have received a copy of the GNU Library General Public |
| 16 |
* License along with the GNU LIBICONV Library; see the file COPYING.LIB. |
| 17 |
* If not, see <http://www.gnu.org/licenses/>. |
| 18 |
*/ |
| 19 |
|
| 20 |
/* |
| 21 |
* UTF-7 |
| 22 |
*/ |
| 23 |
|
| 24 |
/* Specification: RFC 2152 (and old RFC 1641, RFC 1642) */ |
| 25 |
/* The original Base64 encoding is defined in RFC 2045. */ |
| 26 |
|
| 27 |
/* Set of direct characters: |
| 28 |
* A-Z a-z 0-9 ' ( ) , - . / : ? space tab lf cr |
| 29 |
*/ |
| 30 |
static const unsigned char direct_tab[128/8] = { |
| 31 |
0x00, 0x26, 0x00, 0x00, 0x81, 0xf3, 0xff, 0x87, |
| 32 |
0xfe, 0xff, 0xff, 0x07, 0xfe, 0xff, 0xff, 0x07, |
| 33 |
}; |
| 34 |
#define isdirect(ch) ((ch) < 128 && ((direct_tab[(ch)>>3] >> (ch & 7)) & 1)) |
| 35 |
|
| 36 |
/* Set of direct and optional direct characters: |
| 37 |
* A-Z a-z 0-9 ' ( ) , - . / : ? space tab lf cr |
| 38 |
* ! " # $ % & * ; < = > @ [ ] ^ _ ` { | } |
| 39 |
*/ |
| 40 |
static const unsigned char xdirect_tab[128/8] = { |
| 41 |
0x00, 0x26, 0x00, 0x00, 0xff, 0xf7, 0xff, 0xff, |
| 42 |
0xff, 0xff, 0xff, 0xef, 0xff, 0xff, 0xff, 0x3f, |
| 43 |
}; |
| 44 |
#define isxdirect(ch) ((ch) < 128 && ((xdirect_tab[(ch)>>3] >> (ch & 7)) & 1)) |
| 45 |
|
| 46 |
/* Set of base64 characters, extended: |
| 47 |
* A-Z a-z 0-9 + / - |
| 48 |
*/ |
| 49 |
static const unsigned char xbase64_tab[128/8] = { |
| 50 |
0x00, 0x00, 0x00, 0x00, 0x00, 0xa8, 0xff, 0x03, |
| 51 |
0xfe, 0xff, 0xff, 0x07, 0xfe, 0xff, 0xff, 0x07, |
| 52 |
}; |
| 53 |
#define isxbase64(ch) ((ch) < 128 && ((xbase64_tab[(ch)>>3] >> (ch & 7)) & 1)) |
| 54 |
|
| 55 |
/* |
| 56 |
* The state is structured as follows: |
| 57 |
* bit 1..0: shift |
| 58 |
* bit 7..2: data |
| 59 |
* Precise meaning: |
| 60 |
* shift data |
| 61 |
* 0 0 not inside base64 encoding |
| 62 |
* 1 0 inside base64, no pending bits |
| 63 |
* 2 XXXX00 inside base64, 4 bits remain from 2nd byte |
| 64 |
* 3 XX0000 inside base64, 2 bits remain from 3rd byte |
| 65 |
*/ |
| 66 |
|
| 67 |
static int |
| 68 |
utf7_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n) |
| 69 |
{ |
| 70 |
state_t state = conv->istate; |
| 71 |
int count = 0; /* number of input bytes already read */ |
| 72 |
if (state & 3) |
| 73 |
goto active; |
| 74 |
else |
| 75 |
goto inactive; |
| 76 |
|
| 77 |
inactive: |
| 78 |
{ |
| 79 |
/* Here (state & 3) == 0 */ |
| 80 |
if (n < count+1) |
| 81 |
goto none; |
| 82 |
{ |
| 83 |
unsigned char c = *s; |
| 84 |
if (isxdirect(c)) { |
| 85 |
*pwc = (ucs4_t) c; |
| 86 |
conv->istate = state; |
| 87 |
return count+1; |
| 88 |
} |
| 89 |
if (c == '+') { |
| 90 |
if (n < count+2) |
| 91 |
goto none; |
| 92 |
if (s[1] == '-') { |
| 93 |
*pwc = (ucs4_t) '+'; |
| 94 |
