xref: /freebsd/crypto/heimdal/lib/wind/utf8.c (revision 6be3386466ab79a84b48429ae66244f21526d3df)
1 /*
2  * Copyright (c) 2004, 2006, 2007, 2008 Kungliga Tekniska Högskolan
3  * (Royal Institute of Technology, Stockholm, Sweden).
4  * All rights reserved.
5  *
6  * Redistribution and use in source and binary forms, with or without
7  * modification, are permitted provided that the following conditions
8  * are met:
9  *
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  *
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  *
17  * 3. Neither the name of the Institute nor the names of its contributors
18  *    may be used to endorse or promote products derived from this software
19  *    without specific prior written permission.
20  *
21  * THIS SOFTWARE IS PROVIDED BY THE INSTITUTE AND CONTRIBUTORS ``AS IS'' AND
22  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24  * ARE DISCLAIMED.  IN NO EVENT SHALL THE INSTITUTE OR CONTRIBUTORS BE LIABLE
25  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31  * SUCH DAMAGE.
32  */
33 
34 #include <config.h>
35 #include "windlocl.h"
36 
37 static int
38 utf8toutf32(const unsigned char **pp, uint32_t *out)
39 {
40     const unsigned char *p = *pp;
41     unsigned c = *p;
42 
43     if (c & 0x80) {
44 	if ((c & 0xE0) == 0xC0) {
45 	    const unsigned c2 = *++p;
46 	    if ((c2 & 0xC0) == 0x80) {
47 		*out =  ((c  & 0x1F) << 6)
48 		    | (c2 & 0x3F);
49 	    } else {
50 		return WIND_ERR_INVALID_UTF8;
51 	    }
52 	} else if ((c & 0xF0) == 0xE0) {
53 	    const unsigned c2 = *++p;
54 	    if ((c2 & 0xC0) == 0x80) {
55 		const unsigned c3 = *++p;
56 		if ((c3 & 0xC0) == 0x80) {
57 		    *out =   ((c  & 0x0F) << 12)
58 			| ((c2 & 0x3F) << 6)
59 			|  (c3 & 0x3F);
60 		} else {
61 		    return WIND_ERR_INVALID_UTF8;
62 		}
63 	    } else {
64 		return WIND_ERR_INVALID_UTF8;
65 	    }
66 	} else if ((c & 0xF8) == 0xF0) {
67 	    const unsigned c2 = *++p;
68 	    if ((c2 & 0xC0) == 0x80) {
69 		const unsigned c3 = *++p;
70 		if ((c3 & 0xC0) == 0x80) {
71 		    const unsigned c4 = *++p;
72 		    if ((c4 & 0xC0) == 0x80) {
73 			*out =   ((c  & 0x07) << 18)
74 			    | ((c2 & 0x3F) << 12)
75 			    | ((c3 & 0x3F) <<  6)
76 			    |  (c4 & 0x3F);
77 		    } else {
78 			return WIND_ERR_INVALID_UTF8;
79 		    }
80 		} else {
81 		    return WIND_ERR_INVALID_UTF8;
82 		}
83 	    } else {
84 		return WIND_ERR_INVALID_UTF8;
85 	    }
86 	} else {
87 	    return WIND_ERR_INVALID_UTF8;
88 	}
89     } else {
90 	*out = c;
91     }
92 
93     *pp = p;
94 
95     return 0;
96 }
97 
98 /**
99  * Convert an UTF-8 string to an UCS4 string.
100  *
101  * @param in an UTF-8 string to convert.
102  * @param out the resulting UCS4 strint, must be at least
103  * wind_utf8ucs4_length() long.  If out is NULL, the function will
104  * calculate the needed space for the out variable (just like
105  * wind_utf8ucs4_length()).
106  * @param out_len before processing out_len should be the length of
107  * the out variable, after processing it will be the length of the out
108  * string.
109  *
110  * @return returns 0 on success, an wind error code otherwise
111  * @ingroup wind
112  */
113 
114 int
115 wind_utf8ucs4(const char *in, uint32_t *out, size_t *out_len)
116 {
117     const unsigned char *p;
118     size_t o = 0;
119     int ret;
120 
121     for (p = (const unsigned char *)in; *p != '\0'; ++p) {
122 	uint32_t u;
123 
124 	ret = utf8toutf32(&p, &u);
125 	if (ret)
126 	    return ret;
127 
128 	if (out) {
129 	    if (o >= *out_len)
130 		return WIND_ERR_OVERRUN;
131 	    out[o] = u;
132 	}
133 	o++;
134     }
135     *out_len = o;
136     return 0;
137 }
138 
139 /**
140  * Calculate the length of from converting a UTF-8 string to a UCS4
141  * string.
