VirtualBox

source: vbox/trunk/src/VBox/Runtime/common/string/utf-8.cpp@ 34079

Last change on this file since 34079 was 33540, checked in by vboxsync, 14 years ago

*: spelling fixes, thanks Timeless!
  • Property svn:eol-style set to native
  • Property svn:keywords set to Id
File size: 53.7 KB
Line 
1/* $Id: utf-8.cpp 33540 2010-10-28 09:27:05Z vboxsync $ */
2/** @file
3 * IPRT - UTF-8 Decoding.
4 */
5
6/*
7 * Copyright (C) 2006-2010 Oracle Corporation
8 *
9 * This file is part of VirtualBox Open Source Edition (OSE), as
10 * available from http://www.virtualbox.org. This file is free software;
11 * you can redistribute it and/or modify it under the terms of the GNU
12 * General Public License (GPL) as published by the Free Software
13 * Foundation, in version 2 as it comes in the "COPYING" file of the
14 * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15 * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16 *
17 * The contents of this file may alternatively be used under the terms
18 * of the Common Development and Distribution License Version 1.0
19 * (CDDL) only, as it comes in the "COPYING.CDDL" file of the
20 * VirtualBox OSE distribution, in which case the provisions of the
21 * CDDL are applicable instead of those of the GPL.
22 *
23 * You may elect to license modified versions of this file under the
24 * terms and conditions of either the GPL or the CDDL or both.
25 */
26
27
28/*******************************************************************************
29* Header Files *
30*******************************************************************************/
31#include <iprt/string.h>
32#include "internal/iprt.h"
33
34#include <iprt/uni.h>
35#include <iprt/alloc.h>
36#include <iprt/assert.h>
37#include <iprt/err.h>
38#include "internal/string.h"
39
40
41
42/**
43 * Get get length in code points of a UTF-8 encoded string.
44 * The string is validated while doing this.
45 *
46 * @returns IPRT status code.
47 * @param psz Pointer to the UTF-8 string.
48 * @param cch The max length of the string. (btw cch = cb)
49 * Use RTSTR_MAX if all of the string is to be examined.
50 * @param pcuc Where to store the length in unicode code points.
51 * @param pcchActual Where to store the actual size of the UTF-8 string
52 * on success (cch = cb again). Optional.
53 */
54int rtUtf8Length(const char *psz, size_t cch, size_t *pcuc, size_t *pcchActual)
55{
56 const unsigned char *puch = (const unsigned char *)psz;
57 size_t cCodePoints = 0;
58 while (cch > 0)
59 {
60 const unsigned char uch = *puch;
61 if (!uch)
62 break;
63 if (uch & RT_BIT(7))
64 {
65 /* figure sequence length and validate the first byte */
66/** @todo RT_USE_RTC_3629 */
67 unsigned cb;
68 if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5))) == (RT_BIT(7) | RT_BIT(6)))
69 cb = 2;
70 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5)))
71 cb = 3;
72 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4)))
73 cb = 4;
74 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3)))
75 cb = 5;
76 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2) | RT_BIT(1))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2)))
77 cb = 6;
78 else
79 {
80 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(cch, 10), puch));
81 return VERR_INVALID_UTF8_ENCODING;
82 }
83
84 /* check length */
85 if (cb > cch)
86 {
87 RTStrAssertMsgFailed(("Invalid UTF-8 length: cb=%d cch=%d (%.*Rhxs)\n", cb, cch, RT_MIN(cch, 10), puch));
88 return VERR_INVALID_UTF8_ENCODING;
89 }
90
91 /* validate the rest */
92 switch (cb)
93 {
94 case 6:
95 RTStrAssertMsgReturn((puch[5] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("6/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
96 case 5:
97 RTStrAssertMsgReturn((puch[4] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("5/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
98 case 4:
99 RTStrAssertMsgReturn((puch[3] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("4/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
100 case 3:
101 RTStrAssertMsgReturn((puch[2] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("3/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
102 case 2:
103 RTStrAssertMsgReturn((puch[1] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("2/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
104 break;
105 }
106
107 /* validate the code point. */
108 RTUNICP uc;
109 switch (cb)
110 {
111 case 6:
112 uc = (puch[5] & 0x3f)
113 | ((RTUNICP)(puch[4] & 0x3f) << 6)
114 | ((RTUNICP)(puch[3] & 0x3f) << 12)
115 | ((RTUNICP)(puch[2] & 0x3f) << 18)
116 | ((RTUNICP)(puch[1] & 0x3f) << 24)
117 | ((RTUNICP)(uch & 0x01) << 30);
118 RTStrAssertMsgReturn(uc >= 0x04000000 && uc <= 0x7fffffff,
119 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
120 break;
121 case 5:
122 uc = (puch[4] & 0x3f)
123 | ((RTUNICP)(puch[3] & 0x3f) << 6)
124 | ((RTUNICP)(puch[2] & 0x3f) << 12)
125 | ((RTUNICP)(puch[1] & 0x3f) << 18)
126 | ((RTUNICP)(uch & 0x03) << 24);
127 RTStrAssertMsgReturn(uc >= 0x00200000 && uc <= 0x03ffffff,
128 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
129 break;
130 case 4:
131 uc = (puch[3] & 0x3f)
132 | ((RTUNICP)(puch[2] & 0x3f) << 6)
133 | ((RTUNICP)(puch[1] & 0x3f) << 12)
134 | ((RTUNICP)(uch & 0x07) << 18);
135 RTStrAssertMsgReturn(uc >= 0x00010000 && uc <= 0x001fffff,
136 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
137 break;
138 case 3:
139 uc = (puch[2] & 0x3f)
140 | ((RTUNICP)(puch[1] & 0x3f) << 6)
141 | ((RTUNICP)(uch & 0x0f) << 12);
142 RTStrAssertMsgReturn(uc >= 0x00000800 && uc <= 0x0000fffd,
143 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch),
144 uc == 0xffff || uc == 0xfffe ? VERR_CODE_POINT_ENDIAN_INDICATOR : VERR_INVALID_UTF8_ENCODING);
145 RTStrAssertMsgReturn(uc < 0xd800 || uc > 0xdfff,
146 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_CODE_POINT_SURROGATE);
147 break;
148 case 2:
149 uc = (puch[1] & 0x3f)
150 | ((RTUNICP)(uch & 0x1f) << 6);
151 RTStrAssertMsgReturn(uc >= 0x00000080 && uc <= 0x000007ff,
152 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
153 break;
154 }
155
156 /* advance */
157 cch -= cb;
158 puch += cb;
159 }
160 else
161 {
162 /* one ASCII byte */
163 puch++;
164 cch--;
165 }
166 cCodePoints++;
167 }
168
169 /* done */
170 *pcuc = cCodePoints;
171 if (pcchActual)
172 *pcchActual = puch - (unsigned char const *)psz;
173 return VINF_SUCCESS;
174}
175
176
177/**
178 * Decodes and UTF-8 string into an array of unicode code point.
179 *
180 * Since we know the input is valid, we do *not* perform encoding or length checks.
181 *
182 * @returns iprt status code.
183 * @param psz The UTF-8 string to recode. This is a valid encoding.
184 * @param cch The number of chars (the type char, so bytes if you like) to process of the UTF-8 string.
185 * The recoding will stop when cch or '\\0' is reached. Pass RTSTR_MAX to process up to '\\0'.
186 * @param paCps Where to store the code points array.
187 * @param cCps The number of RTUNICP items the paCps buffer can hold, excluding the terminator ('\\0').