conv->istate = state; |
| 95 |
return count+2; |
| 96 |
} |
| 97 |
s++; count++; |
| 98 |
state = 1; |
| 99 |
goto active; |
| 100 |
} |
| 101 |
goto ilseq; |
| 102 |
} |
| 103 |
} |
| 104 |
|
| 105 |
active: |
| 106 |
{ |
| 107 |
/* base64 encoding active */ |
| 108 |
unsigned int wc = 0; |
| 109 |
state_t base64state = state; |
| 110 |
unsigned int kmax = 2; /* number of payload bytes to read */ |
| 111 |
unsigned int k = 0; /* number of payload bytes already read */ |
| 112 |
unsigned int base64count = 0; /* number of base64 bytes already read */ |
| 113 |
for (;;) { |
| 114 |
unsigned char c = *s; |
| 115 |
unsigned int i; |
| 116 |
if (c >= 'A' && c <= 'Z') |
| 117 |
i = c-'A'; |
| 118 |
else if (c >= 'a' && c <= 'z') |
| 119 |
i = c-'a'+26; |
| 120 |
else if (c >= '0' && c <= '9') |
| 121 |
i = c-'0'+52; |
| 122 |
else if (c == '+') |
| 123 |
i = 62; |
| 124 |
else if (c == '/') |
| 125 |
i = 63; |
| 126 |
else { |
| 127 |
/* c terminates base64 encoding */ |
| 128 |
if (base64state & -4) |
| 129 |
goto ilseq; /* data must be 0, otherwise illegal */ |
| 130 |
if (base64count) |
| 131 |
goto ilseq; /* partial UTF-16 characters are invalid */ |
| 132 |
if (c == '-') { |
| 133 |
s++; count++; |
| 134 |
} |
| 135 |
state = 0; |
| 136 |
goto inactive; |
| 137 |
} |
| 138 |
s++; base64count++; |
| 139 |
/* read 6 bits: 0 <= i < 64 */ |
| 140 |
switch (base64state & 3) { |
| 141 |
case 1: /* inside base64, no pending bits */ |
| 142 |
base64state = (i << 2) | 0; break; |
| 143 |
case 0: /* inside base64, 6 bits remain from 1st byte */ |
| 144 |
wc = (wc << 8) | (base64state & -4) | (i >> 4); k++; |
| 145 |
base64state = ((i & 15) << 4) | 2; break; |
| 146 |
case 2: /* inside base64, 4 bits remain from 2nd byte */ |
| 147 |
wc = (wc << 8) | (base64state & -4) | (i >> 2); k++; |
| 148 |
base64state = ((i & 3) << 6) | 3; break; |
| 149 |
case 3: /* inside base64, 2 bits remain from 3rd byte */ |
| 150 |
wc = (wc << 8) | (base64state & -4) | i; k++; |
| 151 |
base64state = 1; break; |
| 152 |
} |
| 153 |
if (k == kmax) { |
| 154 |
/* UTF-16: When we see a High Surrogate, we must also decode |
| 155 |
the following Low Surrogate. */ |
| 156 |
if (kmax == 2 && (wc >= 0xd800 && wc < 0xdc00)) |
| 157 |
kmax = 4; |
| 158 |
else |
| 159 |
break; |
| 160 |
} |
| 161 |
if (n < count+base64count+1) |
| 162 |
goto none; |
| 163 |
} |
| 164 |
/* Here k = kmax > 0, hence base64count > 0. */ |
| 165 |
if ((base64state & 3) == 0) abort(); |
| 166 |
if (kmax == 4) { |
| 167 |
ucs4_t wc1 = wc >> 16; |
| 168 |
ucs4_t wc2 = wc & 0xffff; |
| 169 |
if (!(wc1 >= 0xd800 && wc1 < 0xdc00)) abort(); |
| 170 |
if (!(wc2 >= 0xdc00 && wc2 < 0xe000)) goto ilseq; |
| 171 |
*pwc = 0x10000 + ((wc1 - 0xd800) << 10) + (wc2 - 0xdc00); |