142  *
143  * @param in an UTF-8 string to convert.
144  * @param out_len the length of the resulting UCS4 string.
145  *
146  * @return returns 0 on success, an wind error code otherwise
147  * @ingroup wind
148  */
149 
150 int
151 wind_utf8ucs4_length(const char *in, size_t *out_len)
152 {
153     return wind_utf8ucs4(in, NULL, out_len);
154 }
155 
156 static const char first_char[4] =
157     { 0x00, 0xC0, 0xE0, 0xF0 };
158 
159 /**
160  * Convert an UCS4 string to a UTF-8 string.
161  *
162  * @param in an UCS4 string to convert.
163  * @param in_len the length input array.
164 
165  * @param out the resulting UTF-8 strint, must be at least
166  * wind_ucs4utf8_length() + 1 long (the extra char for the NUL).  If
167  * out is NULL, the function will calculate the needed space for the
168  * out variable (just like wind_ucs4utf8_length()).
169 
170  * @param out_len before processing out_len should be the length of
171  * the out variable, after processing it will be the length of the out
172  * string.
173  *
174  * @return returns 0 on success, an wind error code otherwise
175  * @ingroup wind
176  */
177 
178 int
179 wind_ucs4utf8(const uint32_t *in, size_t in_len, char *out, size_t *out_len)
180 {
181     uint32_t ch;
182     size_t i, len, o;
183 
184     for (o = 0, i = 0; i < in_len; i++) {
185 	ch = in[i];
186 
187 	if (ch < 0x80) {
188 	    len = 1;
189 	} else if (ch < 0x800) {
190 	    len = 2;
191 	} else if (ch < 0x10000) {
192 	    len = 3;
193 	} else if (ch <= 0x10FFFF) {
194 	    len = 4;
195 	} else
196 	    return WIND_ERR_INVALID_UTF32;
197 
198 	o += len;
199 
200 	if (out) {
201 	    if (o >= *out_len)
202 		return WIND_ERR_OVERRUN;
203 
204 	    switch(len) {
205 	    case 4:
206 		out[3] = (ch | 0x80) & 0xbf;
207 		ch = ch >> 6;
208 	    case 3:
209 		out[2] = (ch | 0x80) & 0xbf;
210 		ch = ch >> 6;
211 	    case 2:
212 		out[1] = (ch | 0x80) & 0xbf;
213 		ch = ch >> 6;
214 	    case 1:
215 		out[0] = ch | first_char[len - 1];
216 	    }
217 	}
218 	out += len;
219     }
220     if (out) {
221 	if (o + 1 >= *out_len)
222 	    return WIND_ERR_OVERRUN;
223 	*out = '\0';
224     }
225     *out_len = o;
226     return 0;
227 }
228 
229 /**
230  * Calculate the length of from converting a UCS4 string to an UTF-8 string.
231  *
232  * @param in an UCS4 string to convert.
233  * @param in_len the length of UCS4 string to convert.
234  * @param out_len the length of the resulting UTF-8 string.
235  *
236  * @return returns 0 on success, an wind error code otherwise
237  * @ingroup wind
238  */
239 
240 int
241 wind_ucs4utf8_length(const uint32_t *in, size_t in_len, size_t *out_len)
242 {
243     return wind_ucs4utf8(in, in_len, NULL, out_len);
244 }
245 
246 /**
247  * Read in an UCS2 from a buffer.
248  *
249  * @param ptr The input buffer to read from.
250  * @param len the length of the input buffer.
251  * @param flags Flags to control the behavior of the function.
252  * @param out the output UCS2, the array must be at least out/2 long.
253  * @param out_len the output length
254  *
255  * @return returns 0 on success, an wind error code otherwise.
