1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | #include "clang/Lex/LiteralSupport.h" |
15 | #include "clang/Basic/CharInfo.h" |
16 | #include "clang/Basic/LangOptions.h" |
17 | #include "clang/Basic/SourceLocation.h" |
18 | #include "clang/Basic/TargetInfo.h" |
19 | #include "clang/Lex/LexDiagnostic.h" |
20 | #include "clang/Lex/Lexer.h" |
21 | #include "clang/Lex/Preprocessor.h" |
22 | #include "clang/Lex/Token.h" |
23 | #include "llvm/ADT/APInt.h" |
24 | #include "llvm/ADT/SmallVector.h" |
25 | #include "llvm/ADT/StringExtras.h" |
26 | #include "llvm/ADT/StringSwitch.h" |
27 | #include "llvm/Support/ConvertUTF.h" |
28 | #include "llvm/Support/ErrorHandling.h" |
29 | #include <algorithm> |
30 | #include <cassert> |
31 | #include <cstddef> |
32 | #include <cstdint> |
33 | #include <cstring> |
34 | #include <string> |
35 | |
36 | using namespace clang; |
37 | |
38 | static unsigned getCharWidth(tok::TokenKind kind, const TargetInfo &Target) { |
39 | switch (kind) { |
40 | default: llvm_unreachable("Unknown token type!"); |
41 | case tok::char_constant: |
42 | case tok::string_literal: |
43 | case tok::utf8_char_constant: |
44 | case tok::utf8_string_literal: |
45 | return Target.getCharWidth(); |
46 | case tok::wide_char_constant: |
47 | case tok::wide_string_literal: |
48 | return Target.getWCharWidth(); |
49 | case tok::utf16_char_constant: |
50 | case tok::utf16_string_literal: |
51 | return Target.getChar16Width(); |
52 | case tok::utf32_char_constant: |
53 | case tok::utf32_string_literal: |
54 | return Target.getChar32Width(); |
55 | } |
56 | } |
57 | |
58 | static CharSourceRange MakeCharSourceRange(const LangOptions &Features, |
59 | FullSourceLoc TokLoc, |
60 | const char *TokBegin, |
61 | const char *TokRangeBegin, |
62 | const char *TokRangeEnd) { |
63 | SourceLocation Begin = |
64 | Lexer::AdvanceToTokenCharacter(TokLoc, TokRangeBegin - TokBegin, |
65 | TokLoc.getManager(), Features); |
66 | SourceLocation End = |
67 | Lexer::AdvanceToTokenCharacter(Begin, TokRangeEnd - TokRangeBegin, |
68 | TokLoc.getManager(), Features); |
69 | return CharSourceRange::getCharRange(Begin, End); |
70 | } |
71 | |
72 | |
73 | |
74 | |
75 | |
76 | |
77 | static DiagnosticBuilder Diag(DiagnosticsEngine *Diags, |
78 | const LangOptions &Features, FullSourceLoc TokLoc, |
79 | const char *TokBegin, const char *TokRangeBegin, |
80 | const char *TokRangeEnd, unsigned DiagID) { |
81 | SourceLocation Begin = |
82 | Lexer::AdvanceToTokenCharacter(TokLoc, TokRangeBegin - TokBegin, |
83 | TokLoc.getManager(), Features); |
84 | return Diags->Report(Begin, DiagID) << |
85 | MakeCharSourceRange(Features, TokLoc, TokBegin, TokRangeBegin, TokRangeEnd); |
86 | } |
87 | |
88 | |
89 | |
90 | static unsigned ProcessCharEscape(const char *ThisTokBegin, |
91 | const char *&ThisTokBuf, |
92 | const char *ThisTokEnd, bool &HadError, |
93 | FullSourceLoc Loc, unsigned CharWidth, |
94 | DiagnosticsEngine *Diags, |
95 | const LangOptions &Features) { |
96 | const char *EscapeBegin = ThisTokBuf; |
97 | |
98 | |
99 | ++ThisTokBuf; |
100 | |
101 | |
102 | |
103 | unsigned ResultChar = *ThisTokBuf++; |
104 | switch (ResultChar) { |
105 | |
106 | case '\\': case '\'': case '"': case '?': break; |
107 | |
108 | |
109 | case 'a': |
110 | |
111 | ResultChar = 7; |
112 | break; |
113 | case 'b': |
114 | ResultChar = 8; |
115 | break; |
116 | case 'e': |
117 | if (Diags) |
118 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
119 | diag::ext_nonstandard_escape) << "e"; |
120 | ResultChar = 27; |
121 | break; |
122 | case 'E': |
123 | if (Diags) |
124 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
125 | diag::ext_nonstandard_escape) << "E"; |
126 | ResultChar = 27; |
127 | break; |
128 | case 'f': |
129 | ResultChar = 12; |
130 | break; |
131 | case 'n': |
132 | ResultChar = 10; |
133 | break; |
134 | case 'r': |
135 | ResultChar = 13; |
136 | break; |
137 | case 't': |
138 | ResultChar = 9; |
139 | break; |
140 | case 'v': |
141 | ResultChar = 11; |
142 | break; |
143 | case 'x': { |
144 | ResultChar = 0; |
145 | if (ThisTokBuf == ThisTokEnd || !isHexDigit(*ThisTokBuf)) { |
146 | if (Diags) |
147 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
148 | diag::err_hex_escape_no_digits) << "x"; |
149 | HadError = true; |
150 | break; |
151 | } |
152 | |
153 | |
154 | bool Overflow = false; |
155 | for (; ThisTokBuf != ThisTokEnd; ++ThisTokBuf) { |
156 | int CharVal = llvm::hexDigitValue(ThisTokBuf[0]); |
157 | if (CharVal == -1) break; |
158 | |
159 | if (ResultChar & 0xF0000000) |
160 | Overflow = true; |
161 | ResultChar <<= 4; |
162 | ResultChar |= CharVal; |
163 | } |
164 | |
165 | |
166 | if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) { |
167 | Overflow = true; |
168 | ResultChar &= ~0U >> (32-CharWidth); |
169 | } |
170 | |
171 | |
172 | if (Overflow && Diags) |
173 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
174 | diag::err_escape_too_large) << 0; |
175 | break; |
176 | } |
177 | case '0': case '1': case '2': case '3': |
178 | case '4': case '5': case '6': case '7': { |
179 | |
180 | --ThisTokBuf; |
181 | ResultChar = 0; |
182 | |
183 | |
184 | |
185 | unsigned NumDigits = 0; |
186 | do { |
187 | ResultChar <<= 3; |
188 | ResultChar |= *ThisTokBuf++ - '0'; |
189 | ++NumDigits; |
190 | } while (ThisTokBuf != ThisTokEnd && NumDigits < 3 && |
191 | ThisTokBuf[0] >= '0' && ThisTokBuf[0] <= '7'); |
192 | |
193 | |
194 | if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) { |
195 | if (Diags) |
196 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
197 | diag::err_escape_too_large) << 1; |
198 | ResultChar &= ~0U >> (32-CharWidth); |
199 | } |
200 | break; |
201 | } |
202 | |
203 | |
204 | case '(': case '{': case '[': case '%': |
205 | |
206 | if (Diags) |
207 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
208 | diag::ext_nonstandard_escape) |
209 | << std::string(1, ResultChar); |
210 | break; |
211 | default: |
212 | if (!Diags) |
213 | break; |
214 | |
215 | if (isPrintable(ResultChar)) |
216 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
217 | diag::ext_unknown_escape) |
218 | << std::string(1, ResultChar); |
219 | else |
220 | Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, |
221 | diag::ext_unknown_escape) |
222 | << "x" + llvm::utohexstr(ResultChar); |
223 | break; |
224 | } |
225 | |
226 | return ResultChar; |
227 | } |
228 | |
229 | static void appendCodePoint(unsigned Codepoint, |
230 | llvm::SmallVectorImpl<char> &Str) { |
231 | char ResultBuf[4]; |
232 | char *ResultPtr = ResultBuf; |
233 | bool Res = llvm::ConvertCodePointToUTF8(Codepoint, ResultPtr); |
234 | (void)Res; |
235 | (0) . __assert_fail ("Res && \"Unexpected conversion failure\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 235, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(Res && "Unexpected conversion failure"); |
236 | Str.append(ResultBuf, ResultPtr); |
237 | } |
238 | |
239 | void clang::expandUCNs(SmallVectorImpl<char> &Buf, StringRef Input) { |
240 | for (StringRef::iterator I = Input.begin(), E = Input.end(); I != E; ++I) { |
241 | if (*I != '\\') { |
242 | Buf.push_back(*I); |
243 | continue; |
244 | } |
245 | |
246 | ++I; |
247 | assert(*I == 'u' || *I == 'U'); |
248 | |
249 | unsigned NumHexDigits; |
250 | if (*I == 'u') |
251 | NumHexDigits = 4; |
252 | else |
253 | NumHexDigits = 8; |
254 | |
255 | assert(I + NumHexDigits <= E); |
256 | |
257 | uint32_t CodePoint = 0; |
258 | for (++I; NumHexDigits != 0; ++I, --NumHexDigits) { |
