1/*
2 * Copyright (c) 1998, 2016, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation. Oracle designates this
8 * particular file as subject to the "Classpath" exception as provided
9 * by Oracle in the LICENSE file that accompanied this code.
10 *
11 * This code is distributed in the hope that it will be useful, but WITHOUT
12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
14 * version 2 for more details (a copy is included in the LICENSE file that
15 * accompanied this code).
16 *
17 * You should have received a copy of the GNU General Public License version
18 * 2 along with this work; if not, write to the Free Software Foundation,
19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 *
21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
22 * or visit www.oracle.com if you need additional information or have any
23 * questions.
24 */
25
26#include <stdlib.h>
27#include <ctype.h>
28
29#include "jni.h"
30
31#include "utf_util.h"
32
33
34/* Error and assert macros */
35#define UTF_ERROR(m) utfError(__FILE__, __LINE__, m)
36#define UTF_ASSERT(x) ( (x)==0 ? UTF_ERROR("ASSERT ERROR " #x) : (void)0 )
37
38// Platform independed part
39
40static void utfError(char *file, int line, char *message) {
41 (void)fprintf(stderr, "UTF ERROR [\"%s\":%d]: %s\n", file, line, message);
42 abort();
43}
44
45/* Determine length of this Standard UTF-8 in Modified UTF-8.
46 * Validation is done of the basic UTF encoding rules, returns
47 * length (no change) when errors are detected in the UTF encoding.
48 *
49 * Note: Accepts Modified UTF-8 also, no verification on the
50 * correctness of Standard UTF-8 is done. e,g, 0xC080 input is ok.
51 */
52int JNICALL utf8sToUtf8mLength(jbyte *string, int length) {
53 int newLength;
54 int i;
55
56 newLength = 0;
57 for ( i = 0 ; i < length ; i++ ) {
58 unsigned byte;
59
60 byte = (unsigned char)string[i];
61 if ( (byte & 0x80) == 0 ) { /* 1byte encoding */
62 newLength++;
63 if ( byte == 0 ) {
64 newLength++; /* We gain one byte in length on NULL bytes */
65 }
66 } else if ( (byte & 0xE0) == 0xC0 ) { /* 2byte encoding */
67 /* Check encoding of following bytes */
68 if ( (i+1) >= length || (string[i+1] & 0xC0) != 0x80 ) {
69 break; /* Error condition */
70 }
71 i++; /* Skip next byte */
72 newLength += 2;
73 } else if ( (byte & 0xF0) == 0xE0 ) { /* 3byte encoding */
74 /* Check encoding of following bytes */
75 if ( (i+2) >= length || (string[i+1] & 0xC0) != 0x80
76 || (string[i+2] & 0xC0) != 0x80 ) {
77 break; /* Error condition */
78 }
79 i += 2; /* Skip next two bytes */
80 newLength += 3;
81 } else if ( (byte & 0xF8) == 0xF0 ) { /* 4byte encoding */
82 /* Check encoding of following bytes */
83 if ( (i+3) >= length || (string[i+1] & 0xC0) != 0x80
84 || (string[i+2] & 0xC0) != 0x80
85 || (string[i+3] & 0xC0) != 0x80 ) {
86 break; /* Error condition */
87 }
88 i += 3; /* Skip next 3 bytes */
89 newLength += 6; /* 4byte encoding turns into 2 3byte ones */
90 } else {
91 break; /* Error condition */
92 }
93 }
94 if ( i != length ) {
95 /* Error in finding new length, return old length so no conversion */
96 /* FIXUP: ERROR_MESSAGE? */
97 return length;
98 }
99 return newLength;
100}
101
102/* Convert Standard UTF-8 to Modified UTF-8.
103 * Assumes the UTF-8 encoding was validated by utf8mLength() above.
104 *
105 * Note: Accepts Modified UTF-8 also, no verification on the
106 * correctness of Standard UTF-8 is done. e,g, 0xC080 input is ok.
