1 /*===--- ConvertUTF.c - Universal Character Names conversions ---------------===
2  *
3  *                     The LLVM Compiler Infrastructure
4  *
5  * This file is distributed under the University of Illinois Open Source
6  * License. See LICENSE.TXT for details.
7  *
8  *===------------------------------------------------------------------------=*/
9 /*
10  * Copyright 2001-2004 Unicode, Inc.
11  *
12  * Disclaimer
13  *
14  * This source code is provided as is by Unicode, Inc. No claims are
15  * made as to fitness for any particular purpose. No warranties of any
16  * kind are expressed or implied. The recipient agrees to determine
17  * applicability of information provided. If this file has been
18  * purchased on magnetic or optical media from Unicode, Inc., the
19  * sole remedy for any claim will be exchange of defective media
20  * within 90 days of receipt.
21  *
22  * Limitations on Rights to Redistribute This Code
23  *
24  * Unicode, Inc. hereby grants the right to freely use the information
25  * supplied in this file in the creation of products supporting the
26  * Unicode Standard, and to make copies of this file in any form
27  * for internal or external distribution as long as this notice
28  * remains attached.
29  */
30 
31 /* ---------------------------------------------------------------------
32 
33     Conversions between UTF32, UTF-16, and UTF-8. Source code file.
34     Author: Mark E. Davis, 1994.
35     Rev History: Rick McGowan, fixes & updates May 2001.
36     Sept 2001: fixed const & error conditions per
37         mods suggested by S. Parent & A. Lillich.
38     June 2002: Tim Dodd added detection and handling of incomplete
39         source sequences, enhanced error detection, added casts
40         to eliminate compiler warnings.
41     July 2003: slight mods to back out aggressive FFFE detection.
42     Jan 2004: updated switches in from-UTF8 conversions.
43     Oct 2004: updated to use UNI_MAX_LEGAL_UTF32 in UTF-32 conversions.
44 
45     See the header file "ConvertUTF.h" for complete documentation.
46 
47 ------------------------------------------------------------------------ */
48 
49 
50 #include "llvm/Support/ConvertUTF.h"
51 #ifdef CVTUTF_DEBUG
52 #include <stdio.h>
53 #endif
54 #include <assert.h>
55 
56 static const int halfShift  = 10; /* used for shifting by 10 bits */
57 
58 static const UTF32 halfBase = 0x0010000UL;
59 static const UTF32 halfMask = 0x3FFUL;
60 
61 #define UNI_SUR_HIGH_START  (UTF32)0xD800
62 #define UNI_SUR_HIGH_END    (UTF32)0xDBFF
63 #define UNI_SUR_LOW_START   (UTF32)0xDC00
64 #define UNI_SUR_LOW_END     (UTF32)0xDFFF
65 #define false      0
66 #define true        1
67 
68 /* --------------------------------------------------------------------- */
69 
70 /*
71  * Index into the table below with the first byte of a UTF-8 sequence to
72  * get the number of trailing bytes that are supposed to follow it.
73  * Note that *legal* UTF-8 values can't have 4 or 5-bytes. The table is
74  * left as-is for anyone who may want to do such conversion, which was
75  * allowed in earlier algorithms.
76  */
77 static const char trailingBytesForUTF8[256] = {
78     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,
79     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,
80     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,
81     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,
82     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,
83     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,
84     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,
85     2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5
86 };
87 
88 /*
89  * Magic values subtracted from a buffer value during UTF8 conversion.
90  * This table contains as many values as there might be trailing bytes
91  * in a UTF-8 sequence.
92  */
93 static const UTF32 offsetsFromUTF8[6] = { 0x00000000UL, 0x00003080UL, 0x000E2080UL,
94                      0x03C82080UL, 0xFA082080UL, 0x82082080UL };
95 
96 /*
97  * Once the bits are split out into bytes of UTF-8, this is a mask OR-ed
98  * into the first byte, depending on how many bytes follow.  There are
99  * as many entries in this table as there are UTF-8 sequence types.
100  * (I.e., one byte sequence, two byte... etc.). Remember that sequencs
101  * for *legal* UTF-8 will be 4 or fewer bytes total.