| 172 |
} else { |
| 173 |
*pwc = wc; |
| 174 |
} |
| 175 |
conv->istate = base64state; |
| 176 |
return count+base64count; |
| 177 |
} |
| 178 |
|
| 179 |
none: |
| 180 |
conv->istate = state; |
| 181 |
return RET_TOOFEW(count); |
| 182 |
|
| 183 |
ilseq: |
| 184 |
conv->istate = state; |
| 185 |
return RET_SHIFT_ILSEQ(count); |
| 186 |
} |
| 187 |
|
| 188 |
/* |
| 189 |
* The state is structured as follows: |
| 190 |
* bit 1..0: shift |
| 191 |
* bit 7..2: data |
| 192 |
* Precise meaning: |
| 193 |
* shift data |
| 194 |
* 0 0 not inside base64 encoding |
| 195 |
* 1 0 inside base64, no pending bits |
| 196 |
* 2 XX00 inside base64, 2 bits known for 2nd byte |
| 197 |
* 3 XXXX inside base64, 4 bits known for 3rd byte |
| 198 |
*/ |
| 199 |
|
| 200 |
/* Define this to 1 if you want the so-called "optional direct" characters |
| 201 |
! " # $ % & * ; < = > @ [ ] ^ _ ` { | } |
| 202 |
to be encoded. Define to 0 if you want them to be passed straight through, |
| 203 |
like the so-called "direct" characters. |
| 204 |
We set this to 1 because it's safer. |
| 205 |
*/ |
| 206 |
#define UTF7_ENCODE_OPTIONAL_CHARS 1 |
| 207 |
|
| 208 |
static int |
| 209 |
utf7_wctomb (conv_t conv, unsigned char *r, ucs4_t iwc, size_t n) |
| 210 |
{ |
| 211 |
state_t state = conv->ostate; |
| 212 |
unsigned int wc = iwc; |
| 213 |
int count = 0; |
| 214 |
if (state & 3) |
| 215 |
goto active; |
| 216 |
|
| 217 |
/*inactive:*/ |
| 218 |
{ |
| 219 |
if (UTF7_ENCODE_OPTIONAL_CHARS ? isdirect(wc) : isxdirect(wc)) { |
| 220 |
r[0] = (unsigned char) wc; |
| 221 |
/*conv->ostate = state;*/ |
| 222 |
return 1; |
| 223 |
} else { |
| 224 |
*r++ = '+'; |
| 225 |
if (wc == '+') { |
| 226 |
if (n < 2) |
| 227 |
return RET_TOOSMALL; |
| 228 |
*r = '-'; |
| 229 |
/*conv->ostate = state;*/ |
| 230 |
return 2; |
| 231 |
} |
| 232 |
count = 1; |
| 233 |
state = 1; |
| 234 |
goto active; |
| 235 |
} |
| 236 |
} |
| 237 |
|
| 238 |
active: |
| 239 |
{ |
| 240 |
/* base64 encoding active */ |
| 241 |
if (UTF7_ENCODE_OPTIONAL_CHARS ? isdirect(wc) : isxdirect(wc)) { |
| 242 |
/* deactivate base64 encoding */ |
| 243 |
count += ((state & 3) >= 2 ? 1 : 0) + (isxbase64(wc) ? 1 : 0) + 1; |
| 244 |
if (n < count) |
| 245 |
return RET_TOOSMALL; |
| 246 |
if ((state & 3) >= 2) { |
| 247 |
unsigned int i = state & -4; |
| 248 |
unsigned char c; |
| 249 |
if (i < 26) |
| 250 |
c = i+'A'; |
| 251 |
else if (i < 52) |
| 252 |
c = i-26+'a'; |
| 253 |
else if (i < 62) |
| 254 |
c = i-52+'0'; |
| 255 |
else if (i == 62) |
| 256 |
c = '+'; |
| 257 |
else if (i == 63) |
| 258 |
c = '/'; |
| 259 |
else |
| 260 |
abort(); |
| 261 |
*r++ = c; |
| 262 |
} |
| 263 |
if (isxbase64(wc)) |
| 264 |
*r++ = '-'; |
| 265 |
state = 0; |
| 266 |
*r++ = (unsigned char) wc; |