107 */
108void JNICALL utf8sToUtf8m(jbyte *string, int length, jbyte *newString, int newLength) {
109 int i;
110 int j;
111
112 j = 0;
113 for ( i = 0 ; i < length ; i++ ) {
114 unsigned byte1;
115
116 byte1 = (unsigned char)string[i];
117
118 /* NULL bytes and bytes starting with 11110xxx are special */
119 if ( (byte1 & 0x80) == 0 ) { /* 1byte encoding */
120 if ( byte1 == 0 ) {
121 /* Bits out: 11000000 10000000 */
122 newString[j++] = (jbyte)0xC0;
123 newString[j++] = (jbyte)0x80;
124 } else {
125 /* Single byte */
126 newString[j++] = byte1;
127 }
128 } else if ( (byte1 & 0xE0) == 0xC0 ) { /* 2byte encoding */
129 newString[j++] = byte1;
130 newString[j++] = string[++i];
131 } else if ( (byte1 & 0xF0) == 0xE0 ) { /* 3byte encoding */
132 newString[j++] = byte1;
133 newString[j++] = string[++i];
134 newString[j++] = string[++i];
135 } else if ( (byte1 & 0xF8) == 0xF0 ) { /* 4byte encoding */
136 /* Beginning of 4byte encoding, turn into 2 3byte encodings */
137 unsigned byte2, byte3, byte4, u21;
138
139 /* Bits in: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
140 byte2 = (unsigned char)string[++i];
141 byte3 = (unsigned char)string[++i];
142 byte4 = (unsigned char)string[++i];
143 /* Reconstruct full 21bit value */
144 u21 = (byte1 & 0x07) << 18;
145 u21 += (byte2 & 0x3F) << 12;
146 u21 += (byte3 & 0x3F) << 6;
147 u21 += (byte4 & 0x3F);
148 /* Bits out: 11101101 1010xxxx 10xxxxxx */
149 newString[j++] = (jbyte)0xED;
150 newString[j++] = (jbyte)(0xA0 + (((u21 >> 16) - 1) & 0x0F));
151 newString[j++] = (jbyte)(0x80 + ((u21 >> 10) & 0x3F));
152 /* Bits out: 11101101 1011xxxx 10xxxxxx */
153 newString[j++] = (jbyte)0xED;
154 newString[j++] = (jbyte)(0xB0 + ((u21 >> 6) & 0x0F));
155 newString[j++] = byte4;
156 }
157 }
158 UTF_ASSERT(i==length);
159 UTF_ASSERT(j==newLength);
160 newString[j] = (jbyte)0;
161}
162
163/* Given a Modified UTF-8 string, calculate the Standard UTF-8 length.
164 * Basic validation of the UTF encoding rules is done, and length is
165 * returned (no change) when errors are detected.
166 *
167 * Note: No validation is made that this is indeed Modified UTF-8 coming in.
168 *
169 */
170int JNICALL utf8mToUtf8sLength(jbyte *string, int length) {
171 int newLength;
172 int i;
173
174 newLength = 0;
175 for ( i = 0 ; i < length ; i++ ) {
176 unsigned byte1, byte2, byte3, byte4, byte5, byte6;
177
178 byte1 = (unsigned char)string[i];
179 if ( (byte1 & 0x80) == 0 ) { /* 1byte encoding */
180 newLength++;
181 } else if ( (byte1 & 0xE0) == 0xC0 ) { /* 2byte encoding */
182 /* Check encoding of following bytes */
183 if ( (i+1) >= length || (string[i+1] & 0xC0) != 0x80 ) {
184 break; /* Error condition */
185 }
186 byte2 = (unsigned char)string[++i];
187 if ( byte1 != 0xC0 || byte2 != 0x80 ) {
188 newLength += 2; /* Normal 2byte encoding, not 0xC080 */
189 } else {
190 newLength++; /* We will turn 0xC080 into 0 */
191 }
192 } else if ( (byte1 & 0xF0) == 0xE0 ) { /* 3byte encoding */
193 /* Check encoding of following bytes */
194 if ( (i+2) >= length || (string[i+1] & 0xC0) != 0x80
195 || (string[i+2] & 0xC0) != 0x80 ) {
196 break; /* Error condition */
197 }
198 byte2 = (unsigned char)string[++i];
199 byte3 = (unsigned char)string[++i];
200 newLength += 3;
201 /* Possible process a second 3byte encoding */
202 if ( (i+3) < length && byte1 == 0xED && (byte2 & 0xF0) == 0xA0 ) {
203 /* See if this is a pair of 3byte encodings */
204 byte4 = (unsigned char)string[i+1];
205 byte5 = (unsigned char)string[i+2];
206 byte6 = (unsigned char)string[i+3];
207 if ( byte4 == 0xED && (byte5 & 0xF0) == 0xB0 ) {
208 /* Check encoding of 3rd byte */
209 if ( (byte6 & 0xC0) != 0x80 ) {
210 break; /* Error condition */
211 }
212 newLength++; /* New string will have 4byte encoding */
213 i += 3; /* Skip next 3 bytes */
214 }
215 }
216 } else {
217 break; /* Error condition */
218 }
219 }
220 if ( i != length ) {
221 /* Error in UTF encoding */
222 /* FIXUP: ERROR_MESSAGE()? */
223 return length;
224 }
225 return newLength;
226}
227
228/* Convert a Modified UTF-8 string into a Standard UTF-8 string
229 * It is assumed that this string has been validated in terms of the
230 * basic UTF encoding rules by utf8Length() above.