188 */
189static int rtUtf8Decode(const char *psz, size_t cch, PRTUNICP paCps, size_t cCps)
190{
191 int rc = VINF_SUCCESS;
192 const unsigned char *puch = (const unsigned char *)psz;
193 PRTUNICP pCp = paCps;
194 while (cch > 0)
195 {
196 /* read the next char and check for terminator. */
197 const unsigned char uch = *puch;
198 if (!uch)
199 break;
200
201 /* check for output overflow */
202 if (RT_UNLIKELY(cCps < 1))
203 {
204 rc = VERR_BUFFER_OVERFLOW;
205 break;
206 }
207 cCps--;
208
209 /* decode and recode the code point */
210 if (!(uch & RT_BIT(7)))
211 {
212 *pCp++ = uch;
213 puch++;
214 cch--;
215 }
216#ifdef RT_STRICT
217 else if (!(uch & RT_BIT(6)))
218 AssertMsgFailed(("Internal error!\n"));
219#endif
220 else if (!(uch & RT_BIT(5)))
221 {
222 *pCp++ = (puch[1] & 0x3f)
223 | ((uint16_t)(uch & 0x1f) << 6);
224 puch += 2;
225 cch -= 2;
226 }
227 else if (!(uch & RT_BIT(4)))
228 {
229 *pCp++ = (puch[2] & 0x3f)
230 | ((uint16_t)(puch[1] & 0x3f) << 6)
231 | ((uint16_t)(uch & 0x0f) << 12);
232 puch += 3;
233 cch -= 3;
234 }
235 else if (!(uch & RT_BIT(3)))
236 {
237 *pCp++ = (puch[3] & 0x3f)
238 | ((RTUNICP)(puch[2] & 0x3f) << 6)
239 | ((RTUNICP)(puch[1] & 0x3f) << 12)
240 | ((RTUNICP)(uch & 0x07) << 18);
241 puch += 4;
242 cch -= 4;
243 }
244 else if (!(uch & RT_BIT(2)))
245 {
246 *pCp++ = (puch[4] & 0x3f)
247 | ((RTUNICP)(puch[3] & 0x3f) << 6)
248 | ((RTUNICP)(puch[2] & 0x3f) << 12)
249 | ((RTUNICP)(puch[1] & 0x3f) << 18)
250 | ((RTUNICP)(uch & 0x03) << 24);
251 puch += 5;
252 cch -= 6;
253 }
254 else
255 {
256 Assert(!(uch & RT_BIT(1)));
257 *pCp++ = (puch[5] & 0x3f)
258 | ((RTUNICP)(puch[4] & 0x3f) << 6)
259 | ((RTUNICP)(puch[3] & 0x3f) << 12)
260 | ((RTUNICP)(puch[2] & 0x3f) << 18)
261 | ((RTUNICP)(puch[1] & 0x3f) << 24)
262 | ((RTUNICP)(uch & 0x01) << 30);
263 puch += 6;
264 cch -= 6;
265 }
266 }
267
268 /* done */
269 *pCp = 0;
270 return rc;
271}
272
273
274RTDECL(size_t) RTStrUniLen(const char *psz)
275{
276 size_t cCodePoints;
277 int rc = rtUtf8Length(psz, RTSTR_MAX, &cCodePoints, NULL);
278 return RT_SUCCESS(rc) ? cCodePoints : 0;
279}
280RT_EXPORT_SYMBOL(RTStrUniLen);
281
282
283RTDECL(int) RTStrUniLenEx(const char *psz, size_t cch, size_t *pcCps)
284{
285 size_t cCodePoints;
286 int rc = rtUtf8Length(psz, cch, &cCodePoints, NULL);
287 if (pcCps)
288 *pcCps = RT_SUCCESS(rc) ? cCodePoints : 0;
289 return rc;
290}
291RT_EXPORT_SYMBOL(RTStrUniLenEx);
292
293
294RTDECL(int) RTStrValidateEncoding(const char *psz)
295{
296 return RTStrValidateEncodingEx(psz, RTSTR_MAX, 0);
297}
298RT_EXPORT_SYMBOL(RTStrValidateEncoding);
299
300
301RTDECL(int) RTStrValidateEncodingEx(const char *psz, size_t cch, uint32_t fFlags)
302{
303 AssertReturn(!(fFlags & ~(RTSTR_VALIDATE_ENCODING_ZERO_TERMINATED)), VERR_INVALID_PARAMETER);
304 AssertPtr(psz);
305
306 /*
307 * Use rtUtf8Length for the job.
308 */
309 size_t cchActual;
310 size_t cCpsIgnored;
311 int rc = rtUtf8Length(psz, cch, &cCpsIgnored, &cchActual);
312 if (RT_SUCCESS(rc))
313 {
314 if ( (fFlags & RTSTR_VALIDATE_ENCODING_ZERO_TERMINATED)
315 && cchActual >= cch)
316 rc = VERR_BUFFER_OVERFLOW;
317 }
318 return rc;
319}
320RT_EXPORT_SYMBOL(RTStrValidateEncodingEx);
321
322
323RTDECL(bool) RTStrIsValidEncoding(const char *psz)
324{
325 int rc = RTStrValidateEncodingEx(psz, RTSTR_MAX, 0);
326 return RT_SUCCESS(rc);
327}
328RT_EXPORT_SYMBOL(RTStrIsValidEncoding);
329
330
331RTDECL(size_t) RTStrPurgeEncoding(char *psz)
332{
333 size_t cErrors = 0;
334 for (;;)
335 {
336 RTUNICP Cp;
337 int rc = RTStrGetCpEx((const char **)&psz, &Cp);
338 if (RT_SUCCESS(rc))
339 {
340 if (!Cp)
341 break;
342 }
343 else
344 {
345 psz[-1] = '?';
346 cErrors++;
347 }
348 }
349 return cErrors;
350}
351RT_EXPORT_SYMBOL(RTStrPurgeEncoding);
352
353
354RTDECL(int) RTStrToUni(const char *pszString, PRTUNICP *ppaCps)
355{
356 /*
357 * Validate input.
358 */
359 Assert(VALID_PTR(pszString));
360 Assert(VALID_PTR(ppaCps));
361 *ppaCps = NULL;
362
363 /*
364 * Validate the UTF-8 input and count its code points.
365 */
366 size_t cCps;
367 int rc = rtUtf8Length(pszString, RTSTR_MAX, &cCps, NULL);
368 if (RT_SUCCESS(rc))
369 {
370 /*
371 * Allocate buffer.
372 */
373 PRTUNICP paCps = (PRTUNICP)RTMemAlloc((cCps + 1) * sizeof(RTUNICP));
374 if (paCps)
375 {
376 /*
377 * Decode the string.
378 */
379 rc = rtUtf8Decode(pszString, RTSTR_MAX, paCps, cCps);
380 if (RT_SUCCESS(rc))
381 {
382 *ppaCps = paCps;
383 return rc;
384 }
385 RTMemFree(paCps);
386 }
387 else
388 rc = VERR_NO_CODE_POINT_MEMORY;
389 }
390 return rc;
391}
392RT_EXPORT_SYMBOL(RTStrToUni);
393
394
395RTDECL(int) RTStrToUniEx(const char *pszString, size_t cchString, PRTUNICP *ppaCps, size_t cCps, size_t *pcCps)
396{
397 /*
398 * Validate input.