| 267 |
conv->ostate = state; |
| 268 |
return count; |
| 269 |
} else { |
| 270 |
unsigned int k; /* number of payload bytes to write */ |
| 271 |
if (wc < 0x10000) { |
| 272 |
k = 2; |
| 273 |
count += ((state & 3) >= 2 ? 3 : 2); |
| 274 |
} else if (wc < 0x110000) { |
| 275 |
unsigned int wc1 = 0xd800 + ((wc - 0x10000) >> 10); |
| 276 |
unsigned int wc2 = 0xdc00 + ((wc - 0x10000) & 0x3ff); |
| 277 |
wc = (wc1 << 16) | wc2; |
| 278 |
k = 4; |
| 279 |
count += ((state & 3) >= 3 ? 6 : 5); |
| 280 |
} else |
| 281 |
return RET_ILUNI; |
| 282 |
if (n < count) |
| 283 |
return RET_TOOSMALL; |
| 284 |
for (;;) { |
| 285 |
unsigned int i; |
| 286 |
unsigned char c; |
| 287 |
switch (state & 3) { |
| 288 |
case 0: /* inside base64, 6 bits known for 4th byte */ |
| 289 |
c = (state & -4) >> 2; state = 1; break; |
| 290 |
case 1: /* inside base64, no pending bits */ |
| 291 |
i = (wc >> (8 * --k)) & 0xff; |
| 292 |
c = i >> 2; state = ((i & 3) << 4) | 2; break; |
| 293 |
case 2: /* inside base64, 2 bits known for 2nd byte */ |
| 294 |
i = (wc >> (8 * --k)) & 0xff; |
| 295 |
c = (state & -4) | (i >> 4); state = ((i & 15) << 2) | 3; break; |
| 296 |
case 3: /* inside base64, 4 bits known for 3rd byte */ |
| 297 |
i = (wc >> (8 * --k)) & 0xff; |
| 298 |
c = (state & -4) | (i >> 6); state = ((i & 63) << 2) | 0; break; |
| 299 |
default: abort(); /* stupid gcc */ |
| 300 |
} |
| 301 |
if (c < 26) |
| 302 |
c = c+'A'; |
| 303 |
else if (c < 52) |
| 304 |
c = c-26+'a'; |
| 305 |
else if (c < 62) |
| 306 |
c = c-52+'0'; |
| 307 |
else if (c == 62) |
| 308 |
c = '+'; |
| 309 |
else if (c == 63) |
| 310 |
c = '/'; |
| 311 |
else |
| 312 |
abort(); |
| 313 |
*r++ = c; |
| 314 |
if ((state & 3) && (k == 0)) |
| 315 |
break; |
| 316 |
} |
| 317 |
conv->ostate = state; |
| 318 |
return count; |
| 319 |
} |
| 320 |
} |
| 321 |
} |
| 322 |
|
| 323 |
static int |
| 324 |
utf7_reset (conv_t conv, unsigned char *r, size_t n) |
| 325 |
{ |
| 326 |
state_t state = conv->ostate; |
| 327 |
if (state & 3) { |
| 328 |
/* deactivate base64 encoding */ |
| 329 |
unsigned int count = ((state & 3) >= 2 ? 1 : 0) + 1; |
| 330 |
if (n < count) |
| 331 |
return RET_TOOSMALL; |
| 332 |
if ((state & 3) >= 2) { |
| 333 |
unsigned int i = state & -4; |
| 334 |
unsigned char c; |
| 335 |
if (i < 26) |
| 336 |
c = i+'A'; |
| 337 |
else if (i < 52) |
| 338 |
c = i-26+'a'; |
| 339 |
else if (i < 62) |
| 340 |
c = i-52+'0'; |
| 341 |
else if (i == 62) |
| 342 |
c = '+'; |
| 343 |
else if (i == 63) |
| 344 |
c = '/'; |
| 345 |
else |
| 346 |
abort(); |
| 347 |
*r++ = c; |
| 348 |
} |
| 349 |
*r++ = '-'; |
| 350 |
/* conv->ostate = 0; will be done by the caller */ |
| 351 |
return count; |
| 352 |
} else |
| 353 |
return 0; |
| 354 |
} |