102  */
103 static const UTF8 firstByteMark[7] = { 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
104 
105 /* --------------------------------------------------------------------- */
106 
107 /* The interface converts a whole buffer to avoid function-call overhead.
108  * Constants have been gathered. Loops & conditionals have been removed as
109  * much as possible for efficiency, in favor of drop-through switches.
110  * (See "Note A" at the bottom of the file for equivalent code.)
111  * If your compiler supports it, the "isLegalUTF8" call can be turned
112  * into an inline function.
113  */
114 
115 
116 /* --------------------------------------------------------------------- */
117 
ConvertUTF32toUTF16(const UTF32 ** sourceStart,const UTF32 * sourceEnd,UTF16 ** targetStart,UTF16 * targetEnd,ConversionFlags flags)118 ConversionResult ConvertUTF32toUTF16 (
119         const UTF32** sourceStart, const UTF32* sourceEnd,
120         UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
121     ConversionResult result = conversionOK;
122     const UTF32* source = *sourceStart;
123     UTF16* target = *targetStart;
124     while (source < sourceEnd) {
125         UTF32 ch;
126         if (target >= targetEnd) {
127             result = targetExhausted; break;
128         }
129         ch = *source++;
130         if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
131             /* UTF-16 surrogate values are illegal in UTF-32; 0xffff or 0xfffe are both reserved values */
132             if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
133                 if (flags == strictConversion) {
134                     --source; /* return to the illegal value itself */
135                     result = sourceIllegal;
136                     break;
137                 } else {
138                     *target++ = UNI_REPLACEMENT_CHAR;
139                 }
140             } else {
141                 *target++ = (UTF16)ch; /* normal case */
142             }
143         } else if (ch > UNI_MAX_LEGAL_UTF32) {
144             if (flags == strictConversion) {
145                 result = sourceIllegal;
146             } else {
147                 *target++ = UNI_REPLACEMENT_CHAR;
148             }
149         } else {
150             /* target is a character in range 0xFFFF - 0x10FFFF. */
151             if (target + 1 >= targetEnd) {
152                 --source; /* Back up source pointer! */
153                 result = targetExhausted; break;
154             }
155             ch -= halfBase;
156             *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
157             *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
158         }
159     }
160     *sourceStart = source;
161     *targetStart = target;
162     return result;
163 }
164 
165 /* --------------------------------------------------------------------- */
166 
ConvertUTF16toUTF32(const UTF16 ** sourceStart,const UTF16 * sourceEnd,UTF32 ** targetStart,UTF32 * targetEnd,ConversionFlags flags)167 ConversionResult ConvertUTF16toUTF32 (
168         const UTF16** sourceStart, const UTF16* sourceEnd,
169         UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags) {
170     ConversionResult result = conversionOK;
171     const UTF16* source = *sourceStart;
172     UTF32* target = *targetStart;
173     UTF32 ch, ch2;
174     while (source < sourceEnd) {
175         const UTF16* oldSource = source; /*  In case we have to back up because of target overflow. */
176         ch = *source++;
177         /* If we have a surrogate pair, convert to UTF32 first. */
178         if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
179             /* If the 16 bits following the high surrogate are in the source buffer... */
180             if (source < sourceEnd) {
181                 ch2 = *source;
182                 /* If it's a low surrogate, convert to UTF32. */
183                 if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
184                     ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
185                         + (ch2 - UNI_SUR_LOW_START) + halfBase;
186                     ++source;
187                 } else if (flags == strictConversion) { /* it's an unpaired high surrogate */
188                     --source; /* return to the illegal value itself */
189                     result = sourceIllegal;
190                     break;
191                 }
192             } else { /* We don't have the 16 bits following the high surrogate. */
193                 --source; /* return to the high surrogate */
194                 result = sourceExhausted;
195                 break;
196             }
197         } else if (flags == strictConversion) {
198             /* UTF-16 surrogate values are illegal in UTF-32 */
199             if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
200                 --source; /* return to the illegal value itself */
201                 result = sourceIllegal;
202                 break;
203             }
204         }
205         if (target >= targetEnd) {
206             source = oldSource; /* Back up source pointer! */
