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/Error.h"
29#include "llvm/Support/ErrorHandling.h"
30#include "llvm/Support/Unicode.h"
42 default: llvm_unreachable(
"Unknown token type!");
43 case tok::char_constant:
44 case tok::string_literal:
45 case tok::utf8_char_constant:
46 case tok::utf8_string_literal:
47 return Target.getCharWidth();
48 case tok::wide_char_constant:
49 case tok::wide_string_literal:
50 return Target.getWCharWidth();
51 case tok::utf16_char_constant:
52 case tok::utf16_string_literal:
53 return Target.getChar16Width();
54 case tok::utf32_char_constant:
55 case tok::utf32_string_literal:
56 return Target.getChar32Width();
63 llvm_unreachable(
"Unknown token type!");
64 case tok::char_constant:
65 case tok::string_literal:
67 case tok::utf8_char_constant:
68 case tok::utf8_string_literal:
70 case tok::wide_char_constant:
71 case tok::wide_string_literal:
72 case tok::utf16_char_constant:
73 case tok::utf16_string_literal:
74 case tok::utf32_char_constant:
75 case tok::utf32_string_literal:
83 const char *TokRangeBegin,
84 const char *TokRangeEnd) {
101 const char *TokBegin,
const char *TokRangeBegin,
102 const char *TokRangeEnd,
unsigned DiagID) {
131 const char *&ThisTokBuf,
132 const char *ThisTokEnd,
bool &HadError,
137 const char *EscapeBegin = ThisTokBuf;
138 bool Delimited =
false;
139 bool EndDelimiterFound =
false;
146 unsigned ResultChar = *ThisTokBuf++;
147 char Escape = ResultChar;
148 switch (ResultChar) {
150 case '\\':
case '\'':
case '"':
case '?':
break;
162 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
163 diag::ext_nonstandard_escape) <<
"e";
168 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
169 diag::ext_nonstandard_escape) <<
"E";
189 if (ThisTokBuf != ThisTokEnd && *ThisTokBuf ==
'{') {
192 if (*ThisTokBuf ==
'}') {
193 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
194 diag::err_delimited_escape_empty);
197 }
else if (ThisTokBuf == ThisTokEnd || !
isHexDigit(*ThisTokBuf)) {
199 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
200 diag::err_hex_escape_no_digits) <<
"x";
205 bool Overflow =
false;
206 for (; ThisTokBuf != ThisTokEnd; ++ThisTokBuf) {
207 if (Delimited && *ThisTokBuf ==
'}') {
209 EndDelimiterFound =
true;
212 int CharVal = llvm::hexDigitValue(*ThisTokBuf);
219 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
220 diag::err_delimited_escape_invalid)
221 << StringRef(ThisTokBuf, 1);
225 if (ResultChar & 0xF0000000)
228 ResultChar |= CharVal;
231 if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) {
233 ResultChar &= ~0
U >> (32-CharWidth);
237 if (!HadError && Overflow) {
240 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
241 diag::err_escape_too_large)
246 case '0':
case '1':
case '2':
case '3':
247 case '4':
case '5':
case '6':
case '7': {
254 unsigned NumDigits = 0;
257 ResultChar |= *ThisTokBuf++ -
'0';
259 }
while (ThisTokBuf != ThisTokEnd && NumDigits < 3 &&
260 ThisTokBuf[0] >=
'0' && ThisTokBuf[0] <=
'7');
263 if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) {
265 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
266 diag::err_escape_too_large) << 1;
267 ResultChar &= ~0
U >> (32-CharWidth);
272 bool Overflow =
false;
273 if (ThisTokBuf == ThisTokEnd || *ThisTokBuf !=
'{') {
276 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
277 diag::err_delimited_escape_missing_brace)
285 if (*ThisTokBuf ==
'}') {
286 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
287 diag::err_delimited_escape_empty);
291 while (ThisTokBuf != ThisTokEnd) {
292 if (*ThisTokBuf ==
'}') {
293 EndDelimiterFound =
true;
297 if (*ThisTokBuf <
'0' || *ThisTokBuf >
'7') {
300 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
301 diag::err_delimited_escape_invalid)
302 << StringRef(ThisTokBuf, 1);
307 if (ResultChar & 0xE0000000)
311 ResultChar |= *ThisTokBuf++ -
'0';
315 (Overflow || (CharWidth != 32 && (ResultChar >> CharWidth) != 0))) {
318 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
319 diag::err_escape_too_large)
321 ResultChar &= ~0
U >> (32 - CharWidth);
326 case '(':
case '{':
case '[':
case '%':
329 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
330 diag::ext_nonstandard_escape)
331 << std::string(1, ResultChar);
338 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
339 diag::ext_unknown_escape)
340 << std::string(1, ResultChar);
342 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
343 diag::ext_unknown_escape)
344 <<
"x" + llvm::utohexstr(ResultChar);
348 if (Delimited && Diags) {
349 if (!EndDelimiterFound)
350 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
353 else if (!HadError) {
354 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
355 Features.CPlusPlus23 ? diag::warn_cxx23_delimited_escape_sequence
356 : diag::ext_delimited_escape_sequence)
357 << 0 << (Features.CPlusPlus ? 1 : 0);
361 if (EvalMethod == StringLiteralEvalMethod::Unevaluated &&
363 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
364 diag::err_unevaluated_string_invalid_escape_sequence)
365 << StringRef(EscapeBegin, ThisTokBuf - EscapeBegin);
375 char *ResultPtr = ResultBuf;
376 if (llvm::ConvertCodePointToUTF8(Codepoint, ResultPtr))
377 Str.append(ResultBuf, ResultPtr);
381 for (StringRef::iterator I = Input.begin(), E = Input.end(); I != E; ++I) {
391 assert(Kind ==
'u' || Kind ==
'U' || Kind ==
'N');
392 uint32_t CodePoint = 0;
394 if (Kind ==
'u' && *I ==
'{') {
395 for (++I; *I !=
'}'; ++I) {
396 unsigned Value = llvm::hexDigitValue(*I);
397 assert(
Value != -1U);
408 auto Delim = std::find(I, Input.end(),
'}');
409 assert(Delim != Input.end());
410 StringRef Name(I, std::distance(I, Delim));
411 std::optional<llvm::sys::unicode::LooseMatchingResult> Res =
412 llvm::sys::unicode::nameToCodepointLooseMatching(Name);
413 assert(Res &&
"could not find a codepoint that was previously found");
414 CodePoint = Res->CodePoint;
415 assert(CodePoint != 0xFFFFFFFF);
421 unsigned NumHexDigits;
427 assert(I + NumHexDigits <= E);
429 for (; NumHexDigits != 0; ++I, --NumHexDigits) {
430 unsigned Value = llvm::hexDigitValue(*I);
431 assert(
Value != -1U);
444 return LO.MicrosoftExt &&
445 (K == tok::kw___FUNCTION__ || K == tok::kw_L__FUNCTION__ ||
446 K == tok::kw___FUNCSIG__ || K == tok::kw_L__FUNCSIG__ ||
447 K == tok::kw___FUNCDNAME__);
456 const char *&ThisTokBuf,
457 const char *ThisTokEnd, uint32_t &UcnVal,
458 unsigned short &UcnLen,
bool &Delimited,
461 bool in_char_string_literal =
false) {
462 const char *UcnBegin = ThisTokBuf;
463 bool HasError =
false;
464 bool EndDelimiterFound =
false;
469 if (UcnBegin[1] ==
'u' && in_char_string_literal &&
470 ThisTokBuf != ThisTokEnd && *ThisTokBuf ==
'{') {
473 }
else if (ThisTokBuf == ThisTokEnd || !
isHexDigit(*ThisTokBuf)) {
475 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
476 diag::err_hex_escape_no_digits)
477 << StringRef(&ThisTokBuf[-1], 1);
480 UcnLen = (ThisTokBuf[-1] ==
'u' ? 4 : 8);
482 bool Overflow =
false;
483 unsigned short Count = 0;
484 for (; ThisTokBuf != ThisTokEnd && (Delimited || Count != UcnLen);
486 if (Delimited && *ThisTokBuf ==
'}') {
488 EndDelimiterFound =
true;
491 int CharVal = llvm::hexDigitValue(*ThisTokBuf);
497 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
498 diag::err_delimited_escape_invalid)
499 << StringRef(ThisTokBuf, 1);
504 if (UcnVal & 0xF0000000) {
515 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
516 diag::err_escape_too_large)
521 if (Delimited && !EndDelimiterFound) {
523 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
531 if (Count == 0 || (!Delimited && Count != UcnLen)) {
533 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
534 Delimited ? diag::err_delimited_escape_empty
535 : diag::err_ucn_escape_incomplete);
543 const char *TokBegin,
const char *TokRangeBegin,
const char *TokRangeEnd,
544 llvm::StringRef Name) {
546 Diag(Diags, Features, Loc, TokBegin, TokRangeBegin, TokRangeEnd,
547 diag::err_invalid_ucn_name)
550 namespace u = llvm::sys::unicode;
552 std::optional<u::LooseMatchingResult> Res =
553 u::nameToCodepointLooseMatching(Name);
555 Diag(Diags, Features, Loc, TokBegin, TokRangeBegin, TokRangeEnd,
556 diag::note_invalid_ucn_name_loose_matching)
564 unsigned Distance = 0;
566 u::nearestMatchesForCodepointName(Name, 5);
567 assert(!Matches.empty() &&
"No unicode characters found");
569 for (
const auto &Match : Matches) {
571 Distance = Match.Distance;
572 if (std::max(Distance, Match.Distance) -
573 std::min(Distance, Match.Distance) >
576 Distance = Match.Distance;
579 llvm::UTF32
V = Match.Value;
583 assert(Converted &&
"Found a match wich is not a unicode character");
585 Diag(Diags, Features, Loc, TokBegin, TokRangeBegin, TokRangeEnd,
586 diag::note_invalid_ucn_name_candidate)
587 << Match.Name << llvm::utohexstr(Match.Value)
597 const char *&ThisTokBuf,
598 const char *ThisTokEnd, uint32_t &UcnVal,
602 const char *UcnBegin = ThisTokBuf;
603 assert(UcnBegin[0] ==
'\\' && UcnBegin[1] ==
'N');
605 if (ThisTokBuf == ThisTokEnd || *ThisTokBuf !=
'{') {
607 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
608 diag::err_delimited_escape_missing_brace)
609 << StringRef(&ThisTokBuf[-1], 1);
614 const char *ClosingBrace = std::find_if(ThisTokBuf, ThisTokEnd, [](
char C) {
618 bool Empty = ClosingBrace == ThisTokBuf;
621 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
623 : diag::err_delimited_escape_empty)
624 << StringRef(&UcnBegin[1], 1);
626 ThisTokBuf = ClosingBrace == ThisTokEnd ? ClosingBrace : ClosingBrace + 1;
629 StringRef Name(ThisTokBuf, ClosingBrace - ThisTokBuf);
630 ThisTokBuf = ClosingBrace + 1;
631 std::optional<char32_t> Res = llvm::sys::unicode::nameToCodepointStrict(Name);
635 &UcnBegin[3], ClosingBrace, Name);
639 UcnLen = UcnVal > 0xFFFF ? 8 : 4;
646 const char *ThisTokEnd, uint32_t &UcnVal,
650 bool in_char_string_literal =
false) {
653 const char *UcnBegin = ThisTokBuf;
654 bool IsDelimitedEscapeSequence =
false;
655 bool IsNamedEscapeSequence =
false;
656 if (ThisTokBuf[1] ==
'N') {
657 IsNamedEscapeSequence =
true;
659 UcnVal, UcnLen, Loc, Diags, Features);
663 UcnLen, IsDelimitedEscapeSequence, Loc, Diags,
664 Features, in_char_string_literal);
670 if ((0xD800 <= UcnVal && UcnVal <= 0xDFFF) ||
673 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
674 diag::err_ucn_escape_invalid);
682 (UcnVal != 0x24 && UcnVal != 0x40 && UcnVal != 0x60)) {
684 (!(Features.CPlusPlus11 || Features.C23) || !in_char_string_literal);
686 char BasicSCSChar = UcnVal;
687 if (UcnVal >= 0x20 && UcnVal < 0x7f)
688 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
689 IsError ? diag::err_ucn_escape_basic_scs
691 ? diag::warn_cxx98_compat_literal_ucn_escape_basic_scs
692 : diag::warn_c23_compat_literal_ucn_escape_basic_scs)
693 << StringRef(&BasicSCSChar, 1);
695 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
696 IsError ? diag::err_ucn_control_character
698 ? diag::warn_cxx98_compat_literal_ucn_control_character
699 : diag::warn_c23_compat_literal_ucn_control_character);
705 if (!Features.CPlusPlus && !Features.C99 && Diags)
706 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
707 diag::warn_ucn_not_valid_in_c89_literal);
709 if ((IsDelimitedEscapeSequence || IsNamedEscapeSequence) && Diags)
710 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
711 Features.CPlusPlus23 ? diag::warn_cxx23_delimited_escape_sequence
712 : diag::ext_delimited_escape_sequence)
713 << (IsNamedEscapeSequence ? 1 : 0) << (Features.CPlusPlus ? 1 : 0);
721 const char *ThisTokEnd,
unsigned CharByteWidth,
724 if (CharByteWidth == 4)
728 unsigned short UcnLen = 0;
732 UcnLen, Loc,
nullptr, Features,
true)) {
738 if (CharByteWidth == 2)
739 return UcnVal <= 0xFFFF ? 2 : 4;
746 if (UcnVal < 0x10000)
756 const char *ThisTokEnd,
757 char *&ResultBuf,
bool &HadError,
761 typedef uint32_t
UTF32;
763 unsigned short UcnLen = 0;
765 Loc, Diags, Features,
true)) {
770 assert((CharByteWidth == 1 || CharByteWidth == 2 || CharByteWidth == 4) &&
771 "only character widths of 1, 2, or 4 bytes supported");
774 assert((UcnLen== 4 || UcnLen== 8) &&
"only ucn length of 4 or 8 supported");
776 if (CharByteWidth == 4) {
779 llvm::UTF32 *ResultPtr =
reinterpret_cast<llvm::UTF32*
>(ResultBuf);
785 if (CharByteWidth == 2) {
788 llvm::UTF16 *ResultPtr =
reinterpret_cast<llvm::UTF16*
>(ResultBuf);
790 if (UcnVal <= (
UTF32)0xFFFF) {
798 *ResultPtr = 0xD800 + (UcnVal >> 10);
799 *(ResultPtr+1) = 0xDC00 + (UcnVal & 0x3FF);
804 assert(CharByteWidth == 1 &&
"UTF-8 encoding is only for 1 byte characters");
810 typedef uint8_t
UTF8;
812 unsigned short bytesToWrite = 0;
813 if (UcnVal < (
UTF32)0x80)
815 else if (UcnVal < (
UTF32)0x800)
817 else if (UcnVal < (
UTF32)0x10000)
822 const unsigned byteMask = 0xBF;
823 const unsigned byteMark = 0x80;
827 static const UTF8 firstByteMark[5] = {
828 0x00, 0x00, 0xC0, 0xE0, 0xF0
831 ResultBuf += bytesToWrite;
832 switch (bytesToWrite) {
834 *--ResultBuf = (
UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
837 *--ResultBuf = (
UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
840 *--ResultBuf = (
UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
843 *--ResultBuf = (
UTF8) (UcnVal | firstByteMark[bytesToWrite]);
846 ResultBuf += bytesToWrite;
906 :
SM(
SM), LangOpts(LangOpts), Diags(Diags),
907 ThisTokBegin(TokSpelling.begin()), ThisTokEnd(TokSpelling.end()) {
909 s = DigitsBegin = ThisTokBegin;
910 saw_exponent =
false;
912 saw_ud_suffix =
false;
913 saw_fixed_point_suffix =
false;
937 !(LangOpts.HLSL && *ThisTokEnd ==
'.')) {
938 Diags.
Report(TokLoc, diag::err_lexing_numeric);
944 ParseNumberStartingWithZero(TokLoc);
950 if (
s == ThisTokEnd) {
953 ParseDecimalOrOctalCommon(TokLoc);
960 checkSeparator(TokLoc,
s, CSK_AfterDigits);
963 if (LangOpts.FixedPoint) {
964 for (
const char *
c =
s;
c != ThisTokEnd; ++
c) {
965 if (*
c ==
'r' || *
c ==
'k' || *
c ==
'R' || *
c ==
'K') {
966 saw_fixed_point_suffix =
true;
976 bool HasSize =
false;
977 bool DoubleUnderscore =
false;
981 for (;
s != ThisTokEnd; ++
s) {
985 if (!LangOpts.FixedPoint)
988 if (!(saw_period || saw_exponent))
break;
993 if (!LangOpts.FixedPoint)
996 if (!(saw_period || saw_exponent))
break;
1002 if (!(LangOpts.Half || LangOpts.FixedPoint))
1012 if (!isFPConstant)
break;
1023 if ((
Target.hasFloat16Type() || LangOpts.CUDA ||
1024 (LangOpts.OpenMPIsTargetDevice &&
Target.getTriple().isNVPTX())) &&
1025 s + 2 < ThisTokEnd &&
s[1] ==
'1' &&
s[2] ==
'6') {
1035 if (!isFPConstant)
break;
1043 if (isFPConstant)
break;
1055 assert(
s + 1 < ThisTokEnd &&
"didn't maximally munch?");
1056 if (isFPConstant)
break;
1074 if (LangOpts.MicrosoftExt && !isFPConstant) {
1111 assert(
s <= ThisTokEnd &&
"didn't maximally munch?");
1126 if (DoubleUnderscore)
1128 if (LangOpts.CPlusPlus &&
s + 2 < ThisTokEnd &&
1131 DoubleUnderscore =
true;
1133 if (
s + 1 < ThisTokEnd &&
1134 (*
s ==
'u' || *
s ==
'U')) {
1138 if (
s + 1 < ThisTokEnd &&
1139 ((*
s ==
'w' && *(++
s) ==
'b') || (*
s ==
'W' && *(++
s) ==
'B'))) {
1157 if ((!LangOpts.CPlusPlus || DoubleUnderscore) &&
s + 1 < ThisTokEnd &&
1158 ((
s[0] ==
'w' &&
s[1] ==
'b') || (
s[0] ==
'W' &&
s[1] ==
'B'))) {
1172 expandUCNs(UDSuffixBuf, StringRef(SuffixBegin, ThisTokEnd - SuffixBegin));
1187 saw_fixed_point_suffix =
false;
1192 saw_ud_suffix =
true;
1196 if (
s != ThisTokEnd) {
1199 TokLoc, SuffixBegin - ThisTokBegin,
SM, LangOpts),
1200 diag::err_invalid_suffix_constant)
1201 << StringRef(SuffixBegin, ThisTokEnd - SuffixBegin)
1202 << (isFixedPointConstant ? 2 : isFPConstant);
1207 if (!
hadError && saw_fixed_point_suffix) {
1215void NumericLiteralParser::ParseDecimalOrOctalCommon(
SourceLocation TokLoc){
1216 assert((radix == 8 || radix == 10) &&
"Unexpected radix");
1224 diag::err_invalid_digit)
1225 << StringRef(
s, 1) << (radix == 8 ? 1 : 0);
1231 checkSeparator(TokLoc,
s, CSK_AfterDigits);
1235 checkSeparator(TokLoc,
s, CSK_BeforeDigits);
1238 if (*
s ==
'e' || *
s ==
'E') {
1239 checkSeparator(TokLoc,
s, CSK_AfterDigits);
1240 const char *Exponent =
s;
1243 saw_exponent =
true;
1244 if (
s != ThisTokEnd && (*
s ==
'+' || *
s ==
'-'))
s++;
1245 const char *first_non_digit = SkipDigits(
s);
1246 if (containsDigits(
s, first_non_digit)) {
1247 checkSeparator(TokLoc,
s, CSK_BeforeDigits);
1248 s = first_non_digit;
1252 TokLoc, Exponent - ThisTokBegin,
SM, LangOpts),
1253 diag::err_exponent_has_no_digits);
1266 if (!LangOpts.CPlusPlus11 || Suffix.empty())
1272 if (Suffix.starts_with(
"_") && !Suffix.starts_with(
"__"))
1276 if (!LangOpts.CPlusPlus14)
1282 return llvm::StringSwitch<bool>(Suffix)
1283 .Cases(
"h",
"min",
"s",
true)
1284 .Cases(
"ms",
"us",
"ns",
true)
1285 .Cases(
"il",
"i",
"if",
true)
1286 .Cases(
"d",
"y", LangOpts.CPlusPlus20)
1292 CheckSeparatorKind IsAfterDigits) {
1293 if (IsAfterDigits == CSK_AfterDigits) {
1294 if (Pos == ThisTokBegin)
1297 }
else if (Pos == ThisTokEnd)
1300 if (isDigitSeparator(*Pos)) {
1303 diag::err_digit_separator_not_between_digits)
1314void NumericLiteralParser::ParseNumberStartingWithZero(
SourceLocation TokLoc) {
1315 assert(
s[0] ==
'0' &&
"Invalid method call");
1321 if ((c1 ==
'x' || c1 ==
'X') && (
isHexDigit(
s[1]) ||
s[1] ==
'.')) {
1323 assert(
s < ThisTokEnd &&
"didn't maximally munch?");
1326 s = SkipHexDigits(
s);
1327 bool HasSignificandDigits = containsDigits(DigitsBegin,
s);
1328 if (
s == ThisTokEnd) {
1330 }
else if (*
s ==
'.') {
1333 const char *floatDigitsBegin =
s;
1334 s = SkipHexDigits(
s);
1335 if (containsDigits(floatDigitsBegin,
s))
1336 HasSignificandDigits =
true;
1337 if (HasSignificandDigits)
1338 checkSeparator(TokLoc, floatDigitsBegin, CSK_BeforeDigits);
1341 if (!HasSignificandDigits) {
1344 diag::err_hex_constant_requires)
1345 << LangOpts.CPlusPlus << 1;
1352 if (*
s ==
'p' || *
s ==
'P') {
1353 checkSeparator(TokLoc,
s, CSK_AfterDigits);
1354 const char *Exponent =
s;
1356 saw_exponent =
true;
1357 if (
s != ThisTokEnd && (*
s ==
'+' || *
s ==
'-'))
s++;
1358 const char *first_non_digit = SkipDigits(
s);
1359 if (!containsDigits(
s, first_non_digit)) {
1362 TokLoc, Exponent - ThisTokBegin,
SM, LangOpts),
1363 diag::err_exponent_has_no_digits);
1368 checkSeparator(TokLoc,
s, CSK_BeforeDigits);
1369 s = first_non_digit;
1371 if (!LangOpts.HexFloats)
1372 Diags.
Report(TokLoc, LangOpts.CPlusPlus
1373 ? diag::ext_hex_literal_invalid
1374 : diag::ext_hex_constant_invalid);
1375 else if (LangOpts.CPlusPlus17)
1376 Diags.
Report(TokLoc, diag::warn_cxx17_hex_literal);
1377 }
else if (saw_period) {
1380 diag::err_hex_constant_requires)
1381 << LangOpts.CPlusPlus << 0;
1388 if ((c1 ==
'b' || c1 ==
'B') && (
s[1] ==
'0' ||
s[1] ==
'1')) {
1391 if (LangOpts.CPlusPlus14)
1392 DiagId = diag::warn_cxx11_compat_binary_literal;
1393 else if (LangOpts.C23)
1394 DiagId = diag::warn_c23_compat_binary_literal;
1395 else if (LangOpts.CPlusPlus)
1396 DiagId = diag::ext_binary_literal_cxx14;
1398 DiagId = diag::ext_binary_literal;
1399 Diags.
Report(TokLoc, DiagId);
1401 assert(
s < ThisTokEnd &&
"didn't maximally munch?");
1404 s = SkipBinaryDigits(
s);
1405 if (
s == ThisTokEnd) {
1411 diag::err_invalid_digit)
1412 << StringRef(
s, 1) << 2;
1423 const char *PossibleNewDigitStart =
s;
1424 s = SkipOctalDigits(
s);
1428 if (
s != PossibleNewDigitStart)
1429 DigitsBegin = PossibleNewDigitStart;
1431 if (
s == ThisTokEnd)
1437 const char *EndDecimal = SkipDigits(
s);
1438 if (EndDecimal[0] ==
'.' || EndDecimal[0] ==
'e' || EndDecimal[0] ==
'E') {
1444 ParseDecimalOrOctalCommon(TokLoc);
1450 return NumDigits <= 64;
1452 return NumDigits <= 64 / 3;
1454 return NumDigits <= 19;
1456 return NumDigits <= 64 / 4;
1458 llvm_unreachable(
"impossible Radix");
1472 const unsigned NumDigits = SuffixBegin - DigitsBegin;
1475 for (
const char *Ptr = DigitsBegin; Ptr != SuffixBegin; ++Ptr)
1476 if (!isDigitSeparator(*Ptr))
1477 N = N * radix + llvm::hexDigitValue(*Ptr);
1482 return Val.getZExtValue() != N;
1486 const char *Ptr = DigitsBegin;
1488 llvm::APInt RadixVal(Val.getBitWidth(), radix);
1489 llvm::APInt CharVal(Val.getBitWidth(), 0);
1490 llvm::APInt OldVal = Val;
1492 bool OverflowOccurred =
false;
1493 while (Ptr < SuffixBegin) {
1494 if (isDigitSeparator(*Ptr)) {
1499 unsigned C = llvm::hexDigitValue(*Ptr++);
1502 assert(
C < radix &&
"NumericLiteralParser ctor should have rejected this");
1512 OverflowOccurred |= Val.udiv(RadixVal) != OldVal;
1517 OverflowOccurred |= Val.ult(CharVal);
1519 return OverflowOccurred;
1522llvm::APFloat::opStatus
1524 using llvm::APFloat;
1526 unsigned n = std::min(SuffixBegin - ThisTokBegin, ThisTokEnd - ThisTokBegin);
1529 StringRef Str(ThisTokBegin, n);
1530 if (Str.contains(
'\'')) {
1532 std::remove_copy_if(Str.begin(), Str.end(), std::back_inserter(Buffer),
1538 Result.convertFromString(Str, APFloat::rmNearestTiesToEven);
1539 assert(StatusOrErr &&
"Invalid floating point representation");
1540 return !errorToBool(StatusOrErr.takeError()) ? *StatusOrErr
1541 : APFloat::opInvalidOp;
1546 return c ==
'p' ||
c ==
'P';
1547 return c ==
'e' ||
c ==
'E';
1551 assert(radix == 16 || radix == 10);
1554 unsigned NumDigits = SuffixBegin - DigitsBegin;
1555 if (saw_period) --NumDigits;
1558 bool ExpOverflowOccurred =
false;
1559 bool NegativeExponent =
false;
1560 const char *ExponentBegin;
1561 uint64_t Exponent = 0;
1562 int64_t BaseShift = 0;
1564 const char *Ptr = DigitsBegin;
1568 ExponentBegin = Ptr;
1570 NegativeExponent = *Ptr ==
'-';
1571 if (NegativeExponent) ++Ptr;
1573 unsigned NumExpDigits = SuffixBegin - Ptr;
1575 llvm::StringRef ExpStr(Ptr, NumExpDigits);
1576 llvm::APInt ExpInt(64, ExpStr, 10);
1577 Exponent = ExpInt.getZExtValue();
1579 ExpOverflowOccurred =
true;
1582 if (NegativeExponent) BaseShift -= Exponent;
1583 else BaseShift += Exponent;
1603 uint64_t NumBitsNeeded;
1605 NumBitsNeeded = 4 * (NumDigits + Exponent) + Scale;
1607 NumBitsNeeded = 4 * NumDigits + Exponent + Scale;
1609 if (NumBitsNeeded > std::numeric_limits<unsigned>::max())
1610 ExpOverflowOccurred =
true;
1611 llvm::APInt Val(
static_cast<unsigned>(NumBitsNeeded), 0,
false);
1613 bool FoundDecimal =
false;
1615 int64_t FractBaseShift = 0;
1616 const char *End = saw_exponent ? ExponentBegin : SuffixBegin;
1617 for (
const char *Ptr = DigitsBegin; Ptr < End; ++Ptr) {
1619 FoundDecimal =
true;
1624 unsigned C = llvm::hexDigitValue(*Ptr);
1625 assert(
C < radix &&
"NumericLiteralParser ctor should have rejected this");
1637 if (radix == 16) FractBaseShift *= 4;
1638 BaseShift += FractBaseShift;
1642 uint64_t
Base = (radix == 16) ? 2 : 10;
1643 if (BaseShift > 0) {
1644 for (int64_t i = 0; i < BaseShift; ++i) {
1647 }
else if (BaseShift < 0) {
1648 for (int64_t i = BaseShift; i < 0 && !Val.isZero(); ++i)
1649 Val = Val.udiv(
Base);
1652 bool IntOverflowOccurred =
false;
1653 auto MaxVal = llvm::APInt::getMaxValue(StoreVal.getBitWidth());
1654 if (Val.getBitWidth() > StoreVal.getBitWidth()) {
1655 IntOverflowOccurred |= Val.ugt(MaxVal.zext(Val.getBitWidth()));
1656 StoreVal = Val.trunc(StoreVal.getBitWidth());
1657 }
else if (Val.getBitWidth() < StoreVal.getBitWidth()) {
1658 IntOverflowOccurred |= Val.zext(MaxVal.getBitWidth()).ugt(MaxVal);
1659 StoreVal = Val.zext(StoreVal.getBitWidth());
1664 return IntOverflowOccurred || ExpOverflowOccurred;
1714 const char *TokBegin = begin;
1717 if (Kind != tok::char_constant)
1719 if (Kind == tok::utf8_char_constant)
1723 if (begin[0] !=
'\'') {
1724 PP.
Diag(Loc, diag::err_lexing_char);
1732 if (end[-1] !=
'\'') {
1733 const char *UDSuffixEnd = end;
1736 }
while (end[-1] !=
'\'');
1738 expandUCNs(UDSuffixBuf, StringRef(end, UDSuffixEnd - end));
1739 UDSuffixOffset = end - TokBegin;
1743 assert(end != begin &&
"Invalid token lexed");
1750 "Assumes char is 8 bits");
1753 "Assumes sizeof(int) on target is <= 64 and a multiple of char");
1755 "Assumes sizeof(wchar) on target is <= 64");
1758 codepoint_buffer.resize(end - begin);
1759 uint32_t *buffer_begin = &codepoint_buffer.front();
1760 uint32_t *buffer_end = buffer_begin + codepoint_buffer.size();
1765 uint32_t largest_character_for_kind;
1766 if (tok::wide_char_constant == Kind) {
1767 largest_character_for_kind =
1769 }
else if (tok::utf8_char_constant == Kind) {
1770 largest_character_for_kind = 0x7F;
1771 }
else if (tok::utf16_char_constant == Kind) {
1772 largest_character_for_kind = 0xFFFF;
1773 }
else if (tok::utf32_char_constant == Kind) {
1774 largest_character_for_kind = 0x10FFFF;
1776 largest_character_for_kind = 0x7Fu;
1779 while (begin != end) {
1781 if (begin[0] !=
'\\') {
1782 char const *start = begin;
1785 }
while (begin != end && *begin !=
'\\');
1787 char const *tmp_in_start = start;
1788 uint32_t *tmp_out_start = buffer_begin;
1789 llvm::ConversionResult res =
1790 llvm::ConvertUTF8toUTF32(
reinterpret_cast<llvm::UTF8
const **
>(&start),
1791 reinterpret_cast<llvm::UTF8
const *
>(begin),
1792 &buffer_begin, buffer_end, llvm::strictConversion);
1793 if (res != llvm::conversionOK) {
1798 unsigned Msg = diag::err_bad_character_encoding;
1799 if (NoErrorOnBadEncoding)
1800 Msg = diag::warn_bad_character_encoding;
1802 if (NoErrorOnBadEncoding) {
1803 start = tmp_in_start;
1804 buffer_begin = tmp_out_start;
1805 for (; start != begin; ++start, ++buffer_begin)
1806 *buffer_begin =
static_cast<uint8_t
>(*start);
1811 for (; tmp_out_start < buffer_begin; ++tmp_out_start) {
1812 if (*tmp_out_start > largest_character_for_kind) {
1814 PP.
Diag(Loc, diag::err_character_too_large);
1822 if (begin[1] ==
'u' || begin[1] ==
'U' || begin[1] ==
'N') {
1823 unsigned short UcnLen = 0;
1828 }
else if (*buffer_begin > largest_character_for_kind) {
1830 PP.
Diag(Loc, diag::err_character_too_large);
1842 *buffer_begin++ = result;
1845 unsigned NumCharsSoFar = buffer_begin - &codepoint_buffer.front();
1847 if (NumCharsSoFar > 1) {
1849 PP.
Diag(Loc, diag::warn_four_char_character_literal);
1851 PP.
Diag(Loc, diag::warn_multichar_character_literal);
1853 PP.
Diag(Loc, diag::err_multichar_character_literal) << (
isWide() ? 0 : 1);
1858 IsMultiChar =
false;
1865 bool multi_char_too_long =
false;
1868 for (
size_t i = 0; i < NumCharsSoFar; ++i) {
1870 multi_char_too_long |= (LitVal.countl_zero() < 8);
1872 LitVal = LitVal + (codepoint_buffer[i] & 0xFF);
1874 }
else if (NumCharsSoFar > 0) {
1876 LitVal = buffer_begin[-1];
1879 if (!HadError && multi_char_too_long) {
1880 PP.
Diag(Loc, diag::warn_char_constant_too_large);
1884 Value = LitVal.getZExtValue();
1952 :
SM(PP.getSourceManager()), Features(PP.getLangOpts()),
1953 Target(PP.getTargetInfo()), Diags(&PP.getDiagnostics()),
1954 MaxTokenLength(0), SizeBound(0), CharByteWidth(0), Kind(tok::unknown),
1955 ResultPtr(ResultBuf.data()), EvalMethod(EvalMethod), hadError(
false),
1963 if (StringToks.empty() || StringToks[0].getLength() < 2)
1970 assert(!StringToks.empty() &&
"expected at least one token");
1971 MaxTokenLength = StringToks[0].getLength();
1972 assert(StringToks[0].getLength() >= 2 &&
"literal token is invalid!");
1973 SizeBound = StringToks[0].getLength() - 2;
1977 Kind = tok::string_literal;
1980 for (
const Token &Tok : StringToks) {
1981 if (Tok.getLength() < 2)
1982 return DiagnoseLexingError(Tok.getLocation());
1986 assert(Tok.getLength() >= 2 &&
"literal token is invalid!");
1987 SizeBound += Tok.getLength() - 2;
1990 if (Tok.getLength() > MaxTokenLength)
1991 MaxTokenLength = Tok.getLength();
1995 if (
isUnevaluated() && Tok.getKind() != tok::string_literal) {
2002 StringRef Prefix(
SM.getCharacterData(Tok.getLocation()),
2004 Diags->
Report(Tok.getLocation(),
2005 Features.CPlusPlus26
2006 ? diag::err_unevaluated_string_prefix
2007 : diag::warn_unevaluated_string_prefix)
2010 if (Features.CPlusPlus26)
2012 }
else if (Tok.isNot(Kind) && Tok.isNot(tok::string_literal)) {
2014 Kind = Tok.getKind();
2017 Diags->
Report(Tok.getLocation(), diag::err_unsupported_string_concat);
2030 assert((CharByteWidth & 7) == 0 &&
"Assumes character size is byte multiple");
2035 SizeBound *= CharByteWidth;
2038 ResultBuf.resize(SizeBound);
2042 TokenBuf.resize(MaxTokenLength);
2046 ResultPtr = &ResultBuf[0];
2052 for (
unsigned i = 0, e = StringToks.size(); i != e; ++i) {
2053 const char *ThisTokBuf = &TokenBuf[0];
2057 bool StringInvalid =
false;
2058 unsigned ThisTokLen =
2062 return DiagnoseLexingError(StringToks[i].getLocation());
2064 const char *ThisTokBegin = ThisTokBuf;
2065 const char *ThisTokEnd = ThisTokBuf+ThisTokLen;
2068 if (ThisTokEnd[-1] !=
'"') {
2069 const char *UDSuffixEnd = ThisTokEnd;
2072 }
while (ThisTokEnd[-1] !=
'"');
2074 StringRef UDSuffix(ThisTokEnd, UDSuffixEnd - ThisTokEnd);
2076 if (UDSuffixBuf.empty()) {
2077 if (StringToks[i].hasUCN())
2080 UDSuffixBuf.assign(UDSuffix);
2082 UDSuffixOffset = ThisTokEnd - ThisTokBuf;
2083 UDSuffixTokLoc = StringToks[i].getLocation();
2086 if (StringToks[i].hasUCN()) {
2088 UDSuffix = ExpandedUDSuffix;
2095 bool UnevaluatedStringHasUDL =
isUnevaluated() && !UDSuffix.empty();
2096 if (UDSuffixBuf != UDSuffix || UnevaluatedStringHasUDL) {
2099 if (UnevaluatedStringHasUDL) {
2100 Diags->
Report(TokLoc, diag::err_unevaluated_string_udl)
2103 Diags->
Report(TokLoc, diag::err_string_concat_mixed_suffix)
2104 << UDSuffixBuf << UDSuffix
2119 if (ThisTokBuf[0] ==
'L' || ThisTokBuf[0] ==
'u' || ThisTokBuf[0] ==
'U') {
2122 if (ThisTokBuf[0] ==
'8')
2127 if (ThisTokBuf[0] ==
'R') {
2128 if (ThisTokBuf[1] !=
'"') {
2131 return DiagnoseLexingError(StringToks[i].getLocation());
2137 constexpr unsigned MaxRawStrDelimLen = 16;
2139 const char *Prefix = ThisTokBuf;
2140 while (
static_cast<unsigned>(ThisTokBuf - Prefix) < MaxRawStrDelimLen &&
2141 ThisTokBuf[0] !=
'(')
2143 if (ThisTokBuf[0] !=
'(')
2144 return DiagnoseLexingError(StringToks[i].getLocation());
2148 ThisTokEnd -= ThisTokBuf - Prefix;
2149 if (ThisTokEnd < ThisTokBuf)
2150 return DiagnoseLexingError(StringToks[i].getLocation());
2154 StringRef RemainingTokenSpan(ThisTokBuf, ThisTokEnd - ThisTokBuf);
2155 while (!RemainingTokenSpan.empty()) {
2157 size_t CRLFPos = RemainingTokenSpan.find(
"\r\n");
2158 StringRef BeforeCRLF = RemainingTokenSpan.substr(0, CRLFPos);
2159 StringRef AfterCRLF = RemainingTokenSpan.substr(CRLFPos);
2162 if (CopyStringFragment(StringToks[i], ThisTokBegin, BeforeCRLF))
2167 RemainingTokenSpan = AfterCRLF.substr(1);
2170 if (ThisTokBuf[0] !=
'"') {
2173 return DiagnoseLexingError(StringToks[i].getLocation());
2179 ThisTokBuf + 1 != ThisTokEnd && ThisTokBuf[0] ==
'\\' &&
2180 ThisTokBuf[1] ==
'p') {
2191 while (ThisTokBuf != ThisTokEnd) {
2193 if (ThisTokBuf[0] !=
'\\') {
2194 const char *InStart = ThisTokBuf;
2197 }
while (ThisTokBuf != ThisTokEnd && ThisTokBuf[0] !=
'\\');
2200 if (CopyStringFragment(StringToks[i], ThisTokBegin,
2201 StringRef(InStart, ThisTokBuf - InStart)))
2206 if (ThisTokBuf[1] ==
'u' || ThisTokBuf[1] ==
'U' ||
2207 ThisTokBuf[1] ==
'N') {
2211 CharByteWidth, Diags, Features);
2215 unsigned ResultChar =
2218 CharByteWidth * 8, Diags, Features, EvalMethod);
2220 if (CharByteWidth == 4) {
2223 llvm::UTF32 *ResultWidePtr =
reinterpret_cast<llvm::UTF32*
>(ResultPtr);
2224 *ResultWidePtr = ResultChar;
2226 }
else if (CharByteWidth == 2) {
2229 llvm::UTF16 *ResultWidePtr =
reinterpret_cast<llvm::UTF16*
>(ResultPtr);
2230 *ResultWidePtr = ResultChar & 0xFFFF;
2233 assert(CharByteWidth == 1 &&
"Unexpected char width");
2234 *ResultPtr++ = ResultChar & 0xFF;
2241 "Pascal string in unevaluated context");
2243 if (CharByteWidth == 4) {
2246 llvm::UTF32 *ResultWidePtr =
reinterpret_cast<llvm::UTF32*
>(ResultBuf.data());
2248 }
else if (CharByteWidth == 2) {
2251 llvm::UTF16 *ResultWidePtr =
reinterpret_cast<llvm::UTF16*
>(ResultBuf.data());
2254 assert(CharByteWidth == 1 &&
"Unexpected char width");
2261 Diags->
Report(StringToks.front().getLocation(),
2262 diag::err_pascal_string_too_long)
2264 StringToks.back().getLocation());
2270 unsigned MaxChars = Features.CPlusPlus? 65536 : Features.C99 ? 4095 : 509;
2273 Diags->
Report(StringToks.front().getLocation(),
2274 diag::ext_string_too_long)
2276 << (Features.CPlusPlus ? 2 : Features.C99 ? 1 : 0)
2278 StringToks.back().getLocation());
2285 End = Err + std::min<unsigned>(llvm::getNumBytesForUTF8(*Err), End-Err);
2286 while (++Err != End && (*Err & 0xC0) == 0x80)
2294bool StringLiteralParser::CopyStringFragment(
const Token &Tok,
2295 const char *TokBegin,
2296 StringRef Fragment) {
2297 const llvm::UTF8 *ErrorPtrTmp;
2298 if (ConvertUTF8toWide(CharByteWidth, Fragment, ResultPtr, ErrorPtrTmp))
2305 if (NoErrorOnBadEncoding) {
2306 memcpy(ResultPtr, Fragment.data(), Fragment.size());
2307 ResultPtr += Fragment.size();
2311 const char *ErrorPtr =
reinterpret_cast<const char *
>(ErrorPtrTmp);
2315 Diag(Diags, Features, SourceLoc, TokBegin,
2316 ErrorPtr,
resyncUTF8(ErrorPtr, Fragment.end()),
2317 NoErrorOnBadEncoding ? diag::warn_bad_string_encoding
2318 : diag::err_bad_string_encoding);
2320 const char *NextStart =
resyncUTF8(ErrorPtr, Fragment.end());
2321 StringRef NextFragment(NextStart, Fragment.end()-NextStart);
2325 Dummy.reserve(Fragment.size() * CharByteWidth);
2326 char *Ptr = Dummy.data();
2328 while (!ConvertUTF8toWide(CharByteWidth, NextFragment, Ptr, ErrorPtrTmp)) {
2329 const char *ErrorPtr =
reinterpret_cast<const char *
>(ErrorPtrTmp);
2330 NextStart =
resyncUTF8(ErrorPtr, Fragment.end());
2332 ErrorPtr, NextStart);
2333 NextFragment = StringRef(NextStart, Fragment.end()-NextStart);
2336 return !NoErrorOnBadEncoding;
2339void StringLiteralParser::DiagnoseLexingError(
SourceLocation Loc) {
2342 Diags->
Report(Loc, diag::err_lexing_string);
2349 unsigned ByteNo)
const {
2354 bool StringInvalid =
false;
2355 const char *SpellingPtr = &SpellingBuffer[0];
2361 const char *SpellingStart = SpellingPtr;
2362 const char *SpellingEnd = SpellingPtr+TokLen;
2365 if (SpellingPtr[0] ==
'u' && SpellingPtr[1] ==
'8')
2368 assert(SpellingPtr[0] !=
'L' && SpellingPtr[0] !=
'u' &&
2369 SpellingPtr[0] !=
'U' &&
"Doesn't handle wide or utf strings yet");
2372 if (SpellingPtr[0] ==
'R') {
2373 assert(SpellingPtr[1] ==
'"' &&
"Should be a raw string literal!");
2376 while (*SpellingPtr !=
'(') {
2378 assert(SpellingPtr < SpellingEnd &&
"Missing ( for raw string literal");
2382 return SpellingPtr - SpellingStart + ByteNo;
2386 assert(SpellingPtr[0] ==
'"' &&
"Should be a string literal!");
2391 assert(SpellingPtr < SpellingEnd &&
"Didn't find byte offset!");
2394 if (*SpellingPtr !=
'\\') {
2401 bool HadError =
false;
2402 if (SpellingPtr[1] ==
'u' || SpellingPtr[1] ==
'U' ||
2403 SpellingPtr[1] ==
'N') {
2404 const char *EscapePtr = SpellingPtr;
2406 1, Features, HadError);
2409 SpellingPtr = EscapePtr;
2419 assert(!HadError &&
"This method isn't valid on erroneous strings");
2422 return SpellingPtr-SpellingStart;
Defines the clang::LangOptions interface.
static void EncodeUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, char *&ResultBuf, bool &HadError, FullSourceLoc Loc, unsigned CharByteWidth, DiagnosticsEngine *Diags, const LangOptions &Features)
EncodeUCNEscape - Read the Universal Character Name, check constraints and convert the UTF32 to UTF8 ...
static bool ProcessUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, uint32_t &UcnVal, unsigned short &UcnLen, FullSourceLoc Loc, DiagnosticsEngine *Diags, const LangOptions &Features, bool in_char_string_literal=false)
ProcessUCNEscape - Read the Universal Character Name, check constraints and return the UTF32.
static CharSourceRange MakeCharSourceRange(const LangOptions &Features, FullSourceLoc TokLoc, const char *TokBegin, const char *TokRangeBegin, const char *TokRangeEnd)
static const char * resyncUTF8(const char *Err, const char *End)
static int MeasureUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, unsigned CharByteWidth, const LangOptions &Features, bool &HadError)
MeasureUCNEscape - Determine the number of bytes within the resulting string which this UCN will occu...
static void appendCodePoint(unsigned Codepoint, llvm::SmallVectorImpl< char > &Str)
static unsigned getEncodingPrefixLen(tok::TokenKind kind)
static void DiagnoseInvalidUnicodeCharacterName(DiagnosticsEngine *Diags, const LangOptions &Features, FullSourceLoc Loc, const char *TokBegin, const char *TokRangeBegin, const char *TokRangeEnd, llvm::StringRef Name)
static DiagnosticBuilder Diag(DiagnosticsEngine *Diags, const LangOptions &Features, FullSourceLoc TokLoc, const char *TokBegin, const char *TokRangeBegin, const char *TokRangeEnd, unsigned DiagID)
Produce a diagnostic highlighting some portion of a literal.
static bool IsEscapeValidInUnevaluatedStringLiteral(char Escape)
static bool ProcessNumericUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, uint32_t &UcnVal, unsigned short &UcnLen, bool &Delimited, FullSourceLoc Loc, DiagnosticsEngine *Diags, const LangOptions &Features, bool in_char_string_literal=false)
static bool ProcessNamedUCNEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, uint32_t &UcnVal, unsigned short &UcnLen, FullSourceLoc Loc, DiagnosticsEngine *Diags, const LangOptions &Features)
static bool IsExponentPart(char c, bool isHex)
static bool alwaysFitsInto64Bits(unsigned Radix, unsigned NumDigits)
static unsigned ProcessCharEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, bool &HadError, FullSourceLoc Loc, unsigned CharWidth, DiagnosticsEngine *Diags, const LangOptions &Features, StringLiteralEvalMethod EvalMethod)
ProcessCharEscape - Parse a standard C escape sequence, which can occur in either a character or a st...
static unsigned getCharWidth(tok::TokenKind kind, const TargetInfo &Target)
llvm::MachO::Target Target
Defines the clang::Preprocessor interface.
Defines the clang::SourceLocation class and associated facilities.
__DEVICE__ void * memcpy(void *__a, const void *__b, size_t __c)
__device__ __2f16 float __ockl_bool s
__device__ __2f16 float c
CharLiteralParser(const char *begin, const char *end, SourceLocation Loc, Preprocessor &PP, tok::TokenKind kind)
Represents a character-granular source range.
static CharSourceRange getCharRange(SourceRange R)
A little helper class used to produce diagnostics.
Concrete class used by the front-end to report problems and issues.
DiagnosticBuilder Report(SourceLocation Loc, unsigned DiagID)
Issue the message to the client.
static FixItHint CreateReplacement(CharSourceRange RemoveRange, StringRef Code)
Create a code modification hint that replaces the given source range with the given code string.
static FixItHint CreateRemoval(CharSourceRange RemoveRange)
Create a code modification hint that removes the given source range.
A SourceLocation and its associated SourceManager.
const SourceManager & getManager() const
Keeps track of the various options that can be enabled, which controls the dialect of C or C++ that i...
static SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart, unsigned Characters, const SourceManager &SM, const LangOptions &LangOpts)
AdvanceToTokenCharacter - If the current SourceLocation specifies a location at the start of a token,...
static unsigned getSpelling(const Token &Tok, const char *&Buffer, const SourceManager &SourceMgr, const LangOptions &LangOpts, bool *Invalid=nullptr)
getSpelling - This method is used to get the spelling of a token into a preallocated buffer,...
NumericLiteralParser(StringRef TokSpelling, SourceLocation TokLoc, const SourceManager &SM, const LangOptions &LangOpts, const TargetInfo &Target, DiagnosticsEngine &Diags)
integer-constant: [C99 6.4.4.1] decimal-constant integer-suffix octal-constant integer-suffix hexadec...
bool isFixedPointLiteral() const
bool isFloatingLiteral() const
bool isIntegerLiteral() const
llvm::APFloat::opStatus GetFloatValue(llvm::APFloat &Result)
GetFloatValue - Convert this numeric literal to a floating value, using the specified APFloat fltSema...
static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix)
Determine whether a suffix is a valid ud-suffix.
bool GetIntegerValue(llvm::APInt &Val)
GetIntegerValue - Convert this numeric literal value to an APInt that matches Val's input width.
bool GetFixedPointValue(llvm::APInt &StoreVal, unsigned Scale)
GetFixedPointValue - Convert this numeric literal value into a scaled integer that represents this va...
Engages in a tight little dance with the lexer to efficiently preprocess tokens.
SourceManager & getSourceManager() const
const TargetInfo & getTargetInfo() const
const LangOptions & getLangOpts() const
DiagnosticsEngine & getDiagnostics() const
DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID) const
Forwarding function for diagnostics.
Encodes a location in the source.
This class handles loading and caching of source files into memory.
A trivial tuple used to represent a source range.
unsigned getOffsetOfStringByte(const Token &TheTok, unsigned ByteNo) const
getOffsetOfStringByte - This function returns the offset of the specified byte of the string data rep...
bool isUnevaluated() const
StringLiteralParser(ArrayRef< Token > StringToks, Preprocessor &PP, StringLiteralEvalMethod StringMethod=StringLiteralEvalMethod::Evaluated)
unsigned GetStringLength() const
static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix)
Determine whether a suffix is a valid ud-suffix.
unsigned GetNumStringChars() const
Exposes information about the current target.
unsigned getIntWidth() const
getIntWidth/Align - Return the size of 'signed int' and 'unsigned int' for this target,...
unsigned getWCharWidth() const
getWCharWidth/Align - Return the size of 'wchar_t' for this target, in bits.
unsigned getCharWidth() const
Token - This structure provides full information about a lexed token.
SourceLocation getLocation() const
Return a source location identifier for the specified offset in the current file.
unsigned getLength() const
tok::TokenKind getKind() const
Defines the clang::TargetInfo interface.
bool isStringLiteral(TokenKind K)
Return true if this is a C or C++ string-literal (or C++11 user-defined-string-literal) token.
TokenKind
Provides a simple uniform namespace for tokens from all C languages.
The JSON file list parser is used to communicate input to InstallAPI.
LLVM_READONLY bool isVerticalWhitespace(unsigned char c)
Returns true if this character is vertical ASCII whitespace: '\n', '\r'.
LLVM_READONLY bool isPrintable(unsigned char c)
Return true if this character is an ASCII printable character; that is, a character that should take ...
void expandUCNs(SmallVectorImpl< char > &Buf, StringRef Input)
Copy characters from Input to Buf, expanding any UCNs.
bool tokenIsLikeStringLiteral(const Token &Tok, const LangOptions &LO)
Return true if the token is a string literal, or a function local predefined macro,...
@ Result
The result type of a method or function.
LLVM_READONLY bool isDigit(unsigned char c)
Return true if this character is an ASCII digit: [0-9].
bool isFunctionLocalStringLiteralMacro(tok::TokenKind K, const LangOptions &LO)
Return true if the token corresponds to a function local predefined macro, which expands to a string ...
LLVM_READONLY bool isPreprocessingNumberBody(unsigned char c)
Return true if this is the body character of a C preprocessing number, which is [a-zA-Z0-9_.
LLVM_READONLY bool isHexDigit(unsigned char c)
Return true if this character is an ASCII hex digit: [0-9a-fA-F].
@ Incomplete
Template argument deduction did not deduce a value for every template parameter.