399 */
400 Assert(VALID_PTR(pszString));
401 Assert(VALID_PTR(ppaCps));
402 Assert(!pcCps || VALID_PTR(pcCps));
403
404 /*
405 * Validate the UTF-8 input and count the code points.
406 */
407 size_t cCpsResult;
408 int rc = rtUtf8Length(pszString, cchString, &cCpsResult, NULL);
409 if (RT_SUCCESS(rc))
410 {
411 if (pcCps)
412 *pcCps = cCpsResult;
413
414 /*
415 * Check buffer size / Allocate buffer.
416 */
417 bool fShouldFree;
418 PRTUNICP paCpsResult;
419 if (cCps > 0 && *ppaCps)
420 {
421 fShouldFree = false;
422 if (cCps <= cCpsResult)
423 return VERR_BUFFER_OVERFLOW;
424 paCpsResult = *ppaCps;
425 }
426 else
427 {
428 *ppaCps = NULL;
429 fShouldFree = true;
430 cCps = RT_MAX(cCpsResult + 1, cCps);
431 paCpsResult = (PRTUNICP)RTMemAlloc(cCps * sizeof(RTUNICP));
432 }
433 if (paCpsResult)
434 {
435 /*
436 * Encode the UTF-16 string.
437 */
438 rc = rtUtf8Decode(pszString, cchString, paCpsResult, cCps - 1);
439 if (RT_SUCCESS(rc))
440 {
441 *ppaCps = paCpsResult;
442 return rc;
443 }
444 if (fShouldFree)
445 RTMemFree(paCpsResult);
446 }
447 else
448 rc = VERR_NO_CODE_POINT_MEMORY;
449 }
450 return rc;
451}
452RT_EXPORT_SYMBOL(RTStrToUniEx);
453
454
455/**
456 * Calculates the UTF-16 length of a string, validating the encoding while doing so.
457 *
458 * @returns IPRT status code.
459 * @param psz Pointer to the UTF-8 string.
460 * @param cch The max length of the string. (btw cch = cb)
461 * Use RTSTR_MAX if all of the string is to be examined.
462 * @param pcwc Where to store the length of the UTF-16 string as a number of RTUTF16 characters.
463 */
464static int rtUtf8CalcUtf16Length(const char *psz, size_t cch, size_t *pcwc)
465{
466 const unsigned char *puch = (const unsigned char *)psz;
467 size_t cwc = 0;
468 while (cch > 0)
469 {
470 const unsigned char uch = *puch;
471 if (!uch)
472 break;
473 if (!(uch & RT_BIT(7)))
474 {
475 /* one ASCII byte */
476 cwc++;
477 puch++;
478 cch--;
479 }
480 else
481 {
482 /* figure sequence length and validate the first byte */
483 unsigned cb;
484 if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5))) == (RT_BIT(7) | RT_BIT(6)))
485 cb = 2;
486 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5)))
487 cb = 3;
488 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4)))
489 cb = 4;
490 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3)))
491 cb = 5;
492 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2) | RT_BIT(1))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3) | RT_BIT(2)))
493 cb = 6;
494 else
495 {
496 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(cch, 10), puch));
497 return VERR_INVALID_UTF8_ENCODING;
498 }
499
500 /* check length */
501 if (cb > cch)
502 {
503 RTStrAssertMsgFailed(("Invalid UTF-8 length: cb=%d cch=%d (%.*Rhxs)\n", cb, cch, RT_MIN(cch, 10), puch));
504 return VERR_INVALID_UTF8_ENCODING;
505 }
506
507 /* validate the rest */
508 switch (cb)
509 {
510 case 6:
511 RTStrAssertMsgReturn((puch[5] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("6/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
512 case 5:
513 RTStrAssertMsgReturn((puch[4] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("5/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
514 case 4:
515 RTStrAssertMsgReturn((puch[3] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("4/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
516 case 3:
517 RTStrAssertMsgReturn((puch[2] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("3/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
518 case 2:
519 RTStrAssertMsgReturn((puch[1] & (RT_BIT(7) | RT_BIT(6))) == RT_BIT(7), ("2/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
520 break;
521 }
522
523 /* validate the code point. */
524 RTUNICP uc;
525 switch (cb)
526 {
527 case 6:
528 uc = (puch[5] & 0x3f)
529 | ((RTUNICP)(puch[4] & 0x3f) << 6)
530 | ((RTUNICP)(puch[3] & 0x3f) << 12)
531 | ((RTUNICP)(puch[2] & 0x3f) << 18)
532 | ((RTUNICP)(puch[1] & 0x3f) << 24)
533 | ((RTUNICP)(uch & 0x01) << 30);
534 RTStrAssertMsgReturn(uc >= 0x04000000 && uc <= 0x7fffffff,
535 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
536 RTStrAssertMsgFailed(("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch));
537 return VERR_CANT_RECODE_AS_UTF16;
538 case 5:
539 uc = (puch[4] & 0x3f)
540 | ((RTUNICP)(puch[3] & 0x3f) << 6)
541 | ((RTUNICP)(puch[2] & 0x3f) << 12)
542 | ((RTUNICP)(puch[1] & 0x3f) << 18)
543 | ((RTUNICP)(uch & 0x03) << 24);
544 RTStrAssertMsgReturn(uc >= 0x00200000 && uc <= 0x03ffffff,
545 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
546 RTStrAssertMsgFailed(("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch));
547 return VERR_CANT_RECODE_AS_UTF16;
548 case 4:
549 uc = (puch[3] & 0x3f)
550 | ((RTUNICP)(puch[2] & 0x3f) << 6)
551 | ((RTUNICP)(puch[1] & 0x3f) << 12)
552 | ((RTUNICP)(uch & 0x07) << 18);
553 RTStrAssertMsgReturn(uc >= 0x00010000 && uc <= 0x001fffff,
554 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
555 RTStrAssertMsgReturn(uc <= 0x0010ffff,
556 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_CANT_RECODE_AS_UTF16);
557 cwc++;
558 break;
559 case 3:
560 uc = (puch[2] & 0x3f)
561 | ((RTUNICP)(puch[1] & 0x3f) << 6)
562 | ((RTUNICP)(uch & 0x0f) << 12);
563 RTStrAssertMsgReturn(uc >= 0x00000800 && uc <= 0x0000fffd,
564 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch),
565 uc == 0xffff || uc == 0xfffe ? VERR_CODE_POINT_ENDIAN_INDICATOR : VERR_INVALID_UTF8_ENCODING);
566 RTStrAssertMsgReturn(uc < 0xd800 || uc > 0xdfff,
567 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_CODE_POINT_SURROGATE);
568 break;
569 case 2:
570 uc = (puch[1] & 0x3f)
571 | ((RTUNICP)(uch & 0x1f) << 6);
572 RTStrAssertMsgReturn(uc >= 0x00000080 && uc <= 0x000007ff,
573 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, cch), puch), VERR_INVALID_UTF8_ENCODING);
574 break;
575 }
576
577 /* advance */
578 cch -= cb;
579 puch += cb;
580 cwc++;
581 }
582 }
583
584 /* done */
585 *pcwc = cwc;
586 return VINF_SUCCESS;
587}
588
589
590/**
591 * Recodes a valid UTF-8 string as UTF-16.
592 *
593 * Since we know the input is valid, we do *not* perform encoding or length checks.
594 *
595 * @returns iprt status code.
596 * @param psz The UTF-8 string to recode. This is a valid encoding.
597 * @param cch The number of chars (the type char, so bytes if you like) to process of the UTF-8 string.