256  * @ingroup wind
257  */
258 
259 int
260 wind_ucs2read(const void *ptr, size_t len, unsigned int *flags,
261 	      uint16_t *out, size_t *out_len)
262 {
263     const unsigned char *p = ptr;
264     int little = ((*flags) & WIND_RW_LE);
265     size_t olen = *out_len;
266 
267     /** if len is zero, flags are unchanged */
268     if (len == 0) {
269 	*out_len = 0;
270 	return 0;
271     }
272 
273     /** if len is odd, WIND_ERR_LENGTH_NOT_MOD2 is returned */
274     if (len & 1)
275 	return WIND_ERR_LENGTH_NOT_MOD2;
276 
277     /**
278      * If the flags WIND_RW_BOM is set, check for BOM. If not BOM is
279      * found, check is LE/BE flag is already and use that otherwise
280      * fail with WIND_ERR_NO_BOM. When done, clear WIND_RW_BOM and
281      * the LE/BE flag and set the resulting LE/BE flag.
282      */
283     if ((*flags) & WIND_RW_BOM) {
284 	uint16_t bom = (p[0] << 8) + p[1];
285 	if (bom == 0xfffe || bom == 0xfeff) {
286 	    little = (bom == 0xfffe);
287 	    p += 2;
288 	    len -= 2;
289 	} else if (((*flags) & (WIND_RW_LE|WIND_RW_BE)) != 0) {
290 	    /* little already set */
291 	} else
292 	    return WIND_ERR_NO_BOM;
293 	*flags = ((*flags) & ~(WIND_RW_BOM|WIND_RW_LE|WIND_RW_BE));
294 	*flags |= little ? WIND_RW_LE : WIND_RW_BE;
295     }
296 
297     while (len) {
298 	if (olen < 1)
299 	    return WIND_ERR_OVERRUN;
300 	if (little)
301 	    *out = (p[1] << 8) + p[0];
302 	else
303 	    *out = (p[0] << 8) + p[1];
304 	out++; p += 2; len -= 2; olen--;
305     }
306     *out_len -= olen;
307     return 0;
308 }
309 
310 /**
311  * Write an UCS2 string to a buffer.
312  *
313  * @param in The input UCS2 string.
314  * @param in_len the length of the input buffer.
315  * @param flags Flags to control the behavior of the function.
316  * @param ptr The input buffer to write to, the array must be at least
317  * (in + 1) * 2 bytes long.
318  * @param out_len the output length
319  *
320  * @return returns 0 on success, an wind error code otherwise.
321  * @ingroup wind
322  */
323 
324 int
325 wind_ucs2write(const uint16_t *in, size_t in_len, unsigned int *flags,
326 	       void *ptr, size_t *out_len)
327 {
328     unsigned char *p = ptr;
329     size_t len = *out_len;
330 
331     /** If in buffer is not of length be mod 2, WIND_ERR_LENGTH_NOT_MOD2 is returned*/
332     if (len & 1)
333 	return WIND_ERR_LENGTH_NOT_MOD2;
334 
335     /** On zero input length, flags are preserved */
336     if (in_len == 0) {
337 	*out_len = 0;
338 	return 0;
339     }
340     /** If flags have WIND_RW_BOM set, the byte order mark is written
341      * first to the output data */
342     if ((*flags) & WIND_RW_BOM) {
343 	uint16_t bom = 0xfffe;
344 
345 	if (len < 2)
346 	    return WIND_ERR_OVERRUN;
347 
348 	if ((*flags) & WIND_RW_LE) {
349 	    p[0] = (bom     ) & 0xff;
350 	    p[1] = (bom >> 8) & 0xff;
351 	} else {
352 	    p[1] = (bom     ) & 0xff;
353 	    p[0] = (bom >> 8) & 0xff;
354 	}
355 	len -= 2;
356     }
357 
358     while (in_len) {
359 	/** If the output wont fit into out_len, WIND_ERR_OVERRUN is returned */
360 	if (len < 2)
361 	    return WIND_ERR_OVERRUN;
362 	if ((*flags) & WIND_RW_LE) {
363 	    p[0] = (in[0]     ) & 0xff;
364 	    p[1] = (in[0] >> 8) & 0xff;
365 	} else {
366 	    p[1] = (in[0]     ) & 0xff;
367 	    p[0] = (in[0] >> 8) & 0xff;
368 	}
369 	len -= 2;
370 	in_len--;
371 	p += 2;
372 	in++;
373     }
374     *out_len -= len;
375     return 0;
376 }
377 
378 
379 /**
380  * Convert an UTF-8 string to an UCS2 string.