231 *
232 * Note: No validation is made that this is indeed Modified UTF-8 coming in.
233 *
234 */
235void JNICALL utf8mToUtf8s(jbyte *string, int length, jbyte *newString, int newLength) {
236 int i;
237 int j;
238
239 j = 0;
240 for ( i = 0 ; i < length ; i++ ) {
241 unsigned byte1, byte2, byte3, byte4, byte5, byte6;
242
243 byte1 = (unsigned char)string[i];
244 if ( (byte1 & 0x80) == 0 ) { /* 1byte encoding */
245 /* Single byte */
246 newString[j++] = byte1;
247 } else if ( (byte1 & 0xE0) == 0xC0 ) { /* 2byte encoding */
248 byte2 = (unsigned char)string[++i];
249 if ( byte1 != 0xC0 || byte2 != 0x80 ) {
250 newString[j++] = byte1;
251 newString[j++] = byte2;
252 } else {
253 newString[j++] = 0;
254 }
255 } else if ( (byte1 & 0xF0) == 0xE0 ) { /* 3byte encoding */
256 byte2 = (unsigned char)string[++i];
257 byte3 = (unsigned char)string[++i];
258 if ( i+3 < length && byte1 == 0xED && (byte2 & 0xF0) == 0xA0 ) {
259 /* See if this is a pair of 3byte encodings */
260 byte4 = (unsigned char)string[i+1];
261 byte5 = (unsigned char)string[i+2];
262 byte6 = (unsigned char)string[i+3];
263 if ( byte4 == 0xED && (byte5 & 0xF0) == 0xB0 ) {
264 unsigned u21;
265
266 /* Bits in: 11101101 1010xxxx 10xxxxxx */
267 /* Bits in: 11101101 1011xxxx 10xxxxxx */
268 i += 3;
269
270 /* Reconstruct 21 bit code */
271 u21 = ((byte2 & 0x0F) + 1) << 16;
272 u21 += (byte3 & 0x3F) << 10;
273 u21 += (byte5 & 0x0F) << 6;
274 u21 += (byte6 & 0x3F);
275
276 /* Bits out: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */
277
278 /* Convert to 4byte encoding */
279 newString[j++] = 0xF0 + ((u21 >> 18) & 0x07);
280 newString[j++] = 0x80 + ((u21 >> 12) & 0x3F);
281 newString[j++] = 0x80 + ((u21 >> 6) & 0x3F);
282 newString[j++] = 0x80 + (u21 & 0x3F);
283 continue;
284 }
285 }
286 /* Normal 3byte encoding */
287 newString[j++] = byte1;
288 newString[j++] = byte2;
289 newString[j++] = byte3;
290 }
291 }
292 UTF_ASSERT(i==length);
293 UTF_ASSERT(j==newLength);
294 newString[j] = 0;
295}
296
297#ifdef _WIN32
298// Microsoft Windows specific part
299
300#include <windows.h>
301
302static UINT getCodepage() {
303 LANGID langID;
304 LCID localeID;
305 TCHAR strCodePage[7]; // ANSI code page id
306
307 static UINT intCodePage = -1;
308
309 if (intCodePage == -1) {
310 // Firts call, get codepage from the os
311 langID = LANGIDFROMLCID(GetUserDefaultLCID());
312 localeID = MAKELCID(langID, SORT_DEFAULT);
313 if (GetLocaleInfo(localeID, LOCALE_IDEFAULTANSICODEPAGE,
314 strCodePage, sizeof(strCodePage)/sizeof(TCHAR)) > 0 ) {
315 intCodePage = atoi(strCodePage);
316 }
317 else {
318 intCodePage = GetACP();
319 }
320 }
321
322 return intCodePage;
323}
324
325/*
326 * Get wide string (assumes len>0)
327 */
328static WCHAR* getWideString(UINT codePage, char* str, int len, int *pwlen) {
329 int wlen;
330 WCHAR* wstr;
331
332 /* Convert the string to WIDE string */
333 wlen = MultiByteToWideChar(codePage, 0, str, len, NULL, 0);
334 *pwlen = wlen;
335 if (wlen <= 0) {
336 UTF_ERROR(("Can't get WIDE string length"));
337 return NULL;
338 }
339 wstr = (WCHAR*)malloc(wlen * sizeof(WCHAR));
340 if (wstr == NULL) {
341 UTF_ERROR(("Can't malloc() any space"));
342 return NULL;
343 }
344 if (MultiByteToWideChar(codePage, 0, str, len, wstr, wlen) == 0) {
345 UTF_ERROR(("Can't get WIDE string"));
346 return NULL;
347 }
348 return wstr;
349}
350
351/*
352 * Convert UTF-8 to a platform string
353 * NOTE: outputBufSize includes the space for the trailing 0.
354 */
355int JNICALL utf8ToPlatform(jbyte *utf8, int len, char* output, int outputBufSize) {
356 int wlen;
357 int plen;
358 WCHAR* wstr;
359 UINT codepage;
360 int outputMaxLen;
361
362 UTF_ASSERT(utf8);
363 UTF_ASSERT(output);
364 UTF_ASSERT(len >= 0);
365 UTF_ASSERT(outputBufSize > len);
366 outputMaxLen = outputBufSize - 1; // leave space for trailing 0
367
368 /* Zero length is ok, but we don't need to do much */
369 if ( len == 0 ) {
370 output[0] = 0;
371 return 0;
372 }
373
374 /* Get WIDE string version (assumes len>0) */
375 wstr = getWideString(CP_UTF8, (char*)utf8, len, &wlen);
376 if ( wstr == NULL ) {
377 // Can't allocate WIDE string
378 goto just_copy_bytes;
379 }
380
381 /* Convert WIDE string to MultiByte string */
382 codepage = getCodepage();
383 plen = WideCharToMultiByte(codepage, 0, wstr, wlen,
384 output, outputMaxLen, NULL, NULL);
385 free(wstr);
386 if (plen <= 0) {
387 // Can't convert WIDE string to multi-byte
388 goto just_copy_bytes;
389 }
390 output[plen] = '\0';
391 return plen;
392
393just_copy_bytes:
394 (void)memcpy(output, utf8, len);
395 output[len] = 0;
396 return len;
397}
398
399/*
400 * Convert Platform Encoding to UTF-8.
401 * NOTE: outputBufSize includes the space for the trailing 0.
402 */
403int JNICALL utf8FromPlatform(char *str, int len, jbyte *output, int outputBufSize) {
404 int wlen;
405 int plen;
406 WCHAR* wstr;
407 UINT codepage;
408 int outputMaxLen;
409
410 UTF_ASSERT(str);
411 UTF_ASSERT(output);
412 UTF_ASSERT(len >= 0);
413 UTF_ASSERT(outputBufSize > len);
414 outputMaxLen = outputBufSize - 1; // leave space for trailing 0
415
416 /* Zero length is ok, but we don't need to do much */
417 if ( len == 0 ) {
418 output[0] = 0;
419 return 0;
420 }
421
422 /* Get WIDE string version (assumes len>0) */
423 codepage = getCodepage();
424 wstr = getWideString(codepage, str, len, &wlen);
425 if ( wstr == NULL ) {
426 goto just_copy_bytes;
427 }
428
429 /* Convert WIDE string to UTF-8 string */
430 plen = WideCharToMultiByte(CP_UTF8, 0, wstr, wlen,
431 (char*)output, outputMaxLen, NULL, NULL);
432 free(wstr);
433 if (plen <= 0) {
434 UTF_ERROR(("Can't convert WIDE string to multi-byte"));
435 goto just_copy_bytes;
436 }
437 output[plen] = '\0';
438 return plen;
439
440just_copy_bytes:
441 (void)memcpy(output, str, len);
442 output[len] = 0;
443 return len;
444}
445
446
447#else
448// *NIX specific part
449
450#include <iconv.h>
451#include <locale.h>
452#include <langinfo.h>
453#include <string.h>
454
455typedef enum {TO_UTF8, FROM_UTF8} conv_direction;
456
457/*
458 * Do iconv() conversion.
459 * Returns length or -1 if output overflows.
460 * NOTE: outputBufSize includes the space for the trailing 0.
461 */
462static int iconvConvert(conv_direction drn, char *bytes, size_t len, char *output, size_t outputBufSize) {
463
464 static char *codeset = 0;
465 iconv_t func;
466 size_t bytes_converted;
467 size_t inLeft, outLeft;
468 char *inbuf, *outbuf;
469 int outputMaxLen;
470
471 UTF_ASSERT(bytes);
472 UTF_ASSERT(output);
473 UTF_ASSERT(outputBufSize > len);
474 outputMaxLen = outputBufSize - 1; // leave space for trailing 0
475
476 /* Zero length is ok, but we don't need to do much */
477 if ( len == 0 ) {
478 output[0] = 0;
479 return 0;
480 }
481
482 if (codeset == NULL && codeset != (char *) -1) {
483 // locale is not initialized, do it now
484 if (setlocale(LC_ALL, "") != NULL) {
485 // nl_langinfo returns ANSI_X3.4-1968 by default
486 codeset = (char*)nl_langinfo(CODESET);
487 }
488
489 if (codeset == NULL) {
490 // Not able to intialize process locale from platform one.
491 codeset = (char *) -1;
492 }
493 }
494
495 if (codeset == (char *) -1) {
496 // There was an error during initialization, so just bail out
497 goto just_copy_bytes;
498 }
499
500 func = (drn == TO_UTF8) ? iconv_open(codeset, "UTF-8") : iconv_open("UTF-8", codeset);
501 if (func == (iconv_t) -1) {
502 // Requested charset combination is not supported, conversion couldn't be done.
503 // make sure we will not try it again
504 codeset = (char *) -1;
505 goto just_copy_bytes;
506 }
507
508 // perform conversion
509 inbuf = bytes;
510 outbuf = output;
511 inLeft = len;
512 outLeft = outputMaxLen;
513
514 bytes_converted = iconv(func, (void*)&inbuf, &inLeft, &outbuf, &outLeft);
515 if (bytes_converted == (size_t) -1 || bytes_converted == 0 || inLeft != 0) {
516 // Input string is invalid, not able to convert entire string
517 // or some other iconv error happens.
518 iconv_close(func);
519 goto just_copy_bytes;
520 }
521
522 iconv_close(func);
523 // Overwrite bytes_converted with value of actually stored bytes
524 bytes_converted = outputMaxLen-outLeft;
525 output[bytes_converted] = 0;
526 return bytes_converted;
527
528
529just_copy_bytes:
530 (void)memcpy(output, bytes, len);
531 output[len] = 0;
532 return len;
533 }
534
535/*
536 * Convert UTF-8 to Platform Encoding.
537 * Returns length or -1 if output overflows.
538 * NOTE: outputBufSize includes the space for the trailing 0.
539 */
540int JNICALL utf8ToPlatform(jbyte *utf8, int len, char *output, int outputBufSize) {
541 return iconvConvert(FROM_UTF8, (char*)utf8, len, output, outputBufSize);
542}
543
544/*
545 * Convert Platform Encoding to UTF-8.
546 * Returns length or -1 if output overflows.
547 * NOTE: outputBufSize includes the space for the trailing 0.
548 */
549int JNICALL utf8FromPlatform(char *str, int len, jbyte *output, int outputBufSize) {
550 return iconvConvert(TO_UTF8, str, len, (char*) output, outputBufSize);
551}
552
553#endif
554