259 | unsigned Value = llvm::hexDigitValue(*I); |
260 | assert(Value != -1U); |
261 | |
262 | CodePoint <<= 4; |
263 | CodePoint += Value; |
264 | } |
265 | |
266 | appendCodePoint(CodePoint, Buf); |
267 | --I; |
268 | } |
269 | } |
270 | |
271 | |
272 | |
273 | static bool ProcessUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, |
274 | const char *ThisTokEnd, |
275 | uint32_t &UcnVal, unsigned short &UcnLen, |
276 | FullSourceLoc Loc, DiagnosticsEngine *Diags, |
277 | const LangOptions &Features, |
278 | bool in_char_string_literal = false) { |
279 | const char *UcnBegin = ThisTokBuf; |
280 | |
281 | |
282 | ThisTokBuf += 2; |
283 | |
284 | if (ThisTokBuf == ThisTokEnd || !isHexDigit(*ThisTokBuf)) { |
285 | if (Diags) |
286 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
287 | diag::err_hex_escape_no_digits) << StringRef(&ThisTokBuf[-1], 1); |
288 | return false; |
289 | } |
290 | UcnLen = (ThisTokBuf[-1] == 'u' ? 4 : 8); |
291 | unsigned short UcnLenSave = UcnLen; |
292 | for (; ThisTokBuf != ThisTokEnd && UcnLenSave; ++ThisTokBuf, UcnLenSave--) { |
293 | int CharVal = llvm::hexDigitValue(ThisTokBuf[0]); |
294 | if (CharVal == -1) break; |
295 | UcnVal <<= 4; |
296 | UcnVal |= CharVal; |
297 | } |
298 | |
299 | if (UcnLenSave) { |
300 | if (Diags) |
301 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
302 | diag::err_ucn_escape_incomplete); |
303 | return false; |
304 | } |
305 | |
306 | |
307 | if ((0xD800 <= UcnVal && UcnVal <= 0xDFFF) || |
308 | UcnVal > 0x10FFFF) { |
309 | if (Diags) |
310 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
311 | diag::err_ucn_escape_invalid); |
312 | return false; |
313 | } |
314 | |
315 | |
316 | |
317 | if (UcnVal < 0xa0 && |
318 | (UcnVal != 0x24 && UcnVal != 0x40 && UcnVal != 0x60)) { |
319 | bool IsError = (!Features.CPlusPlus11 || !in_char_string_literal); |
320 | if (Diags) { |
321 | char BasicSCSChar = UcnVal; |
322 | if (UcnVal >= 0x20 && UcnVal < 0x7f) |
323 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
324 | IsError ? diag::err_ucn_escape_basic_scs : |
325 | diag::warn_cxx98_compat_literal_ucn_escape_basic_scs) |
326 | << StringRef(&BasicSCSChar, 1); |
327 | else |
328 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
329 | IsError ? diag::err_ucn_control_character : |
330 | diag::warn_cxx98_compat_literal_ucn_control_character); |
331 | } |
332 | if (IsError) |
333 | return false; |
334 | } |
335 | |
336 | if (!Features.CPlusPlus && !Features.C99 && Diags) |
337 | Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf, |
338 | diag::warn_ucn_not_valid_in_c89_literal); |
339 | |
340 | return true; |
341 | } |
342 | |
343 | |
344 | |
345 | static int MeasureUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, |
346 | const char *ThisTokEnd, unsigned CharByteWidth, |
347 | const LangOptions &Features, bool &HadError) { |
348 | |
349 | if (CharByteWidth == 4) |
350 | return 4; |
351 | |
352 | uint32_t UcnVal = 0; |
353 | unsigned short UcnLen = 0; |
354 | FullSourceLoc Loc; |
355 | |
356 | if (!ProcessUCNEscape(ThisTokBegin, ThisTokBuf, ThisTokEnd, UcnVal, |
357 | UcnLen, Loc, nullptr, Features, true)) { |
358 | HadError = true; |
359 | return 0; |
360 | } |
361 | |
362 | |
363 | if (CharByteWidth == 2) |
364 | return UcnVal <= 0xFFFF ? 2 : 4; |
365 | |
366 | |
367 | if (UcnVal < 0x80) |
368 | return 1; |
369 | if (UcnVal < 0x800) |
370 | return 2; |
371 | if (UcnVal < 0x10000) |
372 | return 3; |
373 | return 4; |
374 | } |
375 | |
376 | |
377 | |
378 | |
379 | |
380 | static void EncodeUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, |
381 | const char *ThisTokEnd, |
382 | char *&ResultBuf, bool &HadError, |
383 | FullSourceLoc Loc, unsigned CharByteWidth, |
384 | DiagnosticsEngine *Diags, |
385 | const LangOptions &Features) { |
386 | typedef uint32_t UTF32; |
387 | UTF32 UcnVal = 0; |
388 | unsigned short UcnLen = 0; |
389 | if (!ProcessUCNEscape(ThisTokBegin, ThisTokBuf, ThisTokEnd, UcnVal, UcnLen, |
390 | Loc, Diags, Features, true)) { |
391 | HadError = true; |
392 | return; |
393 | } |
394 | |
395 | (0) . __assert_fail ("(CharByteWidth == 1 || CharByteWidth == 2 || CharByteWidth == 4) && \"only character widths of 1, 2, or 4 bytes supported\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 396, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert((CharByteWidth == 1 || CharByteWidth == 2 || CharByteWidth == 4) && |
396 | (0) . __assert_fail ("(CharByteWidth == 1 || CharByteWidth == 2 || CharByteWidth == 4) && \"only character widths of 1, 2, or 4 bytes supported\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 396, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> "only character widths of 1, 2, or 4 bytes supported"); |
397 | |
398 | (void)UcnLen; |
399 | (0) . __assert_fail ("(UcnLen== 4 || UcnLen== 8) && \"only ucn length of 4 or 8 supported\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 399, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert((UcnLen== 4 || UcnLen== 8) && "only ucn length of 4 or 8 supported"); |
400 | |
401 | if (CharByteWidth == 4) { |
402 | |
403 | |
404 | llvm::UTF32 *ResultPtr = reinterpret_cast<llvm::UTF32*>(ResultBuf); |
405 | *ResultPtr = UcnVal; |
406 | ResultBuf += 4; |
407 | return; |
408 | } |
409 | |
410 | if (CharByteWidth == 2) { |
411 | |
412 | |
413 | llvm::UTF16 *ResultPtr = reinterpret_cast<llvm::UTF16*>(ResultBuf); |
414 | |
415 | if (UcnVal <= (UTF32)0xFFFF) { |
416 | *ResultPtr = UcnVal; |
417 | ResultBuf += 2; |
418 | return; |
419 | } |
420 | |
421 | |
422 | UcnVal -= 0x10000; |
423 | *ResultPtr = 0xD800 + (UcnVal >> 10); |
424 | *(ResultPtr+1) = 0xDC00 + (UcnVal & 0x3FF); |
425 | ResultBuf += 4; |
426 | return; |
427 | } |
428 | |
429 | (0) . __assert_fail ("CharByteWidth == 1 && \"UTF-8 encoding is only for 1 byte characters\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 429, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(CharByteWidth == 1 && "UTF-8 encoding is only for 1 byte characters"); |
430 | |
431 | |
432 | |
433 | |
434 | |
435 | typedef uint8_t UTF8; |
436 | |
437 | unsigned short bytesToWrite = 0; |
438 | if (UcnVal < (UTF32)0x80) |
439 | bytesToWrite = 1; |
440 | else if (UcnVal < (UTF32)0x800) |
441 | bytesToWrite = 2; |
442 | else if (UcnVal < (UTF32)0x10000) |
443 | bytesToWrite = 3; |
444 | else |
445 | bytesToWrite = 4; |
446 | |
447 | const unsigned byteMask = 0xBF; |
448 | const unsigned byteMark = 0x80; |
449 | |
450 | |
451 | |
452 | static const UTF8 firstByteMark[5] = { |
453 | 0x00, 0x00, 0xC0, 0xE0, 0xF0 |
454 | }; |
455 | |
456 | ResultBuf += bytesToWrite; |
457 | switch (bytesToWrite) { |
458 | case 4: |
459 | *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6; |
460 | LLVM_FALLTHROUGH; |
461 | case 3: |
462 | *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6; |
463 | LLVM_FALLTHROUGH; |
464 | case 2: |
465 | *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6; |
466 | LLVM_FALLTHROUGH; |
467 | case 1: |
468 | *--ResultBuf = (UTF8) (UcnVal | firstByteMark[bytesToWrite]); |
469 | } |
470 | |
471 | ResultBuf += bytesToWrite; |
472 | } |
473 | |
474 | |
475 | |
476 | |
477 | |
478 | |
479 | |
480 | |
481 | |
482 | |
483 | |
484 | |
485 | |
486 | |
487 | |
488 | |
489 | |
490 | |
491 | |
492 | |
493 | |
494 | |
495 | |
496 | |
497 | |
498 | |
499 | |
500 | |
501 | |
502 | |
503 | |
504 | |
505 | |
506 | |
507 | |
508 | |
509 | |
510 | |
511 | |
512 | |
513 | |
514 | |
515 | |
516 | |
517 | |
518 | |
519 | |
520 | |
521 | |
522 | |
523 | |
524 | |
525 | NumericLiteralParser::NumericLiteralParser(StringRef TokSpelling, |
526 | SourceLocation TokLoc, |
527 | Preprocessor &PP) |
528 | : PP(PP), ThisTokBegin(TokSpelling.begin()), ThisTokEnd(TokSpelling.end()) { |
529 | |
530 | |
531 | |
532 | |
533 | |
534 | (0) . __assert_fail ("!isPreprocessingNumberBody(*ThisTokEnd) && \"didn't maximally munch?\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 534, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(!isPreprocessingNumberBody(*ThisTokEnd) && "didn't maximally munch?"); |
535 | |
536 | s = DigitsBegin = ThisTokBegin; |
537 | saw_exponent = false; |
538 | saw_period = false; |
539 | saw_ud_suffix = false; |
540 | saw_fixed_point_suffix = false; |
541 | isLong = false; |
542 | isUnsigned = false; |
543 | isLongLong = false; |
544 | isHalf = false; |
545 | isFloat = false; |
546 | isImaginary = false; |
547 | isFloat16 = false; |
548 | isFloat128 = false; |
549 | MicrosoftInteger = 0; |
550 | isFract = false; |
551 | isAccum = false; |
552 | hadError = false; |
553 | |
554 | if (*s == '0') { |
555 | ParseNumberStartingWithZero(TokLoc); |
556 | if (hadError) |
557 | return; |
558 | } else { |
559 | radix = 10; |
560 | s = SkipDigits(s); |
561 | if (s == ThisTokEnd) { |
562 | |
563 | } else { |
564 | ParseDecimalOrOctalCommon(TokLoc); |
565 | if (hadError) |
566 | return; |
567 | } |
568 | } |
569 | |
570 | SuffixBegin = s; |
571 | checkSeparator(TokLoc, s, CSK_AfterDigits); |
572 | |
573 | |
574 | if (PP.getLangOpts().FixedPoint) { |
575 | for (const char *c = s; c != ThisTokEnd; ++c) { |
576 | if (*c == 'r' || *c == 'k' || *c == 'R' || *c == 'K') { |
577 | saw_fixed_point_suffix = true; |
578 | break; |
579 | } |
580 | } |
581 | } |
582 | |
583 | |
584 | |
585 | bool isFPConstant = isFloatingLiteral(); |
586 | |
587 | |
588 | |
589 | for (; s != ThisTokEnd; ++s) { |
590 | switch (*s) { |
591 | case 'R': |
592 | case 'r': |
593 | if (!PP.getLangOpts().FixedPoint) break; |
594 | if (isFract || isAccum) break; |
595 | if (!(saw_period || saw_exponent)) break; |
596 | isFract = true; |
597 | continue; |
598 | case 'K': |
599 | case 'k': |
600 | if (!PP.getLangOpts().FixedPoint) break; |
601 | if (isFract || isAccum) break; |
602 | if (!(saw_period || saw_exponent)) break; |
603 | isAccum = true; |
604 | continue; |
605 | case 'h': |
606 | case 'H': |
607 | |
608 | if (!(PP.getLangOpts().Half || PP.getLangOpts().FixedPoint)) break; |
609 | if (isIntegerLiteral()) break; |
610 | if (isHalf || isFloat || isLong) break; |
611 | isHalf = true; |
612 | continue; |
613 | case 'f': |
614 | case 'F': |
615 | if (!isFPConstant) break; |
616 | if (isHalf || isFloat || isLong || isFloat128) |
617 | break; |
618 | |
619 | |
620 | |
621 | |
622 | if ((PP.getTargetInfo().hasFloat16Type() || PP.getLangOpts().CUDA) && |
623 | s + 2 < ThisTokEnd && s[1] == '1' && s[2] == '6') { |
624 | s += 2; |
625 | isFloat16 = true; |
626 | continue; |
627 | } |
628 | |
629 | isFloat = true; |
630 | continue; |
631 | case 'q': |
632 | case 'Q': |
633 | if (!isFPConstant) break; |
634 | if (isHalf || isFloat || isLong || isFloat128) |
635 | break; |
636 | isFloat128 = true; |
637 | continue; |
638 | case 'u': |
639 | case 'U': |
640 | if (isFPConstant) break; |
641 | if (isUnsigned) break; |
642 | isUnsigned = true; |
643 | continue; |
644 | case 'l': |
645 | case 'L': |
646 | if (isLong || isLongLong) break; |
647 | if (isHalf || isFloat || isFloat128) break; |
648 | |
649 | |
650 | if (s[1] == s[0]) { |
651 | (0) . __assert_fail ("s + 1 < ThisTokEnd && \"didn't maximally munch?\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 651, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(s + 1 < ThisTokEnd && "didn't maximally munch?"); |
652 | if (isFPConstant) break; |
653 | isLongLong = true; |
654 | ++s; |
655 | } else { |
656 | isLong = true; |
657 | } |
658 | continue; |
659 | case 'i': |
660 | case 'I': |
661 | if (PP.getLangOpts().MicrosoftExt) { |
662 | if (isLong || isLongLong || MicrosoftInteger) |
663 | break; |
664 | |
665 | if (!isFPConstant) { |
666 | |
667 | switch (s[1]) { |
668 | case '8': |
669 | s += 2; |
670 | MicrosoftInteger = 8; |
671 | break; |
672 | case '1': |
673 | if (s[2] == '6') { |
674 | s += 3; |
675 | MicrosoftInteger = 16; |
676 | } |
677 | break; |
678 | case '3': |
679 | if (s[2] == '2') { |
680 | s += 3; |
681 | MicrosoftInteger = 32; |
682 | } |
683 | break; |
684 | case '6': |
685 | if (s[2] == '4') { |
686 | s += 3; |
687 | MicrosoftInteger = 64; |
688 | } |
689 | break; |
690 | default: |
691 | break; |
692 | } |
693 | } |
694 | if (MicrosoftInteger) { |
695 | (0) . __assert_fail ("s <= ThisTokEnd && \"didn't maximally munch?\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 695, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(s <= ThisTokEnd && "didn't maximally munch?"); |
696 | break; |
697 | } |
698 | } |
699 | LLVM_FALLTHROUGH; |
700 | case 'j': |
701 | case 'J': |
702 | if (isImaginary) break; |
703 | isImaginary = true; |
704 | continue; |
705 | } |
706 | |
707 | break; |
708 | } |
709 | |
710 | |
711 | if (s != ThisTokEnd || isImaginary) { |
712 | |
713 | expandUCNs(UDSuffixBuf, StringRef(SuffixBegin, ThisTokEnd - SuffixBegin)); |
714 | if (isValidUDSuffix(PP.getLangOpts(), UDSuffixBuf)) { |
715 | if (!isImaginary) { |
716 | |
717 | |
718 | isLong = false; |
719 | isUnsigned = false; |
720 | isLongLong = false; |
721 | isFloat = false; |
722 | isFloat16 = false; |
723 | isHalf = false; |
724 | isImaginary = false; |
725 | MicrosoftInteger = 0; |
726 | saw_fixed_point_suffix = false; |
727 | isFract = false; |
728 | isAccum = false; |
729 | } |
730 | |
731 | saw_ud_suffix = true; |
732 | return; |
733 | } |
734 | |
735 | if (s != ThisTokEnd) { |
736 | |
737 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, SuffixBegin - ThisTokBegin), |
738 | diag::err_invalid_suffix_constant) |
739 | << StringRef(SuffixBegin, ThisTokEnd - SuffixBegin) << isFPConstant; |
740 | hadError = true; |
741 | } |
742 | } |
743 | |
744 | if (!hadError && saw_fixed_point_suffix) { |
745 | assert(isFract || isAccum); |
746 | } |
747 | } |
748 | |
749 | |
750 | |
751 | |
752 | void NumericLiteralParser::ParseDecimalOrOctalCommon(SourceLocation TokLoc){ |
753 | (0) . __assert_fail ("(radix == 8 || radix == 10) && \"Unexpected radix\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 753, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert((radix == 8 || radix == 10) && "Unexpected radix"); |
754 | |
755 | |
756 | |
757 | if (isHexDigit(*s) && *s != 'e' && *s != 'E' && |
758 | !isValidUDSuffix(PP.getLangOpts(), StringRef(s, ThisTokEnd - s))) { |
759 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, s-ThisTokBegin), |
760 | diag::err_invalid_digit) << StringRef(s, 1) << (radix == 8 ? 1 : 0); |
761 | hadError = true; |
762 | return; |
763 | } |
764 | |
765 | if (*s == '.') { |
766 | checkSeparator(TokLoc, s, CSK_AfterDigits); |
767 | s++; |
768 | radix = 10; |
769 | saw_period = true; |
770 | checkSeparator(TokLoc, s, CSK_BeforeDigits); |
771 | s = SkipDigits(s); |
772 | } |
773 | if (*s == 'e' || *s == 'E') { |
774 | checkSeparator(TokLoc, s, CSK_AfterDigits); |
775 | const char *Exponent = s; |
776 | s++; |
777 | radix = 10; |
778 | saw_exponent = true; |
779 | if (s != ThisTokEnd && (*s == '+' || *s == '-')) s++; |
780 | const char *first_non_digit = SkipDigits(s); |
781 | if (containsDigits(s, first_non_digit)) { |
782 | checkSeparator(TokLoc, s, CSK_BeforeDigits); |
783 | s = first_non_digit; |
784 | } else { |
785 | if (!hadError) { |
786 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, Exponent-ThisTokBegin), |
787 | diag::err_exponent_has_no_digits); |
788 | hadError = true; |
789 | } |
790 | return; |
791 | } |
792 | } |
793 | } |
794 | |
795 | |
796 | |
797 | |
798 | bool NumericLiteralParser::isValidUDSuffix(const LangOptions &LangOpts, |
799 | StringRef Suffix) { |
800 | if (!LangOpts.CPlusPlus11 || Suffix.empty()) |
801 | return false; |
802 | |
803 | |
804 | if (Suffix[0] == '_') |
805 | return true; |
806 | |
807 | |
808 | if (!LangOpts.CPlusPlus14) |
809 | return false; |
810 | |
811 | |
812 | |
813 | |
814 | return llvm::StringSwitch<bool>(Suffix) |
815 | .Cases("h", "min", "s", true) |
816 | .Cases("ms", "us", "ns", true) |
817 | .Cases("il", "i", "if", true) |
818 | .Cases("d", "y", LangOpts.CPlusPlus2a) |
819 | .Default(false); |
820 | } |
821 | |
822 | void NumericLiteralParser::checkSeparator(SourceLocation TokLoc, |
823 | const char *Pos, |
824 | CheckSeparatorKind IsAfterDigits) { |
825 | if (IsAfterDigits == CSK_AfterDigits) { |
826 | if (Pos == ThisTokBegin) |
827 | return; |
828 | --Pos; |
829 | } else if (Pos == ThisTokEnd) |
830 | return; |
831 | |
832 | if (isDigitSeparator(*Pos)) { |
833 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, Pos - ThisTokBegin), |
834 | diag::err_digit_separator_not_between_digits) |
835 | << IsAfterDigits; |
836 | hadError = true; |
837 | } |
838 | } |
839 | |
840 | |
841 | |
842 | |
843 | |
844 | |
845 | void NumericLiteralParser::ParseNumberStartingWithZero(SourceLocation TokLoc) { |
846 | (0) . __assert_fail ("s[0] == '0' && \"Invalid method call\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 846, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(s[0] == '0' && "Invalid method call"); |
847 | s++; |
848 | |
849 | int c1 = s[0]; |
850 | |
851 | |
852 | if ((c1 == 'x' || c1 == 'X') && (isHexDigit(s[1]) || s[1] == '.')) { |
853 | s++; |
854 | (0) . __assert_fail ("s < ThisTokEnd && \"didn't maximally munch?\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 854, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(s < ThisTokEnd && "didn't maximally munch?"); |
855 | radix = 16; |
856 | DigitsBegin = s; |
857 | s = SkipHexDigits(s); |
858 | bool HasSignificandDigits = containsDigits(DigitsBegin, s); |
859 | if (s == ThisTokEnd) { |
860 | |
861 | } else if (*s == '.') { |
862 | s++; |
863 | saw_period = true; |
864 | const char *floatDigitsBegin = s; |
865 | s = SkipHexDigits(s); |
866 | if (containsDigits(floatDigitsBegin, s)) |
867 | HasSignificandDigits = true; |
868 | if (HasSignificandDigits) |
869 | checkSeparator(TokLoc, floatDigitsBegin, CSK_BeforeDigits); |
870 | } |
871 | |
872 | if (!HasSignificandDigits) { |
873 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, s - ThisTokBegin), |
874 | diag::err_hex_constant_requires) |
875 | << PP.getLangOpts().CPlusPlus << 1; |
876 | hadError = true; |
877 | return; |
878 | } |
879 | |
880 | |
881 | |
882 | if (*s == 'p' || *s == 'P') { |
883 | checkSeparator(TokLoc, s, CSK_AfterDigits); |
884 | const char *Exponent = s; |
885 | s++; |
886 | saw_exponent = true; |
887 | if (s != ThisTokEnd && (*s == '+' || *s == '-')) s++; |
888 | const char *first_non_digit = SkipDigits(s); |
889 | if (!containsDigits(s, first_non_digit)) { |
890 | if (!hadError) { |
891 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, Exponent-ThisTokBegin), |
892 | diag::err_exponent_has_no_digits); |
893 | hadError = true; |
894 | } |
895 | return; |
896 | } |
897 | checkSeparator(TokLoc, s, CSK_BeforeDigits); |
898 | s = first_non_digit; |
899 | |
900 | if (!PP.getLangOpts().HexFloats) |
901 | PP.Diag(TokLoc, PP.getLangOpts().CPlusPlus |
902 | ? diag::ext_hex_literal_invalid |
903 | : diag::ext_hex_constant_invalid); |
904 | else if (PP.getLangOpts().CPlusPlus17) |
905 | PP.Diag(TokLoc, diag::warn_cxx17_hex_literal); |
906 | } else if (saw_period) { |
907 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, s - ThisTokBegin), |
908 | diag::err_hex_constant_requires) |
909 | << PP.getLangOpts().CPlusPlus << 0; |
910 | hadError = true; |
911 | } |
912 | return; |
913 | } |
914 | |
915 | |
916 | if ((c1 == 'b' || c1 == 'B') && (s[1] == '0' || s[1] == '1')) { |
917 | |
918 | PP.Diag(TokLoc, |
919 | PP.getLangOpts().CPlusPlus14 |
920 | ? diag::warn_cxx11_compat_binary_literal |
921 | : PP.getLangOpts().CPlusPlus |
922 | ? diag::ext_binary_literal_cxx14 |
923 | : diag::ext_binary_literal); |
924 | ++s; |
925 | (0) . __assert_fail ("s < ThisTokEnd && \"didn't maximally munch?\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 925, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(s < ThisTokEnd && "didn't maximally munch?"); |
926 | radix = 2; |
927 | DigitsBegin = s; |
928 | s = SkipBinaryDigits(s); |
929 | if (s == ThisTokEnd) { |
930 | |
931 | } else if (isHexDigit(*s) && |
932 | !isValidUDSuffix(PP.getLangOpts(), |
933 | StringRef(s, ThisTokEnd - s))) { |
934 | PP.Diag(PP.AdvanceToTokenCharacter(TokLoc, s-ThisTokBegin), |
935 | diag::err_invalid_digit) << StringRef(s, 1) << 2; |
936 | hadError = true; |
937 | } |
938 | |
939 | return; |
940 | } |
941 | |
942 | |
943 | |
944 | |
945 | radix = 8; |
946 | DigitsBegin = s; |
947 | s = SkipOctalDigits(s); |
948 | if (s == ThisTokEnd) |
949 | return; |
950 | |
951 | |
952 | |
953 | if (isDigit(*s)) { |
954 | const char *EndDecimal = SkipDigits(s); |
955 | if (EndDecimal[0] == '.' || EndDecimal[0] == 'e' || EndDecimal[0] == 'E') { |
956 | s = EndDecimal; |
957 | radix = 10; |
958 | } |
959 | } |
960 | |
961 | ParseDecimalOrOctalCommon(TokLoc); |
962 | } |
963 | |
964 | static bool alwaysFitsInto64Bits(unsigned Radix, unsigned NumDigits) { |
965 | switch (Radix) { |
966 | case 2: |
967 | return NumDigits <= 64; |
968 | case 8: |
969 | return NumDigits <= 64 / 3; |
970 | case 10: |
971 | return NumDigits <= 19; |
972 | case 16: |
973 | return NumDigits <= 64 / 4; |
974 | default: |
975 | llvm_unreachable("impossible Radix"); |
976 | } |
977 | } |
978 | |
979 | |
980 | |
981 | |
982 | bool NumericLiteralParser::GetIntegerValue(llvm::APInt &Val) { |
983 | |
984 | |
985 | |
986 | |
987 | |
988 | |
989 | const unsigned NumDigits = SuffixBegin - DigitsBegin; |
990 | if (alwaysFitsInto64Bits(radix, NumDigits)) { |
991 | uint64_t N = 0; |
992 | for (const char *Ptr = DigitsBegin; Ptr != SuffixBegin; ++Ptr) |
993 | if (!isDigitSeparator(*Ptr)) |
994 | N = N * radix + llvm::hexDigitValue(*Ptr); |
995 | |
996 | |
997 | |
998 | Val = N; |
999 | return Val.getZExtValue() != N; |
1000 | } |
1001 | |
1002 | Val = 0; |
1003 | const char *Ptr = DigitsBegin; |
1004 | |
1005 | llvm::APInt RadixVal(Val.getBitWidth(), radix); |
1006 | llvm::APInt CharVal(Val.getBitWidth(), 0); |
1007 | llvm::APInt OldVal = Val; |
1008 | |
1009 | bool OverflowOccurred = false; |
1010 | while (Ptr < SuffixBegin) { |
1011 | if (isDigitSeparator(*Ptr)) { |
1012 | ++Ptr; |
1013 | continue; |
1014 | } |
1015 | |
1016 | unsigned C = llvm::hexDigitValue(*Ptr++); |
1017 | |
1018 | |
1019 | (0) . __assert_fail ("C < radix && \"NumericLiteralParser ctor should have rejected this\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1019, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(C < radix && "NumericLiteralParser ctor should have rejected this"); |
1020 | |
1021 | CharVal = C; |
1022 | |
1023 | |
1024 | |
1025 | OldVal = Val; |
1026 | |
1027 | |
1028 | Val *= RadixVal; |
1029 | OverflowOccurred |= Val.udiv(RadixVal) != OldVal; |
1030 | |
1031 | |
1032 | |
1033 | Val += CharVal; |
1034 | OverflowOccurred |= Val.ult(CharVal); |
1035 | } |
1036 | return OverflowOccurred; |
1037 | } |
1038 | |
1039 | llvm::APFloat::opStatus |
1040 | NumericLiteralParser::GetFloatValue(llvm::APFloat &Result) { |
1041 | using llvm::APFloat; |
1042 | |
1043 | unsigned n = std::min(SuffixBegin - ThisTokBegin, ThisTokEnd - ThisTokBegin); |
1044 | |
1045 | llvm::SmallString<16> Buffer; |
1046 | StringRef Str(ThisTokBegin, n); |
1047 | if (Str.find('\'') != StringRef::npos) { |
1048 | Buffer.reserve(n); |
1049 | std::remove_copy_if(Str.begin(), Str.end(), std::back_inserter(Buffer), |
1050 | &isDigitSeparator); |
1051 | Str = Buffer; |
1052 | } |
1053 | |
1054 | return Result.convertFromString(Str, APFloat::rmNearestTiesToEven); |
1055 | } |
1056 | |
1057 | static inline bool IsExponentPart(char c) { |
1058 | return c == 'p' || c == 'P' || c == 'e' || c == 'E'; |
1059 | } |
1060 | |
1061 | bool NumericLiteralParser::GetFixedPointValue(llvm::APInt &StoreVal, unsigned Scale) { |
1062 | assert(radix == 16 || radix == 10); |
1063 | |
1064 | |
1065 | unsigned NumDigits = SuffixBegin - DigitsBegin; |
1066 | if (saw_period) --NumDigits; |
1067 | |
1068 | |
1069 | bool ExpOverflowOccurred = false; |
1070 | bool NegativeExponent = false; |
1071 | const char *ExponentBegin; |
1072 | uint64_t Exponent = 0; |
1073 | int64_t BaseShift = 0; |
1074 | if (saw_exponent) { |
1075 | const char *Ptr = DigitsBegin; |
1076 | |
1077 | while (!IsExponentPart(*Ptr)) ++Ptr; |
1078 | ExponentBegin = Ptr; |
1079 | ++Ptr; |
1080 | NegativeExponent = *Ptr == '-'; |
1081 | if (NegativeExponent) ++Ptr; |
1082 | |
1083 | unsigned NumExpDigits = SuffixBegin - Ptr; |
1084 | if (alwaysFitsInto64Bits(radix, NumExpDigits)) { |
1085 | llvm::StringRef ExpStr(Ptr, NumExpDigits); |
1086 | llvm::APInt ExpInt(, ExpStr, ); |
1087 | Exponent = ExpInt.getZExtValue(); |
1088 | } else { |
1089 | ExpOverflowOccurred = true; |
1090 | } |
1091 | |
1092 | if (NegativeExponent) BaseShift -= Exponent; |
1093 | else BaseShift += Exponent; |
1094 | } |
1095 | |
1096 | |
1097 | |
1098 | |
1099 | |
1100 | |
1101 | |
1102 | |
1103 | |
1104 | |
1105 | |
1106 | |
1107 | |
1108 | |
1109 | |
1110 | |
1111 | |
1112 | |
1113 | uint64_t NumBitsNeeded; |
1114 | if (radix == 10) |
1115 | NumBitsNeeded = 4 * (NumDigits + Exponent) + Scale; |
1116 | else |
1117 | NumBitsNeeded = 4 * NumDigits + Exponent + Scale; |
1118 | |
1119 | if (NumBitsNeeded > std::numeric_limits<unsigned>::max()) |
1120 | ExpOverflowOccurred = true; |
1121 | llvm::APInt Val(static_cast<unsigned>(NumBitsNeeded), 0, ); |
1122 | |
1123 | bool FoundDecimal = false; |
1124 | |
1125 | int64_t FractBaseShift = 0; |
1126 | const char *End = saw_exponent ? ExponentBegin : SuffixBegin; |
1127 | for (const char *Ptr = DigitsBegin; Ptr < End; ++Ptr) { |
1128 | if (*Ptr == '.') { |
1129 | FoundDecimal = true; |
1130 | continue; |
1131 | } |
1132 | |
1133 | |
1134 | unsigned C = llvm::hexDigitValue(*Ptr); |
1135 | (0) . __assert_fail ("C < radix && \"NumericLiteralParser ctor should have rejected this\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1135, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(C < radix && "NumericLiteralParser ctor should have rejected this"); |
1136 | |
1137 | Val *= radix; |
1138 | Val += C; |
1139 | |
1140 | if (FoundDecimal) |
1141 | |
1142 | |
1143 | --FractBaseShift; |
1144 | } |
1145 | |
1146 | |
1147 | if (radix == 16) FractBaseShift *= 4; |
1148 | BaseShift += FractBaseShift; |
1149 | |
1150 | Val <<= Scale; |
1151 | |
1152 | uint64_t Base = (radix == 16) ? 2 : 10; |
1153 | if (BaseShift > 0) { |
1154 | for (int64_t i = 0; i < BaseShift; ++i) { |
1155 | Val *= Base; |
1156 | } |
1157 | } else if (BaseShift < 0) { |
1158 | for (int64_t i = BaseShift; i < 0 && !Val.isNullValue(); ++i) |
1159 | Val = Val.udiv(Base); |
1160 | } |
1161 | |
1162 | bool IntOverflowOccurred = false; |
1163 | auto MaxVal = llvm::APInt::getMaxValue(StoreVal.getBitWidth()); |
1164 | if (Val.getBitWidth() > StoreVal.getBitWidth()) { |
1165 | IntOverflowOccurred |= Val.ugt(MaxVal.zext(Val.getBitWidth())); |
1166 | StoreVal = Val.trunc(StoreVal.getBitWidth()); |
1167 | } else if (Val.getBitWidth() < StoreVal.getBitWidth()) { |
1168 | IntOverflowOccurred |= Val.zext(MaxVal.getBitWidth()).ugt(MaxVal); |
1169 | StoreVal = Val.zext(StoreVal.getBitWidth()); |
1170 | } else { |
1171 | StoreVal = Val; |
1172 | } |
1173 | |
1174 | return IntOverflowOccurred || ExpOverflowOccurred; |
1175 | } |
1176 | |
1177 | |
1178 | |
1179 | |
1180 | |
1181 | |
1182 | |
1183 | |
1184 | |
1185 | |
1186 | |
1187 | |
1188 | |
1189 | |
1190 | |
1191 | |
1192 | |
1193 | |
1194 | |
1195 | |
1196 | |
1197 | |
1198 | |
1199 | |
1200 | |
1201 | |
1202 | |
1203 | |
1204 | |
1205 | |
1206 | |
1207 | |
1208 | |
1209 | |
1210 | |
1211 | |
1212 | |
1213 | |
1214 | |
1215 | |
1216 | CharLiteralParser::CharLiteralParser(const char *begin, const char *end, |
1217 | SourceLocation Loc, Preprocessor &PP, |
1218 | tok::TokenKind kind) { |
1219 | |
1220 | HadError = false; |
1221 | |
1222 | Kind = kind; |
1223 | |
1224 | const char *TokBegin = begin; |
1225 | |
1226 | |
1227 | if (Kind != tok::char_constant) |
1228 | ++begin; |
1229 | if (Kind == tok::utf8_char_constant) |
1230 | ++begin; |
1231 | |
1232 | |
1233 | (0) . __assert_fail ("begin[0] == '\\'' && \"Invalid token lexed\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1233, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(begin[0] == '\'' && "Invalid token lexed"); |
1234 | ++begin; |
1235 | |
1236 | |
1237 | if (end[-1] != '\'') { |
1238 | const char *UDSuffixEnd = end; |
1239 | do { |
1240 | --end; |
1241 | } while (end[-1] != '\''); |
1242 | |
1243 | expandUCNs(UDSuffixBuf, StringRef(end, UDSuffixEnd - end)); |
1244 | UDSuffixOffset = end - TokBegin; |
1245 | } |
1246 | |
1247 | |
1248 | (0) . __assert_fail ("end != begin && \"Invalid token lexed\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1248, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(end != begin && "Invalid token lexed"); |
1249 | --end; |
1250 | |
1251 | |
1252 | |
1253 | |
1254 | (0) . __assert_fail ("PP.getTargetInfo().getCharWidth() == 8 && \"Assumes char is 8 bits\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1255, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(PP.getTargetInfo().getCharWidth() == 8 && |
1255 | (0) . __assert_fail ("PP.getTargetInfo().getCharWidth() == 8 && \"Assumes char is 8 bits\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1255, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> "Assumes char is 8 bits"); |
1256 | (0) . __assert_fail ("PP.getTargetInfo().getIntWidth() <= 64 && (PP.getTargetInfo().getIntWidth() & 7) == 0 && \"Assumes sizeof(int) on target is <= 64 and a multiple of char\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1258, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(PP.getTargetInfo().getIntWidth() <= 64 && |
1257 | (0) . __assert_fail ("PP.getTargetInfo().getIntWidth() <= 64 && (PP.getTargetInfo().getIntWidth() & 7) == 0 && \"Assumes sizeof(int) on target is <= 64 and a multiple of char\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1258, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> (PP.getTargetInfo().getIntWidth() & 7) == 0 && |