598 * The recoding will stop when cch or '\\0' is reached. Pass RTSTR_MAX to process up to '\\0'.
599 * @param pwsz Where to store the UTF-16 string.
600 * @param cwc The number of RTUTF16 items the pwsz buffer can hold, excluding the terminator ('\\0').
601 */
602static int rtUtf8RecodeAsUtf16(const char *psz, size_t cch, PRTUTF16 pwsz, size_t cwc)
603{
604 int rc = VINF_SUCCESS;
605 const unsigned char *puch = (const unsigned char *)psz;
606 PRTUTF16 pwc = pwsz;
607 while (cch > 0)
608 {
609 /* read the next char and check for terminator. */
610 const unsigned char uch = *puch;
611 if (!uch)
612 break;
613
614 /* check for output overflow */
615 if (RT_UNLIKELY(cwc < 1))
616 {
617 rc = VERR_BUFFER_OVERFLOW;
618 break;
619 }
620 cwc--;
621
622 /* decode and recode the code point */
623 if (!(uch & RT_BIT(7)))
624 {
625 *pwc++ = uch;
626 puch++;
627 cch--;
628 }
629 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5))) == (RT_BIT(7) | RT_BIT(6)))
630 {
631 uint16_t uc = (puch[1] & 0x3f)
632 | ((uint16_t)(uch & 0x1f) << 6);
633 *pwc++ = uc;
634 puch += 2;
635 cch -= 2;
636 }
637 else if ((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5)))
638 {
639 uint16_t uc = (puch[2] & 0x3f)
640 | ((uint16_t)(puch[1] & 0x3f) << 6)
641 | ((uint16_t)(uch & 0x0f) << 12);
642 *pwc++ = uc;
643 puch += 3;
644 cch -= 3;
645 }
646 else
647 {
648 /* generate surrogate pair */
649 Assert((uch & (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4) | RT_BIT(3))) == (RT_BIT(7) | RT_BIT(6) | RT_BIT(5) | RT_BIT(4)));
650 RTUNICP uc = (puch[3] & 0x3f)
651 | ((RTUNICP)(puch[2] & 0x3f) << 6)
652 | ((RTUNICP)(puch[1] & 0x3f) << 12)
653 | ((RTUNICP)(uch & 0x07) << 18);
654 if (RT_UNLIKELY(cwc < 1))
655 {
656 rc = VERR_BUFFER_OVERFLOW;
657 break;
658 }
659 cwc--;
660
661 uc -= 0x10000;
662 *pwc++ = 0xd800 | (uc >> 10);
663 *pwc++ = 0xdc00 | (uc & 0x3ff);
664 puch += 4;
665 cch -= 4;
666 }
667 }
668
669 /* done */
670 *pwc = '\0';
671 return rc;
672}
673
674
675RTDECL(int) RTStrToUtf16Tag(const char *pszString, PRTUTF16 *ppwszString, const char *pszTag)
676{
677 /*
678 * Validate input.
679 */
680 Assert(VALID_PTR(ppwszString));
681 Assert(VALID_PTR(pszString));
682 *ppwszString = NULL;
683
684 /*
685 * Validate the UTF-8 input and calculate the length of the UTF-16 string.
686 */
687 size_t cwc;
688 int rc = rtUtf8CalcUtf16Length(pszString, RTSTR_MAX, &cwc);
689 if (RT_SUCCESS(rc))
690 {
691 /*
692 * Allocate buffer.
693 */
694 PRTUTF16 pwsz = (PRTUTF16)RTMemAllocTag((cwc + 1) * sizeof(RTUTF16), pszTag);
695 if (pwsz)
696 {
697 /*
698 * Encode the UTF-16 string.
699 */
700 rc = rtUtf8RecodeAsUtf16(pszString, RTSTR_MAX, pwsz, cwc);
701 if (RT_SUCCESS(rc))
702 {
703 *ppwszString = pwsz;
704 return rc;
705 }
706 RTMemFree(pwsz);
707 }
708 else
709 rc = VERR_NO_UTF16_MEMORY;
710 }
711 return rc;
712}
713RT_EXPORT_SYMBOL(RTStrToUtf16Tag);
714
715
716RTDECL(int) RTStrToUtf16ExTag(const char *pszString, size_t cchString,
717 PRTUTF16 *ppwsz, size_t cwc, size_t *pcwc, const char *pszTag)
718{
719 /*
720 * Validate input.
721 */
722 Assert(VALID_PTR(pszString));
723 Assert(VALID_PTR(ppwsz));
724 Assert(!pcwc || VALID_PTR(pcwc));
725
726 /*
727 * Validate the UTF-8 input and calculate the length of the UTF-16 string.
728 */
729 size_t cwcResult;
730 int rc = rtUtf8CalcUtf16Length(pszString, cchString, &cwcResult);
731 if (RT_SUCCESS(rc))
732 {
733 if (pcwc)
734 *pcwc = cwcResult;
735
736 /*
737 * Check buffer size / Allocate buffer.
738 */
739 bool fShouldFree;
740 PRTUTF16 pwszResult;
741 if (cwc > 0 && *ppwsz)
742 {
743 fShouldFree = false;
744 if (cwc <= cwcResult)
745 return VERR_BUFFER_OVERFLOW;
746 pwszResult = *ppwsz;
747 }
748 else
749 {
750 *ppwsz = NULL;
751 fShouldFree = true;
752 cwc = RT_MAX(cwcResult + 1, cwc);
753 pwszResult = (PRTUTF16)RTMemAllocTag(cwc * sizeof(RTUTF16), pszTag);
754 }
755 if (pwszResult)
756 {
757 /*
758 * Encode the UTF-16 string.
759 */
760 rc = rtUtf8RecodeAsUtf16(pszString, cchString, pwszResult, cwc - 1);
761 if (RT_SUCCESS(rc))
762 {
763 *ppwsz = pwszResult;
764 return rc;
765 }
766 if (fShouldFree)
767 RTMemFree(pwszResult);
768 }
769 else
770 rc = VERR_NO_UTF16_MEMORY;
771 }
772 return rc;
773}
774RT_EXPORT_SYMBOL(RTStrToUtf16ExTag);
775
776
777RTDECL(size_t) RTStrCalcUtf16Len(const char *psz)
778{
779 size_t cwc;
780 int rc = rtUtf8CalcUtf16Length(psz, RTSTR_MAX, &cwc);
781 return RT_SUCCESS(rc) ? cwc : 0;
782}
783RT_EXPORT_SYMBOL(RTStrCalcUtf16Len);
784
785
786RTDECL(int) RTStrCalcUtf16LenEx(const char *psz, size_t cch, size_t *pcwc)
787{
788 size_t cwc;
789 int rc = rtUtf8CalcUtf16Length(psz, cch, &cwc);
790 if (pcwc)
791 *pcwc = RT_SUCCESS(rc) ? cwc : ~(size_t)0;
792 return rc;
793}
794RT_EXPORT_SYMBOL(RTStrCalcUtf16LenEx);
795
796
797/**
798 * Calculates the length of the UTF-8 encoding of a Latin-1 string.
799 *
800 * @returns iprt status code.
801 * @param psz The Latin-1 string.
802 * @param cchIn The max length of the Latin-1 string to consider.