207             result = targetExhausted; break;
208         }
209         *target++ = ch;
210     }
211     *sourceStart = source;
212     *targetStart = target;
213 #ifdef CVTUTF_DEBUG
214 if (result == sourceIllegal) {
215     fprintf(stderr, "ConvertUTF16toUTF32 illegal seq 0x%04x,%04x\n", ch, ch2);
216     fflush(stderr);
217 }
218 #endif
219     return result;
220 }
ConvertUTF16toUTF8(const UTF16 ** sourceStart,const UTF16 * sourceEnd,UTF8 ** targetStart,UTF8 * targetEnd,ConversionFlags flags)221 ConversionResult ConvertUTF16toUTF8 (
222         const UTF16** sourceStart, const UTF16* sourceEnd,
223         UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
224     ConversionResult result = conversionOK;
225     const UTF16* source = *sourceStart;
226     UTF8* target = *targetStart;
227     while (source < sourceEnd) {
228         UTF32 ch;
229         unsigned short bytesToWrite = 0;
230         const UTF32 byteMask = 0xBF;
231         const UTF32 byteMark = 0x80;
232         const UTF16* oldSource = source; /* In case we have to back up because of target overflow. */
233         ch = *source++;
234         /* If we have a surrogate pair, convert to UTF32 first. */
235         if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_HIGH_END) {
236             /* If the 16 bits following the high surrogate are in the source buffer... */
237             if (source < sourceEnd) {
238                 UTF32 ch2 = *source;
239                 /* If it's a low surrogate, convert to UTF32. */
240                 if (ch2 >= UNI_SUR_LOW_START && ch2 <= UNI_SUR_LOW_END) {
241                     ch = ((ch - UNI_SUR_HIGH_START) << halfShift)
242                         + (ch2 - UNI_SUR_LOW_START) + halfBase;
243                     ++source;
244                 } else if (flags == strictConversion) { /* it's an unpaired high surrogate */
245                     --source; /* return to the illegal value itself */
246                     result = sourceIllegal;
247                     break;
248                 }
249             } else { /* We don't have the 16 bits following the high surrogate. */
250                 --source; /* return to the high surrogate */
251                 result = sourceExhausted;
252                 break;
253             }
254         } else if (flags == strictConversion) {
255             /* UTF-16 surrogate values are illegal in UTF-32 */
256             if (ch >= UNI_SUR_LOW_START && ch <= UNI_SUR_LOW_END) {
257                 --source; /* return to the illegal value itself */
258                 result = sourceIllegal;
259                 break;
260             }
261         }
262         /* Figure out how many bytes the result will require */
263         if (ch < (UTF32)0x80) {      bytesToWrite = 1;
264         } else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
265         } else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
266         } else if (ch < (UTF32)0x110000) {  bytesToWrite = 4;
267         } else {                            bytesToWrite = 3;
268                                             ch = UNI_REPLACEMENT_CHAR;
269         }
270 
271         target += bytesToWrite;
272         if (target > targetEnd) {
273             source = oldSource; /* Back up source pointer! */
274             target -= bytesToWrite; result = targetExhausted; break;
275         }
276         switch (bytesToWrite) { /* note: everything falls through. */
277             case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
278             case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
279             case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
280             case 1: *--target =  (UTF8)(ch | firstByteMark[bytesToWrite]);
281         }
282         target += bytesToWrite;
283     }
284     *sourceStart = source;
285     *targetStart = target;
286     return result;
287 }
288 
289 /* --------------------------------------------------------------------- */
290 
ConvertUTF32toUTF8(const UTF32 ** sourceStart,const UTF32 * sourceEnd,UTF8 ** targetStart,UTF8 * targetEnd,ConversionFlags flags)291 ConversionResult ConvertUTF32toUTF8 (
292         const UTF32** sourceStart, const UTF32* sourceEnd,
293         UTF8** targetStart, UTF8* targetEnd, ConversionFlags flags) {
294     ConversionResult result = conversionOK;
295     const UTF32* source = *sourceStart;
296     UTF8* target = *targetStart;
297     while (source < sourceEnd) {
298         UTF32 ch;
299         unsigned short bytesToWrite = 0;
300         const UTF32 byteMask = 0xBF;
301         const UTF32 byteMark = 0x80;
302         ch = *source++;
303         if (flags == strictConversion ) {
304             /* UTF-16 surrogate values are illegal in UTF-32 */
305             if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
306                 --source; /* return to the illegal value itself */
307                 result = sourceIllegal;
308                 break;
309             }
310         }
311         /*
312          * Figure out how many bytes the result will require. Turn any
313          * illegally large UTF32 things (> Plane 17) into replacement chars.
314          */
315         if (ch < (UTF32)0x80) {      bytesToWrite = 1;
316         } else if (ch < (UTF32)0x800) {     bytesToWrite = 2;
317         } else if (ch < (UTF32)0x10000) {   bytesToWrite = 3;
318         } else if (ch <= UNI_MAX_LEGAL_UTF32) {  bytesToWrite = 4;
319         } else {                            bytesToWrite = 3;
320                                             ch = UNI_REPLACEMENT_CHAR;
321                                             result = sourceIllegal;
322         }
323 
324         target += bytesToWrite;
325         if (target > targetEnd) {
326             --source; /* Back up source pointer! */
327             target -= bytesToWrite; result = targetExhausted; break;
328         }
329         switch (bytesToWrite) { /* note: everything falls through. */
330             case 4: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
331             case 3: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
332             case 2: *--target = (UTF8)((ch | byteMark) & byteMask); ch >>= 6;
333             case 1: *--target = (UTF8) (ch | firstByteMark[bytesToWrite]);
334         }
335         target += bytesToWrite;
336     }
337     *sourceStart = source;
338     *targetStart = target;
339     return result;
340 }
341 
342 /* --------------------------------------------------------------------- */
343 
344 /*
345  * Utility routine to tell whether a sequence of bytes is legal UTF-8.
346  * This must be called with the length pre-determined by the first byte.
347  * If not calling this from ConvertUTF8to*, then the length can be set by:
348  *  length = trailingBytesForUTF8[*source]+1;
349  * and the sequence is illegal right away if there aren't that many bytes
350  * available.
351  * If presented with a length > 4, this returns false.  The Unicode
352  * definition of UTF-8 goes up to 4-byte sequences.
353  */
354 
isLegalUTF8(const UTF8 * source,int length)355 static Boolean isLegalUTF8(const UTF8 *source, int length) {
356     UTF8 a;
357     const UTF8 *srcptr = source+length;
358     switch (length) {
359     default: return false;
360         /* Everything else falls through when "true"... */
361     case 4: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
362     case 3: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
363     case 2: if ((a = (*--srcptr)) < 0x80 || a > 0xBF) return false;
364 
365         switch (*source) {
366             /* no fall-through in this inner switch */
367             case 0xE0: if (a < 0xA0) return false; break;
368             case 0xED: if (a > 0x9F) return false; break;
369             case 0xF0: if (a < 0x90) return false; break;
370             case 0xF4: if (a > 0x8F) return false; break;
371             default:   if (a < 0x80) return false;
372         }
373 
374     case 1: if (*source >= 0x80 && *source < 0xC2) return false;
375     }
376     if (*source > 0xF4) return false;
377     return true;
378 }
379 
380 /* --------------------------------------------------------------------- */
381 
382 /*
383  * Exported function to return whether a UTF-8 sequence is legal or not.
384  * This is not used here; it's just exported.
385  */
isLegalUTF8Sequence(const UTF8 * source,const UTF8 * sourceEnd)386 Boolean isLegalUTF8Sequence(const UTF8 *source, const UTF8 *sourceEnd) {
387     int length = trailingBytesForUTF8[*source]+1;
388     if (length > sourceEnd - source) {
389         return false;
390     }
391     return isLegalUTF8(source, length);
392 }
393 
394 /* --------------------------------------------------------------------- */
395 
396 static unsigned
findMaximalSubpartOfIllFormedUTF8Sequence(const UTF8 * source,const UTF8 * sourceEnd)397 findMaximalSubpartOfIllFormedUTF8Sequence(const UTF8 *source,
398                                           const UTF8 *sourceEnd) {
399   UTF8 b1, b2, b3;
400 
401   assert(!isLegalUTF8Sequence(source, sourceEnd));
402 
403   /*
404    * Unicode 6.3.0, D93b:
405    *
406    *   Maximal subpart of an ill-formed subsequence: The longest code unit
407    *   subsequence starting at an unconvertible offset that is either:
408    *   a. the initial subsequence of a well-formed code unit sequence, or
409    *   b. a subsequence of length one.
410    */
411 
412   if (source == sourceEnd)
413     return 0;
414 
415   /*
416    * Perform case analysis.  See Unicode 6.3.0, Table 3-7. Well-Formed UTF-8
417    * Byte Sequences.
418    */
419 
420   b1 = *source;
421   ++source;
422   if (b1 >= 0xC2 && b1 <= 0xDF) {
423     /*
424      * First byte is valid, but we know that this code unit sequence is
425      * invalid, so the maximal subpart has to end after the first byte.
426      */
427     return 1;
428   }
429 
430   if (source == sourceEnd)
431     return 1;
432 
433   b2 = *source;
434   ++source;
435 
436   if (b1 == 0xE0) {
437     return (b2 >= 0xA0 && b2 <= 0xBF) ? 2 : 1;
438   }
439   if (b1 >= 0xE1 && b1 <= 0xEC) {
440     return (b2 >= 0x80 && b2 <= 0xBF) ? 2 : 1;
441   }
442   if (b1 == 0xED) {
443     return (b2 >= 0x80 && b2 <= 0x9F) ? 2 : 1;
444   }
445   if (b1 >= 0xEE && b1 <= 0xEF) {
446     return (b2 >= 0x80 && b2 <= 0xBF) ? 2 : 1;
447   }
448   if (b1 == 0xF0) {
449     if (b2 >= 0x90 && b2 <= 0xBF) {
450       if (source == sourceEnd)
451         return 2;
452 
453       b3 = *source;
454       return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
455     }
456     return 1;
457   }
458   if (b1 >= 0xF1 && b1 <= 0xF3) {
459     if (b2 >= 0x80 && b2 <= 0xBF) {
460       if (source == sourceEnd)
461         return 2;
462 
463       b3 = *source;
464       return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
465     }
466     return 1;
467   }
468   if (b1 == 0xF4) {
469     if (b2 >= 0x80 && b2 <= 0x8F) {
470       if (source == sourceEnd)
471         return 2;
472 
473       b3 = *source;
474       return (b3 >= 0x80 && b3 <= 0xBF) ? 3 : 2;
475     }
476     return 1;
477   }
478 
479   assert((b1 >= 0x80 && b1 <= 0xC1) || b1 >= 0xF5);
480   /*
481    * There are no valid sequences that start with these bytes.  Maximal subpart
482    * is defined to have length 1 in these cases.
483    */
484   return 1;
485 }
486 
487 /* --------------------------------------------------------------------- */
488 
489 /*
490  * Exported function to return the total number of bytes in a codepoint
491  * represented in UTF-8, given the value of the first byte.
492  */
getNumBytesForUTF8(UTF8 first)493 unsigned getNumBytesForUTF8(UTF8 first) {
494   return trailingBytesForUTF8[first] + 1;
495 }
496 
497 /* --------------------------------------------------------------------- */
498 
499 /*
500  * Exported function to return whether a UTF-8 string is legal or not.
501  * This is not used here; it's just exported.
502  */
isLegalUTF8String(const UTF8 ** source,const UTF8 * sourceEnd)503 Boolean isLegalUTF8String(const UTF8 **source, const UTF8 *sourceEnd) {
504     while (*source != sourceEnd) {
505         int length = trailingBytesForUTF8[**source] + 1;
506         if (length > sourceEnd - *source || !isLegalUTF8(*source, length))
507             return false;
508         *source += length;
509     }
510     return true;
511 }
512 
513 /* --------------------------------------------------------------------- */
514 
ConvertUTF8toUTF16(const UTF8 ** sourceStart,const UTF8 * sourceEnd,UTF16 ** targetStart,UTF16 * targetEnd,ConversionFlags flags)515 ConversionResult ConvertUTF8toUTF16 (
516         const UTF8** sourceStart, const UTF8* sourceEnd,
517         UTF16** targetStart, UTF16* targetEnd, ConversionFlags flags) {
518     ConversionResult result = conversionOK;
519     const UTF8* source = *sourceStart;
520     UTF16* target = *targetStart;
521     while (source < sourceEnd) {
522         UTF32 ch = 0;
523         unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
524         if (extraBytesToRead >= sourceEnd - source) {
525             result = sourceExhausted; break;
526         }
527         /* Do this check whether lenient or strict */
528         if (!isLegalUTF8(source, extraBytesToRead+1)) {
529             result = sourceIllegal;
530             break;
531         }
532         /*
533          * The cases all fall through. See "Note A" below.
534          */
535         switch (extraBytesToRead) {
536             case 5: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
537             case 4: ch += *source++; ch <<= 6; /* remember, illegal UTF-8 */
538             case 3: ch += *source++; ch <<= 6;
539             case 2: ch += *source++; ch <<= 6;
540             case 1: ch += *source++; ch <<= 6;
541             case 0: ch += *source++;
542         }
543         ch -= offsetsFromUTF8[extraBytesToRead];
544 
545         if (target >= targetEnd) {
546             source -= (extraBytesToRead+1); /* Back up source pointer! */
547             result = targetExhausted; break;
548         }
549         if (ch <= UNI_MAX_BMP) { /* Target is a character <= 0xFFFF */
550             /* UTF-16 surrogate values are illegal in UTF-32 */
551             if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
552                 if (flags == strictConversion) {
553                     source -= (extraBytesToRead+1); /* return to the illegal value itself */
554                     result = sourceIllegal;
555                     break;
556                 } else {
557                     *target++ = UNI_REPLACEMENT_CHAR;
558                 }
559             } else {
560                 *target++ = (UTF16)ch; /* normal case */
561             }
562         } else if (ch > UNI_MAX_UTF16) {
563             if (flags == strictConversion) {
564                 result = sourceIllegal;
565                 source -= (extraBytesToRead+1); /* return to the start */
566                 break; /* Bail out; shouldn't continue */
567             } else {
568                 *target++ = UNI_REPLACEMENT_CHAR;
569             }
570         } else {
571             /* target is a character in range 0xFFFF - 0x10FFFF. */
572             if (target + 1 >= targetEnd) {
573                 source -= (extraBytesToRead+1); /* Back up source pointer! */
574                 result = targetExhausted; break;
575             }
576             ch -= halfBase;
577             *target++ = (UTF16)((ch >> halfShift) + UNI_SUR_HIGH_START);
578             *target++ = (UTF16)((ch & halfMask) + UNI_SUR_LOW_START);
579         }
580     }
581     *sourceStart = source;
582     *targetStart = target;
583     return result;
584 }
585 
586 /* --------------------------------------------------------------------- */
587 
ConvertUTF8toUTF32Impl(const UTF8 ** sourceStart,const UTF8 * sourceEnd,UTF32 ** targetStart,UTF32 * targetEnd,ConversionFlags flags,Boolean InputIsPartial)588 static ConversionResult ConvertUTF8toUTF32Impl(
589         const UTF8** sourceStart, const UTF8* sourceEnd,
590         UTF32** targetStart, UTF32* targetEnd, ConversionFlags flags,
591         Boolean InputIsPartial) {
592     ConversionResult result = conversionOK;
593     const UTF8* source = *sourceStart;
594     UTF32* target = *targetStart;
595     while (source < sourceEnd) {
596         UTF32 ch = 0;
597         unsigned short extraBytesToRead = trailingBytesForUTF8[*source];
598         if (extraBytesToRead >= sourceEnd - source) {
599             if (flags == strictConversion || InputIsPartial) {
600                 result = sourceExhausted;
601                 break;
602             } else {
603                 result = sourceIllegal;
604 
605                 /*
606                  * Replace the maximal subpart of ill-formed sequence with
607                  * replacement character.
608                  */
609                 source += findMaximalSubpartOfIllFormedUTF8Sequence(source,
610                                                                     sourceEnd);
611                 *target++ = UNI_REPLACEMENT_CHAR;
612                 continue;
613             }
614         }
615         if (target >= targetEnd) {
616             result = targetExhausted; break;
617         }
618 
619         /* Do this check whether lenient or strict */
620         if (!isLegalUTF8(source, extraBytesToRead+1)) {
621             result = sourceIllegal;
622             if (flags == strictConversion) {
623                 /* Abort conversion. */
624                 break;
625             } else {
626                 /*
627                  * Replace the maximal subpart of ill-formed sequence with
628                  * replacement character.
629                  */
630                 source += findMaximalSubpartOfIllFormedUTF8Sequence(source,
631                                                                     sourceEnd);
632                 *target++ = UNI_REPLACEMENT_CHAR;
633                 continue;
634             }
635         }
636         /*
637          * The cases all fall through. See "Note A" below.
638          */
639         switch (extraBytesToRead) {
640             case 5: ch += *source++; ch <<= 6;
641             case 4: ch += *source++; ch <<= 6;
642             case 3: ch += *source++; ch <<= 6;
643             case 2: ch += *source++; ch <<= 6;
644             case 1: ch += *source++; ch <<= 6;
645             case 0: ch += *source++;
646         }
647         ch -= offsetsFromUTF8[extraBytesToRead];
648 
649         if (ch <= UNI_MAX_LEGAL_UTF32) {
650             /*
651              * UTF-16 surrogate values are illegal in UTF-32, and anything
652              * over Plane 17 (> 0x10FFFF) is illegal.
653              */
654             if (ch >= UNI_SUR_HIGH_START && ch <= UNI_SUR_LOW_END) {
655                 if (flags == strictConversion) {
656                     source -= (extraBytesToRead+1); /* return to the illegal value itself */
657                     result = sourceIllegal;
658                     break;
659                 } else {
660                     *target++ = UNI_REPLACEMENT_CHAR;
661                 }
662             } else {
663                 *target++ = ch;
664             }
665         } else { /* i.e., ch > UNI_MAX_LEGAL_UTF32 */
666             result = sourceIllegal;
667             *target++ = UNI_REPLACEMENT_CHAR;
668         }
669     }
670     *sourceStart = source;
671     *targetStart = target;
672     return result;
673 }
674 
ConvertUTF8toUTF32Partial(const UTF8 ** sourceStart,const UTF8 * sourceEnd,UTF32 ** targetStart,UTF32 * targetEnd,ConversionFlags flags)675 ConversionResult ConvertUTF8toUTF32Partial(const UTF8 **sourceStart,
676                                            const UTF8 *sourceEnd,
677                                            UTF32 **targetStart,
678                                            UTF32 *targetEnd,
679                                            ConversionFlags flags) {
680   return ConvertUTF8toUTF32Impl(sourceStart, sourceEnd, targetStart, targetEnd,
681                                 flags, /*InputIsPartial=*/true);
682 }
683 
ConvertUTF8toUTF32(const UTF8 ** sourceStart,const UTF8 * sourceEnd,UTF32 ** targetStart,UTF32 * targetEnd,ConversionFlags flags)684 ConversionResult ConvertUTF8toUTF32(const UTF8 **sourceStart,
685                                     const UTF8 *sourceEnd, UTF32 **targetStart,
686                                     UTF32 *targetEnd, ConversionFlags flags) {
687   return ConvertUTF8toUTF32Impl(sourceStart, sourceEnd, targetStart, targetEnd,
688                                 flags, /*InputIsPartial=*/false);
689 }
690 
691 /* ---------------------------------------------------------------------
692 
693     Note A.
694     The fall-through switches in UTF-8 reading code save a
695     temp variable, some decrements & conditionals.  The switches
696     are equivalent to the following loop:
697         {
698             int tmpBytesToRead = extraBytesToRead+1;
699             do {
700                 ch += *source++;
701                 --tmpBytesToRead;
702                 if (tmpBytesToRead) ch <<= 6;
703             } while (tmpBytesToRead > 0);
704         }
705     In UTF-8 writing code, the switches on "bytesToWrite" are
706     similarly unrolled loops.
707 
708    --------------------------------------------------------------------- */
709