1258 | (0) . __assert_fail ("PP.getTargetInfo().getIntWidth() <= 64 && (PP.getTargetInfo().getIntWidth() & 7) == 0 && \"Assumes sizeof(int) on target is <= 64 and a multiple of char\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1258, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> "Assumes sizeof(int) on target is <= 64 and a multiple of char"); |
1259 | (0) . __assert_fail ("PP.getTargetInfo().getWCharWidth() <= 64 && \"Assumes sizeof(wchar) on target is <= 64\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1260, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(PP.getTargetInfo().getWCharWidth() <= 64 && |
1260 | (0) . __assert_fail ("PP.getTargetInfo().getWCharWidth() <= 64 && \"Assumes sizeof(wchar) on target is <= 64\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1260, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> "Assumes sizeof(wchar) on target is <= 64"); |
1261 | |
1262 | SmallVector<uint32_t, 4> codepoint_buffer; |
1263 | codepoint_buffer.resize(end - begin); |
1264 | uint32_t *buffer_begin = &codepoint_buffer.front(); |
1265 | uint32_t *buffer_end = buffer_begin + codepoint_buffer.size(); |
1266 | |
1267 | |
1268 | |
1269 | |
1270 | uint32_t largest_character_for_kind; |
1271 | if (tok::wide_char_constant == Kind) { |
1272 | largest_character_for_kind = |
1273 | 0xFFFFFFFFu >> (32-PP.getTargetInfo().getWCharWidth()); |
1274 | } else if (tok::utf8_char_constant == Kind) { |
1275 | largest_character_for_kind = 0x7F; |
1276 | } else if (tok::utf16_char_constant == Kind) { |
1277 | largest_character_for_kind = 0xFFFF; |
1278 | } else if (tok::utf32_char_constant == Kind) { |
1279 | largest_character_for_kind = 0x10FFFF; |
1280 | } else { |
1281 | largest_character_for_kind = 0x7Fu; |
1282 | } |
1283 | |
1284 | while (begin != end) { |
1285 | |
1286 | if (begin[0] != '\\') { |
1287 | char const *start = begin; |
1288 | do { |
1289 | ++begin; |
1290 | } while (begin != end && *begin != '\\'); |
1291 | |
1292 | char const *tmp_in_start = start; |
1293 | uint32_t *tmp_out_start = buffer_begin; |
1294 | llvm::ConversionResult res = |
1295 | llvm::ConvertUTF8toUTF32(reinterpret_cast<llvm::UTF8 const **>(&start), |
1296 | reinterpret_cast<llvm::UTF8 const *>(begin), |
1297 | &buffer_begin, buffer_end, llvm::strictConversion); |
1298 | if (res != llvm::conversionOK) { |
1299 | |
1300 | |
1301 | |
1302 | bool NoErrorOnBadEncoding = isAscii(); |
1303 | unsigned Msg = diag::err_bad_character_encoding; |
1304 | if (NoErrorOnBadEncoding) |
1305 | Msg = diag::warn_bad_character_encoding; |
1306 | PP.Diag(Loc, Msg); |
1307 | if (NoErrorOnBadEncoding) { |
1308 | start = tmp_in_start; |
1309 | buffer_begin = tmp_out_start; |
1310 | for (; start != begin; ++start, ++buffer_begin) |
1311 | *buffer_begin = static_cast<uint8_t>(*start); |
1312 | } else { |
1313 | HadError = true; |
1314 | } |
1315 | } else { |
1316 | for (; tmp_out_start < buffer_begin; ++tmp_out_start) { |
1317 | if (*tmp_out_start > largest_character_for_kind) { |
1318 | HadError = true; |
1319 | PP.Diag(Loc, diag::err_character_too_large); |
1320 | } |
1321 | } |
1322 | } |
1323 | |
1324 | continue; |
1325 | } |
1326 | |
1327 | if (begin[1] == 'u' || begin[1] == 'U') { |
1328 | unsigned short UcnLen = 0; |
1329 | if (!ProcessUCNEscape(TokBegin, begin, end, *buffer_begin, UcnLen, |
1330 | FullSourceLoc(Loc, PP.getSourceManager()), |
1331 | &PP.getDiagnostics(), PP.getLangOpts(), true)) { |
1332 | HadError = true; |
1333 | } else if (*buffer_begin > largest_character_for_kind) { |
1334 | HadError = true; |
1335 | PP.Diag(Loc, diag::err_character_too_large); |
1336 | } |
1337 | |
1338 | ++buffer_begin; |
1339 | continue; |
1340 | } |
1341 | unsigned CharWidth = getCharWidth(Kind, PP.getTargetInfo()); |
1342 | uint64_t result = |
1343 | ProcessCharEscape(TokBegin, begin, end, HadError, |
1344 | FullSourceLoc(Loc,PP.getSourceManager()), |
1345 | CharWidth, &PP.getDiagnostics(), PP.getLangOpts()); |
1346 | *buffer_begin++ = result; |
1347 | } |
1348 | |
1349 | unsigned = buffer_begin - &codepoint_buffer.front(); |
1350 | |
1351 | if (NumCharsSoFar > 1) { |
1352 | if (isWide()) |
1353 | PP.Diag(Loc, diag::warn_extraneous_char_constant); |
1354 | else if (isAscii() && NumCharsSoFar == 4) |
1355 | PP.Diag(Loc, diag::ext_four_char_character_literal); |
1356 | else if (isAscii()) |
1357 | PP.Diag(Loc, diag::ext_multichar_character_literal); |
1358 | else |
1359 | PP.Diag(Loc, diag::err_multichar_utf_character_literal); |
1360 | IsMultiChar = true; |
1361 | } else { |
1362 | IsMultiChar = false; |
1363 | } |
1364 | |
1365 | llvm::APInt LitVal(PP.getTargetInfo().getIntWidth(), 0); |
1366 | |
1367 | |
1368 | |
1369 | bool multi_char_too_long = false; |
1370 | if (isAscii() && isMultiChar()) { |
1371 | LitVal = 0; |
1372 | for (size_t i = 0; i < NumCharsSoFar; ++i) { |
1373 | |
1374 | multi_char_too_long |= (LitVal.countLeadingZeros() < 8); |
1375 | LitVal <<= 8; |
1376 | LitVal = LitVal + (codepoint_buffer[i] & 0xFF); |
1377 | } |
1378 | } else if (NumCharsSoFar > 0) { |
1379 | |
1380 | LitVal = buffer_begin[-1]; |
1381 | } |
1382 | |
1383 | if (!HadError && multi_char_too_long) { |
1384 | PP.Diag(Loc, diag::warn_char_constant_too_large); |
1385 | } |
1386 | |
1387 | |
1388 | Value = LitVal.getZExtValue(); |
1389 | |
1390 | |
1391 | |
1392 | |
1393 | |
1394 | if (isAscii() && NumCharsSoFar == 1 && (Value & 128) && |
1395 | PP.getLangOpts().CharIsSigned) |
1396 | Value = (signed char)Value; |
1397 | } |
1398 | |
1399 | |
1400 | |
1401 | |
1402 | |
1403 | |
1404 | |
1405 | |
1406 | |
1407 | |
1408 | |
1409 | |
1410 | |
1411 | |
1412 | |
1413 | |
1414 | |
1415 | |
1416 | |
1417 | |
1418 | |
1419 | |
1420 | |
1421 | |
1422 | |
1423 | |
1424 | |
1425 | |
1426 | |
1427 | |
1428 | |
1429 | |
1430 | |
1431 | |
1432 | |
1433 | |
1434 | |
1435 | |
1436 | |
1437 | |
1438 | |
1439 | |
1440 | |
1441 | |
1442 | |
1443 | |
1444 | |
1445 | |
1446 | |
1447 | |
1448 | |
1449 | |
1450 | |
1451 | |
1452 | |
1453 | StringLiteralParser:: |
1454 | StringLiteralParser(ArrayRef<Token> StringToks, |
1455 | Preprocessor &PP, bool Complain) |
1456 | : SM(PP.getSourceManager()), Features(PP.getLangOpts()), |
1457 | Target(PP.getTargetInfo()), Diags(Complain ? &PP.getDiagnostics() :nullptr), |
1458 | MaxTokenLength(0), SizeBound(0), CharByteWidth(0), Kind(tok::unknown), |
1459 | ResultPtr(ResultBuf.data()), hadError(false), Pascal(false) { |
1460 | init(StringToks); |
1461 | } |
1462 | |
1463 | void StringLiteralParser::init(ArrayRef<Token> StringToks){ |
1464 | |
1465 | |
1466 | if (StringToks.empty() || StringToks[0].getLength() < 2) |
1467 | return DiagnoseLexingError(SourceLocation()); |
1468 | |
1469 | |
1470 | |
1471 | |
1472 | |
1473 | (0) . __assert_fail ("!StringToks.empty() && \"expected at least one token\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1473, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(!StringToks.empty() && "expected at least one token"); |
1474 | MaxTokenLength = StringToks[0].getLength(); |
1475 | (0) . __assert_fail ("StringToks[0].getLength() >= 2 && \"literal token is invalid!\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1475, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(StringToks[0].getLength() >= 2 && "literal token is invalid!"); |
1476 | SizeBound = StringToks[0].getLength()-2; |
1477 | Kind = StringToks[0].getKind(); |
1478 | |
1479 | hadError = false; |
1480 | |
1481 | |
1482 | |
1483 | for (unsigned i = 1; i != StringToks.size(); ++i) { |
1484 | if (StringToks[i].getLength() < 2) |
1485 | return DiagnoseLexingError(StringToks[i].getLocation()); |