803 * @param pcch Where to store the length (excluding '\\0') of the UTF-8 string. (cch == cb, btw)
804 */
805static int rtLatin1CalcUtf8Length(const char *psz, size_t cchIn, size_t *pcch)
806{
807 size_t cch = 0;
808 for (;;)
809 {
810 RTUNICP Cp;
811 int rc = RTLatin1GetCpNEx(&psz, &cchIn, &Cp);
812 if (Cp == 0 || rc == VERR_END_OF_STRING)
813 break;
814 if (RT_FAILURE(rc))
815 return rc;
816 cch += RTStrCpSize(Cp); /* cannot fail */
817 }
818
819 /* done */
820 *pcch = cch;
821 return VINF_SUCCESS;
822}
823
824
825/**
826 * Recodes a Latin-1 string as UTF-8.
827 *
828 * @returns iprt status code.
829 * @param psz The Latin-1 string.
830 * @param cchIn The number of characters to process from psz. The recoding
831 * will stop when cch or '\\0' is reached.
832 * @param psz Where to store the UTF-8 string.
833 * @param cch The size of the UTF-8 buffer, excluding the terminator.
834 */
835static int rtLatin1RecodeAsUtf8(const char *pszIn, size_t cchIn, char *psz, size_t cch)
836{
837 int rc;
838 for (;;)
839 {
840 RTUNICP Cp;
841 size_t cchCp;
842 rc = RTLatin1GetCpNEx(&pszIn, &cchIn, &Cp);
843 if (Cp == 0 || RT_FAILURE(rc))
844 break;
845 cchCp = RTStrCpSize(Cp);
846 if (RT_UNLIKELY(cch < cchCp))
847 {
848 RTStrAssertMsgFailed(("Buffer overflow! 1\n"));
849 rc = VERR_BUFFER_OVERFLOW;
850 break;
851 }
852 cch -= cchCp;
853 psz = RTStrPutCp(psz, Cp);
854 }
855
856 /* done */
857 if (rc == VERR_END_OF_STRING)
858 rc = VINF_SUCCESS;
859 *psz = '\0';
860 return rc;
861}
862
863
864
865RTDECL(int) RTLatin1ToUtf8Tag(const char *pszString, char **ppszString, const char *pszTag)
866{
867 /*
868 * Validate input.
869 */
870 Assert(VALID_PTR(ppszString));
871 Assert(VALID_PTR(pszString));
872 *ppszString = NULL;
873
874 /*
875 * Calculate the length of the UTF-8 encoding of the Latin-1 string.
876 */
877 size_t cch;
878 int rc = rtLatin1CalcUtf8Length(pszString, RTSTR_MAX, &cch);
879 if (RT_SUCCESS(rc))
880 {
881 /*
882 * Allocate buffer and recode it.
883 */
884 char *pszResult = (char *)RTMemAllocTag(cch + 1, pszTag);
885 if (pszResult)
886 {
887 rc = rtLatin1RecodeAsUtf8(pszString, RTSTR_MAX, pszResult, cch);
888 if (RT_SUCCESS(rc))
889 {
890 *ppszString = pszResult;
891 return rc;
892 }
893
894 RTMemFree(pszResult);
895 }
896 else
897 rc = VERR_NO_STR_MEMORY;
898 }
899 return rc;
900}
901RT_EXPORT_SYMBOL(RTLatin1ToUtf8Tag);
902
903
904RTDECL(int) RTLatin1ToUtf8ExTag(const char *pszString, size_t cchString, char **ppsz, size_t cch, size_t *pcch, const char *pszTag)
905{
906 /*
907 * Validate input.
908 */
909 Assert(VALID_PTR(pszString));
910 Assert(VALID_PTR(ppsz));
911 Assert(!pcch || VALID_PTR(pcch));
912
913 /*
914 * Calculate the length of the UTF-8 encoding of the Latin-1 string.
915 */
916 size_t cchResult;
917 int rc = rtLatin1CalcUtf8Length(pszString, cchString, &cchResult);
918 if (RT_SUCCESS(rc))
919 {
920 if (pcch)
921 *pcch = cchResult;
922
923 /*
924 * Check buffer size / Allocate buffer and recode it.
925 */
926 bool fShouldFree;
927 char *pszResult;
928 if (cch > 0 && *ppsz)
929 {
930 fShouldFree = false;
931 if (RT_UNLIKELY(cch <= cchResult))
932 return VERR_BUFFER_OVERFLOW;
933 pszResult = *ppsz;
934 }
935 else
936 {
937 *ppsz = NULL;
938 fShouldFree = true;
939 cch = RT_MAX(cch, cchResult + 1);
940 pszResult = (char *)RTStrAllocTag(cch, pszTag);
941 }
942 if (pszResult)
943 {
944 rc = rtLatin1RecodeAsUtf8(pszString, cchString, pszResult, cch - 1);
945 if (RT_SUCCESS(rc))
946 {
947 *ppsz = pszResult;
948 return rc;
949 }
950
951 if (fShouldFree)
952 RTStrFree(pszResult);
953 }
954 else
955 rc = VERR_NO_STR_MEMORY;
956 }
957 return rc;
958}
959RT_EXPORT_SYMBOL(RTLatin1ToUtf8ExTag);
960
961
962RTDECL(size_t) RTLatin1CalcUtf8Len(const char *psz)
963{
964 size_t cch;
965 int rc = rtLatin1CalcUtf8Length(psz, RTSTR_MAX, &cch);
966 return RT_SUCCESS(rc) ? cch : 0;
967}
968RT_EXPORT_SYMBOL(RTLatin1CalcUtf8Len);
969
970
971RTDECL(int) RTLatin1CalcUtf8LenEx(const char *psz, size_t cchIn, size_t *pcch)
972{
973 size_t cch;
974 int rc = rtLatin1CalcUtf8Length(psz, cchIn, &cch);
975 if (pcch)
976 *pcch = RT_SUCCESS(rc) ? cch : ~(size_t)0;
977 return rc;
978}
979RT_EXPORT_SYMBOL(RTLatin1CalcUtf8LenEx);
980
981
982/**
983 * Calculates the Latin-1 length of a string, validating the encoding while
984 * doing so.
985 *
986 * @returns IPRT status code.
987 * @param psz Pointer to the UTF-8 string.
988 * @param cchIn The max length of the string. (btw cch = cb)
989 * Use RTSTR_MAX if all of the string is to be examined.
990 * @param pcch Where to store the length of the Latin-1 string in bytes.
991 */
992static int rtUtf8CalcLatin1Length(const char *psz, size_t cchIn, size_t *pcch)
993{
994 size_t cch = 0;
995 for (;;)
996 {
997 RTUNICP Cp;
998 size_t cchCp;
999 int rc = RTStrGetCpNEx(&psz, &cchIn, &Cp);
1000 if (Cp == 0 || rc == VERR_END_OF_STRING)
1001 break;
1002 if (RT_FAILURE(rc))
1003 return rc;
1004 cchCp = RTLatin1CpSize(Cp);
1005 if (cchCp == 0)
1006 return VERR_NO_TRANSLATION;
1007 cch += cchCp;
1008 }
1009
1010 /* done */
1011 *pcch = cch;
1012 return VINF_SUCCESS;
1013}
1014
1015
1016/**
1017 * Recodes a valid UTF-8 string as Latin-1.
1018 *
1019 * Since we know the input is valid, we do *not* perform encoding or length checks.
1020 *
1021 * @returns iprt status code.
1022 * @param pszIn The UTF-8 string to recode. This is a valid encoding.
1023 * @param cchIn The number of chars (the type char, so bytes if you like) to process of the UTF-8 string.
1024 * The recoding will stop when cch or '\\0' is reached. Pass RTSTR_MAX to process up to '\\0'.