381  *
382  * @param in an UTF-8 string to convert.
383  * @param out the resulting UCS2 strint, must be at least
384  * wind_utf8ucs2_length() long.  If out is NULL, the function will
385  * calculate the needed space for the out variable (just like
386  * wind_utf8ucs2_length()).
387  * @param out_len before processing out_len should be the length of
388  * the out variable, after processing it will be the length of the out
389  * string.
390  *
391  * @return returns 0 on success, an wind error code otherwise
392  * @ingroup wind
393  */
394 
395 int
396 wind_utf8ucs2(const char *in, uint16_t *out, size_t *out_len)
397 {
398     const unsigned char *p;
399     size_t o = 0;
400     int ret;
401 
402     for (p = (const unsigned char *)in; *p != '\0'; ++p) {
403 	uint32_t u;
404 
405 	ret = utf8toutf32(&p, &u);
406 	if (ret)
407 	    return ret;
408 
409 	if (u & 0xffff0000)
410 	    return WIND_ERR_NOT_UTF16;
411 
412 	if (out) {
413 	    if (o >= *out_len)
414 		return WIND_ERR_OVERRUN;
415 	    out[o] = u;
416 	}
417 	o++;
418     }
419     *out_len = o;
420     return 0;
421 }
422 
423 /**
424  * Calculate the length of from converting a UTF-8 string to a UCS2
425  * string.
426  *
427  * @param in an UTF-8 string to convert.
428  * @param out_len the length of the resulting UCS4 string.
429  *
430  * @return returns 0 on success, an wind error code otherwise
431  * @ingroup wind
432  */
433 
434 int
435 wind_utf8ucs2_length(const char *in, size_t *out_len)
436 {
437     return wind_utf8ucs2(in, NULL, out_len);
438 }
439 
440 /**
441  * Convert an UCS2 string to a UTF-8 string.
442  *
443  * @param in an UCS2 string to convert.
444  * @param in_len the length of the in UCS2 string.
445  * @param out the resulting UTF-8 strint, must be at least
446  * wind_ucs2utf8_length() long.  If out is NULL, the function will
447  * calculate the needed space for the out variable (just like
448  * wind_ucs2utf8_length()).
449  * @param out_len before processing out_len should be the length of
450  * the out variable, after processing it will be the length of the out
451  * string.
452  *
453  * @return returns 0 on success, an wind error code otherwise
454  * @ingroup wind
455  */
456 
457 int
458 wind_ucs2utf8(const uint16_t *in, size_t in_len, char *out, size_t *out_len)
459 {
460     uint16_t ch;
461     size_t i, len, o;
462 
463     for (o = 0, i = 0; i < in_len; i++) {
464 	ch = in[i];
465 
466 	if (ch < 0x80) {
467 	    len = 1;
468 	} else if (ch < 0x800) {
469 	    len = 2;
470 	} else
471 	    len = 3;
472 
473 	o += len;
474 
475 	if (out) {
476 	    if (o >= *out_len)
477 		return WIND_ERR_OVERRUN;
478 
479 	    switch(len) {
480 	    case 3:
481 		out[2] = (ch | 0x80) & 0xbf;
482 		ch = ch >> 6;
483 	    case 2:
484 		out[1] = (ch | 0x80) & 0xbf;
485 		ch = ch >> 6;
486 	    case 1:
487 		out[0] = ch | first_char[len - 1];
488 	    }
489 	    out += len;
490 	}
491     }
492     if (out) {
493 	if (o >= *out_len)
494 	    return WIND_ERR_OVERRUN;
495 	*out = '\0';
496     }
497     *out_len = o;
498     return 0;
499 }
500 
501 /**
502  * Calculate the length of from converting a UCS2 string to an UTF-8 string.
503  *
504  * @param in an UCS2 string to convert.
505  * @param in_len an UCS2 string length to convert.
506  * @param out_len the length of the resulting UTF-8 string.
507  *
508  * @return returns 0 on success, an wind error code otherwise
509  * @ingroup wind
510  */
511 
512 int
513 wind_ucs2utf8_length(const uint16_t *in, size_t in_len, size_t *out_len)
514 {
515     return wind_ucs2utf8(in, in_len, NULL, out_len);
516 }
517