1486 | |
1487 | |
1488 | |
1489 | (0) . __assert_fail ("StringToks[i].getLength() >= 2 && \"literal token is invalid!\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1489, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(StringToks[i].getLength() >= 2 && "literal token is invalid!"); |
1490 | SizeBound += StringToks[i].getLength()-2; |
1491 | |
1492 | |
1493 | if (StringToks[i].getLength() > MaxTokenLength) |
1494 | MaxTokenLength = StringToks[i].getLength(); |
1495 | |
1496 | |
1497 | |
1498 | if (StringToks[i].isNot(Kind) && StringToks[i].isNot(tok::string_literal)) { |
1499 | if (isAscii()) { |
1500 | Kind = StringToks[i].getKind(); |
1501 | } else { |
1502 | if (Diags) |
1503 | Diags->Report(StringToks[i].getLocation(), |
1504 | diag::err_unsupported_string_concat); |
1505 | hadError = true; |
1506 | } |
1507 | } |
1508 | } |
1509 | |
1510 | |
1511 | ++SizeBound; |
1512 | |
1513 | |
1514 | |
1515 | |
1516 | CharByteWidth = getCharWidth(Kind, Target); |
1517 | (0) . __assert_fail ("(CharByteWidth & 7) == 0 && \"Assumes character size is byte multiple\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1517, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert((CharByteWidth & 7) == 0 && "Assumes character size is byte multiple"); |
1518 | CharByteWidth /= 8; |
1519 | |
1520 | |
1521 | |
1522 | SizeBound *= CharByteWidth; |
1523 | |
1524 | |
1525 | ResultBuf.resize(SizeBound); |
1526 | |
1527 | |
1528 | SmallString<512> TokenBuf; |
1529 | TokenBuf.resize(MaxTokenLength); |
1530 | |
1531 | |
1532 | |
1533 | ResultPtr = &ResultBuf[0]; |
1534 | |
1535 | Pascal = false; |
1536 | |
1537 | SourceLocation UDSuffixTokLoc; |
1538 | |
1539 | for (unsigned i = 0, e = StringToks.size(); i != e; ++i) { |
1540 | const char *ThisTokBuf = &TokenBuf[0]; |
1541 | |
1542 | |
1543 | |
1544 | bool StringInvalid = false; |
1545 | unsigned ThisTokLen = |
1546 | Lexer::getSpelling(StringToks[i], ThisTokBuf, SM, Features, |
1547 | &StringInvalid); |
1548 | if (StringInvalid) |
1549 | return DiagnoseLexingError(StringToks[i].getLocation()); |
1550 | |
1551 | const char *ThisTokBegin = ThisTokBuf; |
1552 | const char *ThisTokEnd = ThisTokBuf+ThisTokLen; |
1553 | |
1554 | |
1555 | if (ThisTokEnd[-1] != '"') { |
1556 | const char *UDSuffixEnd = ThisTokEnd; |
1557 | do { |
1558 | --ThisTokEnd; |
1559 | } while (ThisTokEnd[-1] != '"'); |
1560 | |
1561 | StringRef UDSuffix(ThisTokEnd, UDSuffixEnd - ThisTokEnd); |
1562 | |
1563 | if (UDSuffixBuf.empty()) { |
1564 | if (StringToks[i].hasUCN()) |
1565 | expandUCNs(UDSuffixBuf, UDSuffix); |
1566 | else |
1567 | UDSuffixBuf.assign(UDSuffix); |
1568 | UDSuffixToken = i; |
1569 | UDSuffixOffset = ThisTokEnd - ThisTokBuf; |
1570 | UDSuffixTokLoc = StringToks[i].getLocation(); |
1571 | } else { |
1572 | SmallString<32> ExpandedUDSuffix; |
1573 | if (StringToks[i].hasUCN()) { |
1574 | expandUCNs(ExpandedUDSuffix, UDSuffix); |
1575 | UDSuffix = ExpandedUDSuffix; |
1576 | } |
1577 | |
1578 | |
1579 | |
1580 | |
1581 | |
1582 | if (UDSuffixBuf != UDSuffix) { |
1583 | if (Diags) { |
1584 | SourceLocation TokLoc = StringToks[i].getLocation(); |
1585 | Diags->Report(TokLoc, diag::err_string_concat_mixed_suffix) |
1586 | << UDSuffixBuf << UDSuffix |
1587 | << SourceRange(UDSuffixTokLoc, UDSuffixTokLoc) |
1588 | << SourceRange(TokLoc, TokLoc); |
1589 | } |
1590 | hadError = true; |
1591 | } |
1592 | } |
1593 | } |
1594 | |
1595 | |
1596 | --ThisTokEnd; |
1597 | |
1598 | |
1599 | |
1600 | |
1601 | if (ThisTokBuf[0] == 'L' || ThisTokBuf[0] == 'u' || ThisTokBuf[0] == 'U') { |
1602 | ++ThisTokBuf; |
1603 | |
1604 | if (ThisTokBuf[0] == '8') |
1605 | ++ThisTokBuf; |
1606 | } |
1607 | |
1608 | |
1609 | if (ThisTokBuf[0] == 'R') { |
1610 | ThisTokBuf += 2; |
1611 | |
1612 | const char *Prefix = ThisTokBuf; |
1613 | while (ThisTokBuf[0] != '(') |
1614 | ++ThisTokBuf; |
1615 | ++ThisTokBuf; |
1616 | |
1617 | |
1618 | ThisTokEnd -= ThisTokBuf - Prefix; |
1619 | (0) . __assert_fail ("ThisTokEnd >= ThisTokBuf && \"malformed raw string literal\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1619, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(ThisTokEnd >= ThisTokBuf && "malformed raw string literal"); |
1620 | |
1621 | |
1622 | |
1623 | StringRef RemainingTokenSpan(ThisTokBuf, ThisTokEnd - ThisTokBuf); |
1624 | while (!RemainingTokenSpan.empty()) { |
1625 | |
1626 | size_t CRLFPos = RemainingTokenSpan.find("\r\n"); |
1627 | StringRef BeforeCRLF = RemainingTokenSpan.substr(0, CRLFPos); |
1628 | StringRef AfterCRLF = RemainingTokenSpan.substr(CRLFPos); |
1629 | |
1630 | |
1631 | if (CopyStringFragment(StringToks[i], ThisTokBegin, BeforeCRLF)) |
1632 | hadError = true; |
1633 | |
1634 | |
1635 | |
1636 | RemainingTokenSpan = AfterCRLF.substr(1); |
1637 | } |
1638 | } else { |
1639 | if (ThisTokBuf[0] != '"') { |
1640 | |
1641 | |
1642 | return DiagnoseLexingError(StringToks[i].getLocation()); |
1643 | } |
1644 | ++ThisTokBuf; |
1645 | |
1646 | |
1647 | if (Features.PascalStrings && ThisTokBuf + 1 != ThisTokEnd && |
1648 | ThisTokBuf[0] == '\\' && ThisTokBuf[1] == 'p') { |
1649 | |
1650 | |
1651 | |
1652 | if (i == 0) { |
1653 | ++ThisTokBuf; |
1654 | Pascal = true; |
1655 | } else if (Pascal) |
1656 | ThisTokBuf += 2; |
1657 | } |
1658 | |
1659 | while (ThisTokBuf != ThisTokEnd) { |
1660 | |
1661 | if (ThisTokBuf[0] != '\\') { |
1662 | const char *InStart = ThisTokBuf; |
1663 | do { |
1664 | ++ThisTokBuf; |
1665 | } while (ThisTokBuf != ThisTokEnd && ThisTokBuf[0] != '\\'); |
1666 | |
1667 | |
1668 | if (CopyStringFragment(StringToks[i], ThisTokBegin, |
1669 | StringRef(InStart, ThisTokBuf - InStart))) |
1670 | hadError = true; |
1671 | continue; |
1672 | } |
1673 | |
1674 | if (ThisTokBuf[1] == 'u' || ThisTokBuf[1] == 'U') { |
1675 | EncodeUCNEscape(ThisTokBegin, ThisTokBuf, ThisTokEnd, |
1676 | ResultPtr, hadError, |
1677 | FullSourceLoc(StringToks[i].getLocation(), SM), |
1678 | CharByteWidth, Diags, Features); |
1679 | continue; |
1680 | } |
1681 | |
1682 | unsigned ResultChar = |
1683 | ProcessCharEscape(ThisTokBegin, ThisTokBuf, ThisTokEnd, hadError, |
1684 | FullSourceLoc(StringToks[i].getLocation(), SM), |
1685 | CharByteWidth*8, Diags, Features); |
1686 | |
1687 | if (CharByteWidth == 4) { |
1688 | |
1689 | |
1690 | llvm::UTF32 *ResultWidePtr = reinterpret_cast<llvm::UTF32*>(ResultPtr); |
1691 | *ResultWidePtr = ResultChar; |
1692 | ResultPtr += 4; |
1693 | } else if (CharByteWidth == 2) { |
1694 | |
1695 | |
1696 | llvm::UTF16 *ResultWidePtr = reinterpret_cast<llvm::UTF16*>(ResultPtr); |
1697 | *ResultWidePtr = ResultChar & 0xFFFF; |
1698 | ResultPtr += 2; |
1699 | } else { |
1700 | (0) . __assert_fail ("CharByteWidth == 1 && \"Unexpected char width\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1700, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(CharByteWidth == 1 && "Unexpected char width"); |
1701 | *ResultPtr++ = ResultChar & 0xFF; |
1702 | } |
1703 | } |
1704 | } |
1705 | } |
1706 | |
1707 | if (Pascal) { |
1708 | if (CharByteWidth == 4) { |
1709 | |
1710 | |
1711 | llvm::UTF32 *ResultWidePtr = reinterpret_cast<llvm::UTF32*>(ResultBuf.data()); |
1712 | ResultWidePtr[0] = GetNumStringChars() - 1; |
1713 | } else if (CharByteWidth == 2) { |
1714 | |
1715 | |
1716 | llvm::UTF16 *ResultWidePtr = reinterpret_cast<llvm::UTF16*>(ResultBuf.data()); |
1717 | ResultWidePtr[0] = GetNumStringChars() - 1; |
1718 | } else { |
1719 | (0) . __assert_fail ("CharByteWidth == 1 && \"Unexpected char width\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1719, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(CharByteWidth == 1 && "Unexpected char width"); |
1720 | ResultBuf[0] = GetNumStringChars() - 1; |
1721 | } |
1722 | |
1723 | |
1724 | if (GetStringLength() > 256) { |
1725 | if (Diags) |
1726 | Diags->Report(StringToks.front().getLocation(), |
1727 | diag::err_pascal_string_too_long) |
1728 | << SourceRange(StringToks.front().getLocation(), |
1729 | StringToks.back().getLocation()); |
1730 | hadError = true; |
1731 | return; |
1732 | } |
1733 | } else if (Diags) { |
1734 | |
1735 | unsigned MaxChars = Features.CPlusPlus? 65536 : Features.C99 ? 4095 : 509; |
1736 | |
1737 | if (GetNumStringChars() > MaxChars) |
1738 | Diags->Report(StringToks.front().getLocation(), |
1739 | diag::ext_string_too_long) |
1740 | << GetNumStringChars() << MaxChars |
1741 | << (Features.CPlusPlus ? 2 : Features.C99 ? 1 : 0) |
1742 | << SourceRange(StringToks.front().getLocation(), |
1743 | StringToks.back().getLocation()); |
1744 | } |
1745 | } |
1746 | |
1747 | static const char *resyncUTF8(const char *Err, const char *End) { |
1748 | if (Err == End) |
1749 | return End; |
1750 | End = Err + std::min<unsigned>(llvm::getNumBytesForUTF8(*Err), End-Err); |
1751 | while (++Err != End && (*Err & 0xC0) == 0x80) |
1752 | ; |
1753 | return Err; |
1754 | } |
1755 | |
1756 | |
1757 | |
1758 | |
1759 | bool StringLiteralParser::CopyStringFragment(const Token &Tok, |
1760 | const char *TokBegin, |
1761 | StringRef Fragment) { |
1762 | const llvm::UTF8 *ErrorPtrTmp; |
1763 | if (ConvertUTF8toWide(CharByteWidth, Fragment, ResultPtr, ErrorPtrTmp)) |
1764 | return false; |
1765 | |
1766 | |
1767 | |
1768 | |
1769 | bool NoErrorOnBadEncoding = isAscii(); |
1770 | if (NoErrorOnBadEncoding) { |
1771 | memcpy(ResultPtr, Fragment.data(), Fragment.size()); |
1772 | ResultPtr += Fragment.size(); |
1773 | } |
1774 | |
1775 | if (Diags) { |
1776 | const char *ErrorPtr = reinterpret_cast<const char *>(ErrorPtrTmp); |
1777 | |
1778 | FullSourceLoc SourceLoc(Tok.getLocation(), SM); |
1779 | const DiagnosticBuilder &Builder = |
1780 | Diag(Diags, Features, SourceLoc, TokBegin, |
1781 | ErrorPtr, resyncUTF8(ErrorPtr, Fragment.end()), |
1782 | NoErrorOnBadEncoding ? diag::warn_bad_string_encoding |
1783 | : diag::err_bad_string_encoding); |
1784 | |
1785 | const char *NextStart = resyncUTF8(ErrorPtr, Fragment.end()); |
1786 | StringRef NextFragment(NextStart, Fragment.end()-NextStart); |
1787 | |
1788 | |
1789 | SmallString<512> Dummy; |
1790 | Dummy.reserve(Fragment.size() * CharByteWidth); |
1791 | char *Ptr = Dummy.data(); |
1792 | |
1793 | while (!ConvertUTF8toWide(CharByteWidth, NextFragment, Ptr, ErrorPtrTmp)) { |
1794 | const char *ErrorPtr = reinterpret_cast<const char *>(ErrorPtrTmp); |
1795 | NextStart = resyncUTF8(ErrorPtr, Fragment.end()); |
1796 | Builder << MakeCharSourceRange(Features, SourceLoc, TokBegin, |
1797 | ErrorPtr, NextStart); |
1798 | NextFragment = StringRef(NextStart, Fragment.end()-NextStart); |
1799 | } |
1800 | } |
1801 | return !NoErrorOnBadEncoding; |
1802 | } |
1803 | |
1804 | void StringLiteralParser::DiagnoseLexingError(SourceLocation Loc) { |
1805 | hadError = true; |
1806 | if (Diags) |
1807 | Diags->Report(Loc, diag::err_lexing_string); |
1808 | } |
1809 | |
1810 | |
1811 | |
1812 | |
1813 | unsigned StringLiteralParser::getOffsetOfStringByte(const Token &Tok, |
1814 | unsigned ByteNo) const { |
1815 | |
1816 | SmallString<32> SpellingBuffer; |
1817 | SpellingBuffer.resize(Tok.getLength()); |
1818 | |
1819 | bool StringInvalid = false; |
1820 | const char *SpellingPtr = &SpellingBuffer[0]; |
1821 | unsigned TokLen = Lexer::getSpelling(Tok, SpellingPtr, SM, Features, |
1822 | &StringInvalid); |
1823 | if (StringInvalid) |
1824 | return 0; |
1825 | |
1826 | const char *SpellingStart = SpellingPtr; |
1827 | const char *SpellingEnd = SpellingPtr+TokLen; |
1828 | |
1829 | |
1830 | if (SpellingPtr[0] == 'u' && SpellingPtr[1] == '8') |
1831 | SpellingPtr += 2; |
1832 | |
1833 | (0) . __assert_fail ("SpellingPtr[0] != 'L' && SpellingPtr[0] != 'u' && SpellingPtr[0] != 'U' && \"Doesn't handle wide or utf strings yet\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1834, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(SpellingPtr[0] != 'L' && SpellingPtr[0] != 'u' && |
1834 | (0) . __assert_fail ("SpellingPtr[0] != 'L' && SpellingPtr[0] != 'u' && SpellingPtr[0] != 'U' && \"Doesn't handle wide or utf strings yet\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1834, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true"> SpellingPtr[0] != 'U' && "Doesn't handle wide or utf strings yet"); |
1835 | |
1836 | |
1837 | if (SpellingPtr[0] == 'R') { |
1838 | (0) . __assert_fail ("SpellingPtr[1] == '\"' && \"Should be a raw string literal!\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1838, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(SpellingPtr[1] == '"' && "Should be a raw string literal!"); |
1839 | |
1840 | SpellingPtr += 2; |
1841 | while (*SpellingPtr != '(') { |
1842 | ++SpellingPtr; |
1843 | (0) . __assert_fail ("SpellingPtr < SpellingEnd && \"Missing ( for raw string literal\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1843, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(SpellingPtr < SpellingEnd && "Missing ( for raw string literal"); |
1844 | } |
1845 | |
1846 | ++SpellingPtr; |
1847 | return SpellingPtr - SpellingStart + ByteNo; |
1848 | } |
1849 | |
1850 | |
1851 | (0) . __assert_fail ("SpellingPtr[0] == '\"' && \"Should be a string literal!\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1851, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(SpellingPtr[0] == '"' && "Should be a string literal!"); |
1852 | ++SpellingPtr; |
1853 | |
1854 | |
1855 | while (ByteNo) { |
1856 | (0) . __assert_fail ("SpellingPtr < SpellingEnd && \"Didn't find byte offset!\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1856, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(SpellingPtr < SpellingEnd && "Didn't find byte offset!"); |
1857 | |
1858 | |
1859 | if (*SpellingPtr != '\\') { |
1860 | ++SpellingPtr; |
1861 | --ByteNo; |
1862 | continue; |
1863 | } |
1864 | |
1865 | |
1866 | bool HadError = false; |
1867 | if (SpellingPtr[1] == 'u' || SpellingPtr[1] == 'U') { |
1868 | const char *EscapePtr = SpellingPtr; |
1869 | unsigned Len = MeasureUCNEscape(SpellingStart, SpellingPtr, SpellingEnd, |
1870 | 1, Features, HadError); |
1871 | if (Len > ByteNo) { |
1872 | |
1873 | SpellingPtr = EscapePtr; |
1874 | break; |
1875 | } |
1876 | ByteNo -= Len; |
1877 | } else { |
1878 | ProcessCharEscape(SpellingStart, SpellingPtr, SpellingEnd, HadError, |
1879 | FullSourceLoc(Tok.getLocation(), SM), |
1880 | CharByteWidth*8, Diags, Features); |
1881 | --ByteNo; |
1882 | } |
1883 | (0) . __assert_fail ("!HadError && \"This method isn't valid on erroneous strings\"", "/home/seafit/code_projects/clang_source/clang/lib/Lex/LiteralSupport.cpp", 1883, __PRETTY_FUNCTION__))" file_link="../../../include/assert.h.html#88" macro="true">assert(!HadError && "This method isn't valid on erroneous strings"); |
1884 | } |
1885 | |
1886 | return SpellingPtr-SpellingStart; |
1887 | } |
1888 | |
1889 | |
1890 | |
1891 | |
1892 | bool StringLiteralParser::isValidUDSuffix(const LangOptions &LangOpts, |
1893 | StringRef Suffix) { |
1894 | return NumericLiteralParser::isValidUDSuffix(LangOpts, Suffix) || |
1895 | Suffix == "sv"; |
1896 | } |
1897 | |