1025 * @param psz Where to store the Latin-1 string.
1026 * @param cch The number of characters the pszOut buffer can hold, excluding the terminator ('\\0').
1027 */
1028static int rtUtf8RecodeAsLatin1(const char *pszIn, size_t cchIn, char *psz, size_t cch)
1029{
1030 int rc;
1031 for (;;)
1032 {
1033 RTUNICP Cp;
1034 size_t cchCp;
1035 rc = RTStrGetCpNEx(&pszIn, &cchIn, &Cp);
1036 if (Cp == 0 || RT_FAILURE(rc))
1037 break;
1038 cchCp = RTLatin1CpSize(Cp);
1039 if (RT_UNLIKELY(cch < cchCp))
1040 {
1041 RTStrAssertMsgFailed(("Buffer overflow! 1\n"));
1042 rc = VERR_BUFFER_OVERFLOW;
1043 break;
1044 }
1045 cch -= cchCp;
1046 psz = RTLatin1PutCp(psz, Cp);
1047 }
1048
1049 /* done */
1050 if (rc == VERR_END_OF_STRING)
1051 rc = VINF_SUCCESS;
1052 *psz = '\0';
1053 return rc;
1054}
1055
1056
1057
1058RTDECL(int) RTStrToLatin1Tag(const char *pszString, char **ppszString, const char *pszTag)
1059{
1060 /*
1061 * Validate input.
1062 */
1063 Assert(VALID_PTR(ppszString));
1064 Assert(VALID_PTR(pszString));
1065 *ppszString = NULL;
1066
1067 /*
1068 * Validate the UTF-8 input and calculate the length of the Latin-1 string.
1069 */
1070 size_t cch;
1071 int rc = rtUtf8CalcLatin1Length(pszString, RTSTR_MAX, &cch);
1072 if (RT_SUCCESS(rc))
1073 {
1074 /*
1075 * Allocate buffer.
1076 */
1077 char *psz = (char *)RTMemAllocTag(cch + 1, pszTag);
1078 if (psz)
1079 {
1080 /*
1081 * Encode the UTF-16 string.
1082 */
1083 rc = rtUtf8RecodeAsLatin1(pszString, RTSTR_MAX, psz, cch);
1084 if (RT_SUCCESS(rc))
1085 {
1086 *ppszString = psz;
1087 return rc;
1088 }
1089 RTMemFree(psz);
1090 }
1091 else
1092 rc = VERR_NO_STR_MEMORY;
1093 }
1094 return rc;
1095}
1096RT_EXPORT_SYMBOL(RTStrToLatin1Tag);
1097
1098
1099RTDECL(int) RTStrToLatin1ExTag(const char *pszString, size_t cchString,
1100 char **ppsz, size_t cch, size_t *pcch, const char *pszTag)
1101{
1102 /*
1103 * Validate input.
1104 */
1105 Assert(VALID_PTR(pszString));
1106 Assert(VALID_PTR(ppsz));
1107 Assert(!pcch || VALID_PTR(pcch));
1108
1109 /*
1110 * Validate the UTF-8 input and calculate the length of the UTF-16 string.
1111 */
1112 size_t cchResult;
1113 int rc = rtUtf8CalcLatin1Length(pszString, cchString, &cchResult);
1114 if (RT_SUCCESS(rc))
1115 {
1116 if (pcch)
1117 *pcch = cchResult;
1118
1119 /*
1120 * Check buffer size / Allocate buffer.
1121 */
1122 bool fShouldFree;
1123 char *pszResult;
1124 if (cch > 0 && *ppsz)
1125 {
1126 fShouldFree = false;
1127 if (cch <= cchResult)
1128 return VERR_BUFFER_OVERFLOW;
1129 pszResult = *ppsz;
1130 }
1131 else
1132 {
1133 *ppsz = NULL;
1134 fShouldFree = true;
1135 cch = RT_MAX(cchResult + 1, cch);
1136 pszResult = (char *)RTMemAllocTag(cch, pszTag);
1137 }
1138 if (pszResult)
1139 {
1140 /*
1141 * Encode the Latin-1 string.
1142 */
1143 rc = rtUtf8RecodeAsLatin1(pszString, cchString, pszResult, cch - 1);
1144 if (RT_SUCCESS(rc))
1145 {
1146 *ppsz = pszResult;
1147 return rc;
1148 }
1149 if (fShouldFree)
1150 RTMemFree(pszResult);
1151 }
1152 else
1153 rc = VERR_NO_STR_MEMORY;
1154 }
1155 return rc;
1156}
1157RT_EXPORT_SYMBOL(RTStrToLatin1Tag);
1158
1159
1160RTDECL(size_t) RTStrCalcLatin1Len(const char *psz)
1161{
1162 size_t cch;
1163 int rc = rtUtf8CalcLatin1Length(psz, RTSTR_MAX, &cch);
1164 return RT_SUCCESS(rc) ? cch : 0;
1165}
1166RT_EXPORT_SYMBOL(RTStrCalcLatin1Len);
1167
1168
1169RTDECL(int) RTStrCalcLatin1LenEx(const char *psz, size_t cchIn, size_t *pcch)
1170{
1171 size_t cch;
1172 int rc = rtUtf8CalcLatin1Length(psz, cchIn, &cch);
1173 if (pcch)
1174 *pcch = RT_SUCCESS(rc) ? cch : ~(size_t)0;
1175 return rc;
1176}
1177RT_EXPORT_SYMBOL(RTStrCalcLatin1LenEx);
1178
1179
1180/**
1181 * Handle invalid encodings passed to RTStrGetCp() and RTStrGetCpEx().
1182 * @returns rc
1183 * @param ppsz The pointer to the string position point.
1184 * @param pCp Where to store RTUNICP_INVALID.
1185 * @param rc The iprt error code.
1186 */
1187static int rtStrGetCpExFailure(const char **ppsz, PRTUNICP pCp, int rc)
1188{
1189 /*
1190 * Try find a valid encoding.
1191 */
1192 (*ppsz)++; /** @todo code this! */
1193 *pCp = RTUNICP_INVALID;
1194 return rc;
1195}
1196
1197
1198RTDECL(RTUNICP) RTStrGetCpInternal(const char *psz)
1199{
1200 RTUNICP Cp;
1201 RTStrGetCpExInternal(&psz, &Cp);
1202 return Cp;
1203}
1204RT_EXPORT_SYMBOL(RTStrGetCpInternal);
1205
1206
1207RTDECL(int) RTStrGetCpExInternal(const char **ppsz, PRTUNICP pCp)
1208{
1209 const unsigned char *puch = (const unsigned char *)*ppsz;
1210 const unsigned char uch = *puch;
1211 RTUNICP uc;
1212
1213 /* ASCII ? */
1214 if (!(uch & RT_BIT(7)))
1215 {
1216 uc = uch;
1217 puch++;
1218 }
1219 else if (uch & RT_BIT(6))
1220 {
1221 /* figure the length and validate the first octet. */
1222/** @todo RT_USE_RTC_3629 */
1223 unsigned cb;
1224 if (!(uch & RT_BIT(5)))
1225 cb = 2;
1226 else if (!(uch & RT_BIT(4)))
1227 cb = 3;
1228 else if (!(uch & RT_BIT(3)))
1229 cb = 4;
1230 else if (!(uch & RT_BIT(2)))
1231 cb = 5;
1232 else if (!(uch & RT_BIT(1)))
1233 cb = 6;
1234 else
1235 {
1236 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(strlen((char *)puch), 10), puch));
1237 return rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING);
1238 }
1239
1240 /* validate the rest */
1241 switch (cb)
1242 {
1243 case 6:
1244 RTStrAssertMsgReturn((puch[5] & 0xc0) == 0x80, ("6/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1245 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1246 case 5:
1247 RTStrAssertMsgReturn((puch[4] & 0xc0) == 0x80, ("5/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1248 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1249 case 4:
1250 RTStrAssertMsgReturn((puch[3] & 0xc0) == 0x80, ("4/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1251 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1252 case 3:
1253 RTStrAssertMsgReturn((puch[2] & 0xc0) == 0x80, ("3/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1254 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1255 case 2:
1256 RTStrAssertMsgReturn((puch[1] & 0xc0) == 0x80, ("2/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1257 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1258 break;
1259 }
1260
1261 /* get and validate the code point. */
1262 switch (cb)
1263 {
1264 case 6:
1265 uc = (puch[5] & 0x3f)
1266 | ((RTUNICP)(puch[4] & 0x3f) << 6)
1267 | ((RTUNICP)(puch[3] & 0x3f) << 12)
1268 | ((RTUNICP)(puch[2] & 0x3f) << 18)
1269 | ((RTUNICP)(puch[1] & 0x3f) << 24)
1270 | ((RTUNICP)(uch & 0x01) << 30);
1271 RTStrAssertMsgReturn(uc >= 0x04000000 && uc <= 0x7fffffff,
1272 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1273 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1274 break;
1275 case 5:
1276 uc = (puch[4] & 0x3f)
1277 | ((RTUNICP)(puch[3] & 0x3f) << 6)
1278 | ((RTUNICP)(puch[2] & 0x3f) << 12)
1279 | ((RTUNICP)(puch[1] & 0x3f) << 18)
1280 | ((RTUNICP)(uch & 0x03) << 24);
1281 RTStrAssertMsgReturn(uc >= 0x00200000 && uc <= 0x03ffffff,
1282 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1283 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1284 break;
1285 case 4:
1286 uc = (puch[3] & 0x3f)
1287 | ((RTUNICP)(puch[2] & 0x3f) << 6)
1288 | ((RTUNICP)(puch[1] & 0x3f) << 12)
1289 | ((RTUNICP)(uch & 0x07) << 18);
1290 RTStrAssertMsgReturn(uc >= 0x00010000 && uc <= 0x001fffff,
1291 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1292 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1293 break;
1294 case 3:
1295 uc = (puch[2] & 0x3f)
1296 | ((RTUNICP)(puch[1] & 0x3f) << 6)
1297 | ((RTUNICP)(uch & 0x0f) << 12);
1298 RTStrAssertMsgReturn(uc >= 0x00000800 && uc <= 0x0000fffd,
1299 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1300 rtStrGetCpExFailure(ppsz, pCp, uc == 0xffff || uc == 0xfffe ? VERR_CODE_POINT_ENDIAN_INDICATOR : VERR_INVALID_UTF8_ENCODING));
1301 RTStrAssertMsgReturn(uc < 0xd800 || uc > 0xdfff,
1302 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1303 rtStrGetCpExFailure(ppsz, pCp, VERR_CODE_POINT_SURROGATE));
1304 break;
1305 case 2:
1306 uc = (puch[1] & 0x3f)
1307 | ((RTUNICP)(uch & 0x1f) << 6);
1308 RTStrAssertMsgReturn(uc >= 0x00000080 && uc <= 0x000007ff,
1309 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1310 rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING));
1311 break;
1312 default: /* impossible, but GCC is bitching. */
1313 uc = RTUNICP_INVALID;
1314 break;
1315 }
1316 puch += cb;
1317 }
1318 else
1319 {
1320 /* 6th bit is always set. */
1321 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(strlen((char *)puch), 10), puch));
1322 return rtStrGetCpExFailure(ppsz, pCp, VERR_INVALID_UTF8_ENCODING);
1323 }
1324 *pCp = uc;
1325 *ppsz = (const char *)puch;
1326 return VINF_SUCCESS;
1327}
1328RT_EXPORT_SYMBOL(RTStrGetCpExInternal);
1329
1330
1331/**
1332 * Handle invalid encodings passed to RTStrGetCpNEx().
1333 * @returns rc
1334 * @param ppsz The pointer to the string position point.
1335 * @param pcch Pointer to the string length.
1336 * @param pCp Where to store RTUNICP_INVALID.
1337 * @param rc The iprt error code.
1338 */
1339static int rtStrGetCpNExFailure(const char **ppsz, size_t *pcch, PRTUNICP pCp, int rc)
1340{
1341 /*
1342 * Try find a valid encoding.
1343 */
1344 (*ppsz)++; /** @todo code this! */
1345 (*pcch)--;
1346 *pCp = RTUNICP_INVALID;
1347 return rc;
1348}
1349
1350
1351RTDECL(int) RTStrGetCpNExInternal(const char **ppsz, size_t *pcch, PRTUNICP pCp)
1352{
1353 const unsigned char *puch = (const unsigned char *)*ppsz;
1354 const unsigned char uch = *puch;
1355 size_t cch = *pcch;
1356 RTUNICP uc;
1357
1358 if (cch == 0)
1359 {
1360 *pCp = RTUNICP_INVALID;
1361 return VERR_END_OF_STRING;
1362 }
1363
1364 /* ASCII ? */
1365 if (!(uch & RT_BIT(7)))
1366 {
1367 uc = uch;
1368 puch++;
1369 cch--;
1370 }
1371 else if (uch & RT_BIT(6))
1372 {
1373 /* figure the length and validate the first octet. */
1374/** @todo RT_USE_RTC_3629 */
1375 unsigned cb;
1376 if (!(uch & RT_BIT(5)))
1377 cb = 2;
1378 else if (!(uch & RT_BIT(4)))
1379 cb = 3;
1380 else if (!(uch & RT_BIT(3)))
1381 cb = 4;
1382 else if (!(uch & RT_BIT(2)))
1383 cb = 5;
1384 else if (!(uch & RT_BIT(1)))
1385 cb = 6;
1386 else
1387 {
1388 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(strlen((char *)puch), 10), puch));
1389 return rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING);
1390 }
1391
1392 if (cb > cch)
1393 return rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING);
1394
1395 /* validate the rest */
1396 switch (cb)
1397 {
1398 case 6:
1399 RTStrAssertMsgReturn((puch[5] & 0xc0) == 0x80, ("6/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1400 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1401 case 5:
1402 RTStrAssertMsgReturn((puch[4] & 0xc0) == 0x80, ("5/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1403 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1404 case 4:
1405 RTStrAssertMsgReturn((puch[3] & 0xc0) == 0x80, ("4/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1406 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1407 case 3:
1408 RTStrAssertMsgReturn((puch[2] & 0xc0) == 0x80, ("3/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1409 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1410 case 2:
1411 RTStrAssertMsgReturn((puch[1] & 0xc0) == 0x80, ("2/%u: %.*Rhxs\n", cb, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1412 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1413 break;
1414 }
1415
1416 /* get and validate the code point. */
1417 switch (cb)
1418 {
1419 case 6:
1420 uc = (puch[5] & 0x3f)
1421 | ((RTUNICP)(puch[4] & 0x3f) << 6)
1422 | ((RTUNICP)(puch[3] & 0x3f) << 12)
1423 | ((RTUNICP)(puch[2] & 0x3f) << 18)
1424 | ((RTUNICP)(puch[1] & 0x3f) << 24)
1425 | ((RTUNICP)(uch & 0x01) << 30);
1426 RTStrAssertMsgReturn(uc >= 0x04000000 && uc <= 0x7fffffff,
1427 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1428 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1429 break;
1430 case 5:
1431 uc = (puch[4] & 0x3f)
1432 | ((RTUNICP)(puch[3] & 0x3f) << 6)
1433 | ((RTUNICP)(puch[2] & 0x3f) << 12)
1434 | ((RTUNICP)(puch[1] & 0x3f) << 18)
1435 | ((RTUNICP)(uch & 0x03) << 24);
1436 RTStrAssertMsgReturn(uc >= 0x00200000 && uc <= 0x03ffffff,
1437 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1438 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1439 break;
1440 case 4:
1441 uc = (puch[3] & 0x3f)
1442 | ((RTUNICP)(puch[2] & 0x3f) << 6)
1443 | ((RTUNICP)(puch[1] & 0x3f) << 12)
1444 | ((RTUNICP)(uch & 0x07) << 18);
1445 RTStrAssertMsgReturn(uc >= 0x00010000 && uc <= 0x001fffff,
1446 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1447 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1448 break;
1449 case 3:
1450 uc = (puch[2] & 0x3f)
1451 | ((RTUNICP)(puch[1] & 0x3f) << 6)
1452 | ((RTUNICP)(uch & 0x0f) << 12);
1453 RTStrAssertMsgReturn(uc >= 0x00000800 && uc <= 0x0000fffd,
1454 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1455 rtStrGetCpNExFailure(ppsz, pcch, pCp, uc == 0xffff || uc == 0xfffe ? VERR_CODE_POINT_ENDIAN_INDICATOR : VERR_INVALID_UTF8_ENCODING));
1456 RTStrAssertMsgReturn(uc < 0xd800 || uc > 0xdfff,
1457 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1458 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_CODE_POINT_SURROGATE));
1459 break;
1460 case 2:
1461 uc = (puch[1] & 0x3f)
1462 | ((RTUNICP)(uch & 0x1f) << 6);
1463 RTStrAssertMsgReturn(uc >= 0x00000080 && uc <= 0x000007ff,
1464 ("%u: cp=%#010RX32: %.*Rhxs\n", cb, uc, RT_MIN(cb + 10, strlen((char *)puch)), puch),
1465 rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING));
1466 break;
1467 default: /* impossible, but GCC is bitching. */
1468 uc = RTUNICP_INVALID;
1469 break;
1470 }
1471 puch += cb;
1472 cch -= cb;
1473 }
1474 else
1475 {
1476 /* 6th bit is always set. */
1477 RTStrAssertMsgFailed(("Invalid UTF-8 first byte: %.*Rhxs\n", RT_MIN(strlen((char *)puch), 10), puch));
1478 return rtStrGetCpNExFailure(ppsz, pcch, pCp, VERR_INVALID_UTF8_ENCODING);
1479 }
1480 *pCp = uc;
1481 *ppsz = (const char *)puch;
1482 (*pcch) = cch;
1483 return VINF_SUCCESS;
1484}
1485RT_EXPORT_SYMBOL(RTStrGetCpNExInternal);
1486
1487
1488RTDECL(char *) RTStrPutCpInternal(char *psz, RTUNICP uc)
1489{
1490 unsigned char *puch = (unsigned char *)psz;
1491 if (uc < 0x80)
1492 *puch++ = (unsigned char )uc;
1493 else if (uc < 0x00000800)
1494 {
1495 *puch++ = 0xc0 | (uc >> 6);
1496 *puch++ = 0x80 | (uc & 0x3f);
1497 }
1498 else if (uc < 0x00010000)
1499 {
1500/** @todo RT_USE_RTC_3629 */
1501 if ( uc < 0x0000d8000
1502 || ( uc > 0x0000dfff
1503 && uc < 0x0000fffe))
1504 {
1505 *puch++ = 0xe0 | (uc >> 12);
1506 *puch++ = 0x80 | ((uc >> 6) & 0x3f);
1507 *puch++ = 0x80 | (uc & 0x3f);
1508 }
1509 else
1510 {
1511 AssertMsgFailed(("Invalid code point U+%05x!\n", uc));
1512 *puch++ = 0x7f;
1513 }
1514 }
1515/** @todo RT_USE_RTC_3629 */
1516 else if (uc < 0x00200000)
1517 {
1518 *puch++ = 0xf0 | (uc >> 18);
1519 *puch++ = 0x80 | ((uc >> 12) & 0x3f);
1520 *puch++ = 0x80 | ((uc >> 6) & 0x3f);
1521 *puch++ = 0x80 | (uc & 0x3f);
1522 }
1523 else if (uc < 0x04000000)
1524 {
1525 *puch++ = 0xf8 | (uc >> 24);
1526 *puch++ = 0x80 | ((uc >> 18) & 0x3f);
1527 *puch++ = 0x80 | ((uc >> 12) & 0x3f);
1528 *puch++ = 0x80 | ((uc >> 6) & 0x3f);
1529 *puch++ = 0x80 | (uc & 0x3f);
1530 }
1531 else if (uc <= 0x7fffffff)
1532 {
1533 *puch++ = 0xfc | (uc >> 30);
1534 *puch++ = 0x80 | ((uc >> 24) & 0x3f);
1535 *puch++ = 0x80 | ((uc >> 18) & 0x3f);
1536 *puch++ = 0x80 | ((uc >> 12) & 0x3f);
1537 *puch++ = 0x80 | ((uc >> 6) & 0x3f);
1538 *puch++ = 0x80 | (uc & 0x3f);
1539 }
1540 else
1541 {
1542 AssertMsgFailed(("Invalid code point U+%08x!\n", uc));
1543 *puch++ = 0x7f;
1544 }
1545
1546 return (char *)puch;
1547}
1548RT_EXPORT_SYMBOL(RTStrPutCpInternal);
1549
1550
1551RTDECL(char *) RTStrPrevCp(const char *pszStart, const char *psz)
1552{
1553 if (pszStart < psz)
1554 {
1555 /* simple char? */
1556 const unsigned char *puch = (const unsigned char *)psz;
1557 unsigned uch = *--puch;
1558 if (!(uch & RT_BIT(7)))
1559 return (char *)puch;
1560 RTStrAssertMsgReturn(!(uch & RT_BIT(6)), ("uch=%#x\n", uch), (char *)pszStart);
1561
1562 /* two or more. */
1563 uint32_t uMask = 0xffffffc0;
1564 while ( (const unsigned char *)pszStart < puch
1565 && !(uMask & 1))
1566 {
1567 uch = *--puch;
1568 if ((uch & 0xc0) != 0x80)
1569 {
1570 RTStrAssertMsgReturn((uch & (uMask >> 1)) == (uMask & 0xff),
1571 ("Invalid UTF-8 encoding: %.*Rhxs puch=%p psz=%p\n", psz - (char *)puch, puch, psz),
1572 (char *)pszStart);
1573 return (char *)puch;
1574 }
1575 uMask >>= 1;
1576 }
1577 RTStrAssertMsgFailed(("Invalid UTF-8 encoding: %.*Rhxs puch=%p psz=%p\n", psz - (char *)puch, puch, psz));
1578 }
1579 return (char *)pszStart;
1580}
1581RT_EXPORT_SYMBOL(RTStrPrevCp);
1582
Note: See TracBrowser for help on using the repository browser.

© 2024 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette