24 #include "llvm/ADT/APInt.h"
25 #include "llvm/ADT/SmallVector.h"
26 #include "llvm/ADT/StringExtras.h"
27 #include "llvm/ADT/StringSwitch.h"
28 #include "llvm/Support/ConvertUTF.h"
29 #include "llvm/Support/ErrorHandling.h"
37 using namespace clang;
41 default: llvm_unreachable(
"Unknown token type!");
42 case tok::char_constant:
43 case tok::string_literal:
44 case tok::utf8_char_constant:
45 case tok::utf8_string_literal:
47 case tok::wide_char_constant:
48 case tok::wide_string_literal:
50 case tok::utf16_char_constant:
51 case tok::utf16_string_literal:
53 case tok::utf32_char_constant:
54 case tok::utf32_string_literal:
62 const char *TokRangeBegin,
63 const char *TokRangeEnd) {
80 const char *TokBegin,
const char *TokRangeBegin,
81 const char *TokRangeEnd,
unsigned DiagID) {
85 return Diags->
Report(Begin, DiagID) <<
92 const char *&ThisTokBuf,
93 const char *ThisTokEnd,
bool &HadError,
97 const char *EscapeBegin = ThisTokBuf;
104 unsigned ResultChar = *ThisTokBuf++;
105 switch (ResultChar) {
107 case '\\':
case '\'':
case '"':
case '?':
break;
119 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
120 diag::ext_nonstandard_escape) <<
"e";
125 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
126 diag::ext_nonstandard_escape) <<
"E";
146 if (ThisTokBuf == ThisTokEnd || !
isHexDigit(*ThisTokBuf)) {
148 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
149 diag::err_hex_escape_no_digits) <<
"x";
155 bool Overflow =
false;
156 for (; ThisTokBuf != ThisTokEnd; ++ThisTokBuf) {
157 int CharVal = llvm::hexDigitValue(ThisTokBuf[0]);
158 if (CharVal == -1)
break;
160 if (ResultChar & 0xF0000000)
163 ResultChar |= CharVal;
167 if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) {
169 ResultChar &= ~0U >> (32-CharWidth);
173 if (Overflow && Diags)
174 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
175 diag::err_escape_too_large) << 0;
178 case '0':
case '1':
case '2':
case '3':
179 case '4':
case '5':
case '6':
case '7': {
186 unsigned NumDigits = 0;
189 ResultChar |= *ThisTokBuf++ -
'0';
191 }
while (ThisTokBuf != ThisTokEnd && NumDigits < 3 &&
192 ThisTokBuf[0] >=
'0' && ThisTokBuf[0] <=
'7');
195 if (CharWidth != 32 && (ResultChar >> CharWidth) != 0) {
197 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
198 diag::err_escape_too_large) << 1;
199 ResultChar &= ~0U >> (32-CharWidth);
205 case '(':
case '{':
case '[':
case '%':
208 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
209 diag::ext_nonstandard_escape)
210 << std::string(1, ResultChar);
217 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
218 diag::ext_unknown_escape)
219 << std::string(1, ResultChar);
221 Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf,
222 diag::ext_unknown_escape)
223 <<
"x" + llvm::utohexstr(ResultChar);
233 char *ResultPtr = ResultBuf;
234 bool Res = llvm::ConvertCodePointToUTF8(Codepoint, ResultPtr);
236 assert(Res &&
"Unexpected conversion failure");
237 Str.append(ResultBuf, ResultPtr);
241 for (StringRef::iterator
I = Input.begin(),
E = Input.end();
I !=
E; ++
I) {
248 assert(*
I ==
'u' || *
I ==
'U');
250 unsigned NumHexDigits;
256 assert(
I + NumHexDigits <=
E);
258 uint32_t CodePoint = 0;
259 for (++
I; NumHexDigits != 0; ++
I, --NumHexDigits) {
260 unsigned Value = llvm::hexDigitValue(*
I);
261 assert(Value != -1U);
275 const char *ThisTokEnd,
276 uint32_t &UcnVal,
unsigned short &UcnLen,
279 bool in_char_string_literal =
false) {
280 const char *UcnBegin = ThisTokBuf;
285 if (ThisTokBuf == ThisTokEnd || !
isHexDigit(*ThisTokBuf)) {
287 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
288 diag::err_hex_escape_no_digits) << StringRef(&ThisTokBuf[-1], 1);
291 UcnLen = (ThisTokBuf[-1] ==
'u' ? 4 : 8);
292 unsigned short UcnLenSave = UcnLen;
293 for (; ThisTokBuf != ThisTokEnd && UcnLenSave; ++ThisTokBuf, UcnLenSave--) {
294 int CharVal = llvm::hexDigitValue(ThisTokBuf[0]);
295 if (CharVal == -1)
break;
302 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
303 diag::err_ucn_escape_incomplete);
308 if ((0xD800 <= UcnVal && UcnVal <= 0xDFFF) ||
311 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
312 diag::err_ucn_escape_invalid);
319 (UcnVal != 0x24 && UcnVal != 0x40 && UcnVal != 0x60)) {
320 bool IsError = (!Features.CPlusPlus11 || !in_char_string_literal);
322 char BasicSCSChar = UcnVal;
323 if (UcnVal >= 0x20 && UcnVal < 0x7f)
324 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
325 IsError ? diag::err_ucn_escape_basic_scs :
326 diag::warn_cxx98_compat_literal_ucn_escape_basic_scs)
327 << StringRef(&BasicSCSChar, 1);
329 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
330 IsError ? diag::err_ucn_control_character :
331 diag::warn_cxx98_compat_literal_ucn_control_character);
337 if (!Features.CPlusPlus && !Features.C99 && Diags)
338 Diag(Diags, Features, Loc, ThisTokBegin, UcnBegin, ThisTokBuf,
339 diag::warn_ucn_not_valid_in_c89_literal);
347 const char *ThisTokEnd,
unsigned CharByteWidth,
350 if (CharByteWidth == 4)
354 unsigned short UcnLen = 0;
358 UcnLen, Loc,
nullptr, Features,
true)) {
364 if (CharByteWidth == 2)
365 return UcnVal <= 0xFFFF ? 2 : 4;
372 if (UcnVal < 0x10000)
382 const char *ThisTokEnd,
383 char *&ResultBuf,
bool &HadError,
387 typedef uint32_t UTF32;
389 unsigned short UcnLen = 0;
391 Loc, Diags, Features,
true)) {
396 assert((CharByteWidth == 1 || CharByteWidth == 2 || CharByteWidth == 4) &&
397 "only character widths of 1, 2, or 4 bytes supported");
400 assert((UcnLen== 4 || UcnLen== 8) &&
"only ucn length of 4 or 8 supported");
402 if (CharByteWidth == 4) {
405 llvm::UTF32 *ResultPtr =
reinterpret_cast<llvm::UTF32*
>(ResultBuf);
411 if (CharByteWidth == 2) {
414 llvm::UTF16 *ResultPtr =
reinterpret_cast<llvm::UTF16*
>(ResultBuf);
416 if (UcnVal <= (UTF32)0xFFFF) {
424 *ResultPtr = 0xD800 + (UcnVal >> 10);
425 *(ResultPtr+1) = 0xDC00 + (UcnVal & 0x3FF);
430 assert(CharByteWidth == 1 &&
"UTF-8 encoding is only for 1 byte characters");
436 typedef uint8_t UTF8;
438 unsigned short bytesToWrite = 0;
439 if (UcnVal < (UTF32)0x80)
441 else if (UcnVal < (UTF32)0x800)
443 else if (UcnVal < (UTF32)0x10000)
448 const unsigned byteMask = 0xBF;
449 const unsigned byteMark = 0x80;
453 static const UTF8 firstByteMark[5] = {
454 0x00, 0x00, 0xC0, 0xE0, 0xF0
457 ResultBuf += bytesToWrite;
458 switch (bytesToWrite) {
460 *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
463 *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
466 *--ResultBuf = (UTF8)((UcnVal | byteMark) & byteMask); UcnVal >>= 6;
469 *--ResultBuf = (UTF8) (UcnVal | firstByteMark[bytesToWrite]);
472 ResultBuf += bytesToWrite;
529 : PP(PP), ThisTokBegin(TokSpelling.begin()), ThisTokEnd(TokSpelling.end()) {
537 s = DigitsBegin = ThisTokBegin;
538 saw_exponent =
false;
540 saw_ud_suffix =
false;
552 ParseNumberStartingWithZero(TokLoc);
558 if (s == ThisTokEnd) {
561 ParseDecimalOrOctalCommon(TokLoc);
568 checkSeparator(TokLoc, s, CSK_AfterDigits);
576 for (; s != ThisTokEnd; ++s) {
582 if (!isFPConstant)
break;
588 if (!isFPConstant)
break;
595 if (!isFPConstant)
break;
602 if (isFPConstant)
break;
613 assert(s + 1 < ThisTokEnd &&
"didn't maximally munch?");
614 if (isFPConstant)
break;
657 assert(s <= ThisTokEnd &&
"didn't maximally munch?");
675 if (s != ThisTokEnd) {
677 expandUCNs(UDSuffixBuf, StringRef(SuffixBegin, ThisTokEnd - SuffixBegin));
689 saw_ud_suffix =
true;
695 diag::err_invalid_suffix_constant)
696 << StringRef(SuffixBegin, ThisTokEnd-SuffixBegin) << isFPConstant;
703 diag::ext_imaginary_constant);
710 void NumericLiteralParser::ParseDecimalOrOctalCommon(
SourceLocation TokLoc){
711 assert((radix == 8 || radix == 10) &&
"Unexpected radix");
715 if (
isHexDigit(*s) && *s !=
'e' && *s !=
'E') {
717 diag::err_invalid_digit) << StringRef(s, 1) << (radix == 8 ? 1 : 0);
723 checkSeparator(TokLoc, s, CSK_AfterDigits);
727 checkSeparator(TokLoc, s, CSK_BeforeDigits);
730 if (*s ==
'e' || *s ==
'E') {
731 checkSeparator(TokLoc, s, CSK_AfterDigits);
732 const char *Exponent = s;
736 if (*s ==
'+' || *s ==
'-') s++;
737 const char *first_non_digit = SkipDigits(s);
738 if (containsDigits(s, first_non_digit)) {
739 checkSeparator(TokLoc, s, CSK_BeforeDigits);
743 diag::err_exponent_has_no_digits);
755 if (!LangOpts.CPlusPlus11 || Suffix.empty())
759 if (Suffix[0] ==
'_')
763 if (!LangOpts.CPlusPlus14)
768 return llvm::StringSwitch<bool>(Suffix)
769 .Cases(
"h",
"min",
"s",
true)
770 .Cases(
"ms",
"us",
"ns",
true)
771 .Cases(
"il",
"i",
"if",
true)
777 CheckSeparatorKind IsAfterDigits) {
778 if (IsAfterDigits == CSK_AfterDigits) {
779 if (Pos == ThisTokBegin)
782 }
else if (Pos == ThisTokEnd)
785 if (isDigitSeparator(*Pos))
787 diag::err_digit_separator_not_between_digits)
796 void NumericLiteralParser::ParseNumberStartingWithZero(
SourceLocation TokLoc) {
797 assert(s[0] ==
'0' &&
"Invalid method call");
803 if ((c1 ==
'x' || c1 ==
'X') && (
isHexDigit(s[1]) || s[1] ==
'.')) {
805 assert(s < ThisTokEnd &&
"didn't maximally munch?");
808 s = SkipHexDigits(s);
809 bool HasSignificandDigits = containsDigits(DigitsBegin, s);
810 if (s == ThisTokEnd) {
812 }
else if (*s ==
'.') {
815 const char *floatDigitsBegin = s;
816 s = SkipHexDigits(s);
817 if (containsDigits(floatDigitsBegin, s))
818 HasSignificandDigits =
true;
819 if (HasSignificandDigits)
820 checkSeparator(TokLoc, floatDigitsBegin, CSK_BeforeDigits);
823 if (!HasSignificandDigits) {
825 diag::err_hex_constant_requires)
833 if (*s ==
'p' || *s ==
'P') {
834 checkSeparator(TokLoc, s, CSK_AfterDigits);
835 const char *Exponent = s;
838 if (*s ==
'+' || *s ==
'-') s++;
839 const char *first_non_digit = SkipDigits(s);
840 if (!containsDigits(s, first_non_digit)) {
842 diag::err_exponent_has_no_digits);
846 checkSeparator(TokLoc, s, CSK_BeforeDigits);
851 ? diag::ext_hex_literal_invalid
852 : diag::ext_hex_constant_invalid);
854 PP.
Diag(TokLoc, diag::warn_cxx1z_hex_literal);
855 }
else if (saw_period) {
857 diag::err_hex_constant_requires)
865 if ((c1 ==
'b' || c1 ==
'B') && (s[1] ==
'0' || s[1] ==
'1')) {
869 ? diag::warn_cxx11_compat_binary_literal
871 ? diag::ext_binary_literal_cxx14
872 : diag::ext_binary_literal);
874 assert(s < ThisTokEnd &&
"didn't maximally munch?");
877 s = SkipBinaryDigits(s);
878 if (s == ThisTokEnd) {
882 diag::err_invalid_digit) << StringRef(s, 1) << 2;
894 s = SkipOctalDigits(s);
901 const char *EndDecimal = SkipDigits(s);
902 if (EndDecimal[0] ==
'.' || EndDecimal[0] ==
'e' || EndDecimal[0] ==
'E') {
908 ParseDecimalOrOctalCommon(TokLoc);
914 return NumDigits <= 64;
916 return NumDigits <= 64 / 3;
918 return NumDigits <= 19;
920 return NumDigits <= 64 / 4;
922 llvm_unreachable(
"impossible Radix");
936 const unsigned NumDigits = SuffixBegin - DigitsBegin;
939 for (
const char *Ptr = DigitsBegin; Ptr != SuffixBegin; ++Ptr)
940 if (!isDigitSeparator(*Ptr))
941 N = N * radix + llvm::hexDigitValue(*Ptr);
946 return Val.getZExtValue() != N;
950 const char *Ptr = DigitsBegin;
952 llvm::APInt RadixVal(Val.getBitWidth(), radix);
953 llvm::APInt CharVal(Val.getBitWidth(), 0);
954 llvm::APInt OldVal = Val;
956 bool OverflowOccurred =
false;
957 while (Ptr < SuffixBegin) {
958 if (isDigitSeparator(*Ptr)) {
963 unsigned C = llvm::hexDigitValue(*Ptr++);
966 assert(C < radix &&
"NumericLiteralParser ctor should have rejected this");
976 OverflowOccurred |= Val.udiv(RadixVal) != OldVal;
981 OverflowOccurred |= Val.ult(CharVal);
983 return OverflowOccurred;
986 llvm::APFloat::opStatus
990 unsigned n =
std::min(SuffixBegin - ThisTokBegin, ThisTokEnd - ThisTokBegin);
993 StringRef Str(ThisTokBegin, n);
994 if (Str.find(
'\'') != StringRef::npos) {
996 std::remove_copy_if(Str.begin(), Str.end(), std::back_inserter(Buffer),
1001 return Result.convertFromString(Str, APFloat::rmNearestTiesToEven);
1051 const char *TokBegin = begin;
1054 if (
Kind != tok::char_constant)
1056 if (
Kind == tok::utf8_char_constant)
1060 assert(begin[0] ==
'\'' &&
"Invalid token lexed");
1064 if (end[-1] !=
'\'') {
1065 const char *UDSuffixEnd = end;
1068 }
while (end[-1] !=
'\'');
1070 expandUCNs(UDSuffixBuf, StringRef(end, UDSuffixEnd - end));
1071 UDSuffixOffset = end - TokBegin;
1075 assert(end != begin &&
"Invalid token lexed");
1082 "Assumes char is 8 bits");
1085 "Assumes sizeof(int) on target is <= 64 and a multiple of char");
1087 "Assumes sizeof(wchar) on target is <= 64");
1090 codepoint_buffer.resize(end - begin);
1091 uint32_t *buffer_begin = &codepoint_buffer.front();
1092 uint32_t *buffer_end = buffer_begin + codepoint_buffer.size();
1097 uint32_t largest_character_for_kind;
1098 if (tok::wide_char_constant ==
Kind) {
1099 largest_character_for_kind =
1101 }
else if (tok::utf8_char_constant ==
Kind) {
1102 largest_character_for_kind = 0x7F;
1103 }
else if (tok::utf16_char_constant ==
Kind) {
1104 largest_character_for_kind = 0xFFFF;
1105 }
else if (tok::utf32_char_constant ==
Kind) {
1106 largest_character_for_kind = 0x10FFFF;
1108 largest_character_for_kind = 0x7Fu;
1111 while (begin != end) {
1113 if (begin[0] !=
'\\') {
1114 char const *start = begin;
1117 }
while (begin != end && *begin !=
'\\');
1119 char const *tmp_in_start = start;
1120 uint32_t *tmp_out_start = buffer_begin;
1121 llvm::ConversionResult res =
1122 llvm::ConvertUTF8toUTF32(reinterpret_cast<llvm::UTF8 const **>(&start),
1123 reinterpret_cast<llvm::UTF8 const *>(begin),
1124 &buffer_begin, buffer_end, llvm::strictConversion);
1125 if (res != llvm::conversionOK) {
1129 bool NoErrorOnBadEncoding =
isAscii();
1130 unsigned Msg = diag::err_bad_character_encoding;
1131 if (NoErrorOnBadEncoding)
1132 Msg = diag::warn_bad_character_encoding;
1134 if (NoErrorOnBadEncoding) {
1135 start = tmp_in_start;
1136 buffer_begin = tmp_out_start;
1137 for (; start != begin; ++start, ++buffer_begin)
1138 *buffer_begin = static_cast<uint8_t>(*start);
1143 for (; tmp_out_start < buffer_begin; ++tmp_out_start) {
1144 if (*tmp_out_start > largest_character_for_kind) {
1146 PP.
Diag(Loc, diag::err_character_too_large);
1154 if (begin[1] ==
'u' || begin[1] ==
'U') {
1155 unsigned short UcnLen = 0;
1160 }
else if (*buffer_begin > largest_character_for_kind) {
1162 PP.
Diag(Loc, diag::err_character_too_large);
1173 *buffer_begin++ = result;
1176 unsigned NumCharsSoFar = buffer_begin - &codepoint_buffer.front();
1178 if (NumCharsSoFar > 1) {
1180 PP.
Diag(Loc, diag::warn_extraneous_char_constant);
1181 else if (
isAscii() && NumCharsSoFar == 4)
1182 PP.
Diag(Loc, diag::ext_four_char_character_literal);
1184 PP.
Diag(Loc, diag::ext_multichar_character_literal);
1186 PP.
Diag(Loc, diag::err_multichar_utf_character_literal);
1189 IsMultiChar =
false;
1196 bool multi_char_too_long =
false;
1199 for (
size_t i = 0; i < NumCharsSoFar; ++i) {
1201 multi_char_too_long |= (LitVal.countLeadingZeros() < 8);
1203 LitVal = LitVal + (codepoint_buffer[i] & 0xFF);
1205 }
else if (NumCharsSoFar > 0) {
1207 LitVal = buffer_begin[-1];
1210 if (!HadError && multi_char_too_long) {
1211 PP.
Diag(Loc, diag::warn_char_constant_too_large);
1215 Value = LitVal.getZExtValue();
1221 if (
isAscii() && NumCharsSoFar == 1 && (
Value & 128) &&
1283 :
SM(PP.getSourceManager()), Features(PP.getLangOpts()),
1284 Target(PP.getTargetInfo()), Diags(Complain ? &PP.getDiagnostics() :nullptr),
1285 MaxTokenLength(0), SizeBound(0), CharByteWidth(0),
Kind(tok::unknown),
1286 ResultPtr(ResultBuf.data()), hadError(
false), Pascal(
false) {
1293 if (StringToks.empty() || StringToks[0].getLength() < 2)
1300 assert(!StringToks.empty() &&
"expected at least one token");
1301 MaxTokenLength = StringToks[0].getLength();
1302 assert(StringToks[0].getLength() >= 2 &&
"literal token is invalid!");
1303 SizeBound = StringToks[0].getLength()-2;
1304 Kind = StringToks[0].getKind();
1310 for (
unsigned i = 1; i != StringToks.size(); ++i) {
1311 if (StringToks[i].getLength() < 2)
1312 return DiagnoseLexingError(StringToks[i].getLocation());
1316 assert(StringToks[i].getLength() >= 2 &&
"literal token is invalid!");
1317 SizeBound += StringToks[i].getLength()-2;
1320 if (StringToks[i].getLength() > MaxTokenLength)
1321 MaxTokenLength = StringToks[i].getLength();
1325 if (StringToks[i].isNot(
Kind) && StringToks[i].isNot(tok::string_literal)) {
1327 Kind = StringToks[i].getKind();
1330 Diags->
Report(StringToks[i].getLocation(),
1331 diag::err_unsupported_string_concat);
1344 assert((CharByteWidth & 7) == 0 &&
"Assumes character size is byte multiple");
1349 SizeBound *= CharByteWidth;
1352 ResultBuf.resize(SizeBound);
1356 TokenBuf.resize(MaxTokenLength);
1360 ResultPtr = &ResultBuf[0];
1366 for (
unsigned i = 0, e = StringToks.size(); i != e; ++i) {
1367 const char *ThisTokBuf = &TokenBuf[0];
1371 bool StringInvalid =
false;
1372 unsigned ThisTokLen =
1376 return DiagnoseLexingError(StringToks[i].getLocation());
1378 const char *ThisTokBegin = ThisTokBuf;
1379 const char *ThisTokEnd = ThisTokBuf+ThisTokLen;
1382 if (ThisTokEnd[-1] !=
'"') {
1383 const char *UDSuffixEnd = ThisTokEnd;
1386 }
while (ThisTokEnd[-1] !=
'"');
1388 StringRef UDSuffix(ThisTokEnd, UDSuffixEnd - ThisTokEnd);
1390 if (UDSuffixBuf.empty()) {
1391 if (StringToks[i].hasUCN())
1394 UDSuffixBuf.assign(UDSuffix);
1396 UDSuffixOffset = ThisTokEnd - ThisTokBuf;
1397 UDSuffixTokLoc = StringToks[i].getLocation();
1400 if (StringToks[i].hasUCN()) {
1402 UDSuffix = ExpandedUDSuffix;
1409 if (UDSuffixBuf != UDSuffix) {
1412 Diags->
Report(TokLoc, diag::err_string_concat_mixed_suffix)
1413 << UDSuffixBuf << UDSuffix
1428 if (ThisTokBuf[0] ==
'L' || ThisTokBuf[0] ==
'u' || ThisTokBuf[0] ==
'U') {
1431 if (ThisTokBuf[0] ==
'8')
1436 if (ThisTokBuf[0] ==
'R') {
1439 const char *Prefix = ThisTokBuf;
1440 while (ThisTokBuf[0] !=
'(')
1445 ThisTokEnd -= ThisTokBuf - Prefix;
1446 assert(ThisTokEnd >= ThisTokBuf &&
"malformed raw string literal");
1450 StringRef RemainingTokenSpan(ThisTokBuf, ThisTokEnd - ThisTokBuf);
1451 while (!RemainingTokenSpan.empty()) {
1453 size_t CRLFPos = RemainingTokenSpan.find(
"\r\n");
1454 StringRef BeforeCRLF = RemainingTokenSpan.substr(0, CRLFPos);
1455 StringRef AfterCRLF = RemainingTokenSpan.substr(CRLFPos);
1458 if (CopyStringFragment(StringToks[i], ThisTokBegin, BeforeCRLF))
1463 RemainingTokenSpan = AfterCRLF.substr(1);
1466 if (ThisTokBuf[0] !=
'"') {
1469 return DiagnoseLexingError(StringToks[i].getLocation());
1474 if (Features.PascalStrings && ThisTokBuf + 1 != ThisTokEnd &&
1475 ThisTokBuf[0] ==
'\\' && ThisTokBuf[1] ==
'p') {
1486 while (ThisTokBuf != ThisTokEnd) {
1488 if (ThisTokBuf[0] !=
'\\') {
1489 const char *InStart = ThisTokBuf;
1492 }
while (ThisTokBuf != ThisTokEnd && ThisTokBuf[0] !=
'\\');
1495 if (CopyStringFragment(StringToks[i], ThisTokBegin,
1496 StringRef(InStart, ThisTokBuf - InStart)))
1501 if (ThisTokBuf[1] ==
'u' || ThisTokBuf[1] ==
'U') {
1505 CharByteWidth, Diags, Features);
1509 unsigned ResultChar =
1512 CharByteWidth*8, Diags, Features);
1514 if (CharByteWidth == 4) {
1517 llvm::UTF32 *ResultWidePtr =
reinterpret_cast<llvm::UTF32*
>(ResultPtr);
1518 *ResultWidePtr = ResultChar;
1520 }
else if (CharByteWidth == 2) {
1523 llvm::UTF16 *ResultWidePtr =
reinterpret_cast<llvm::UTF16*
>(ResultPtr);
1524 *ResultWidePtr = ResultChar & 0xFFFF;
1527 assert(CharByteWidth == 1 &&
"Unexpected char width");
1528 *ResultPtr++ = ResultChar & 0xFF;
1535 if (CharByteWidth == 4) {
1538 llvm::UTF32 *ResultWidePtr =
reinterpret_cast<llvm::UTF32*
>(ResultBuf.data());
1540 }
else if (CharByteWidth == 2) {
1543 llvm::UTF16 *ResultWidePtr =
reinterpret_cast<llvm::UTF16*
>(ResultBuf.data());
1546 assert(CharByteWidth == 1 &&
"Unexpected char width");
1553 Diags->
Report(StringToks.front().getLocation(),
1554 diag::err_pascal_string_too_long)
1556 StringToks.back().getLocation());
1562 unsigned MaxChars = Features.CPlusPlus? 65536 : Features.C99 ? 4095 : 509;
1565 Diags->
Report(StringToks.front().getLocation(),
1566 diag::ext_string_too_long)
1568 << (Features.CPlusPlus ? 2 : Features.C99 ? 1 : 0)
1570 StringToks.back().getLocation());
1577 End = Err + std::min<unsigned>(llvm::getNumBytesForUTF8(*Err), End-Err);
1578 while (++Err != End && (*Err & 0xC0) == 0x80)
1586 bool StringLiteralParser::CopyStringFragment(
const Token &Tok,
1587 const char *TokBegin,
1588 StringRef Fragment) {
1589 const llvm::UTF8 *ErrorPtrTmp;
1590 if (ConvertUTF8toWide(CharByteWidth, Fragment, ResultPtr, ErrorPtrTmp))
1596 bool NoErrorOnBadEncoding =
isAscii();
1597 if (NoErrorOnBadEncoding) {
1598 memcpy(ResultPtr, Fragment.data(), Fragment.size());
1599 ResultPtr += Fragment.size();
1603 const char *ErrorPtr =
reinterpret_cast<const char *
>(ErrorPtrTmp);
1607 Diag(Diags, Features, SourceLoc, TokBegin,
1608 ErrorPtr,
resyncUTF8(ErrorPtr, Fragment.end()),
1609 NoErrorOnBadEncoding ? diag::warn_bad_string_encoding
1610 : diag::err_bad_string_encoding);
1612 const char *NextStart =
resyncUTF8(ErrorPtr, Fragment.end());
1613 StringRef NextFragment(NextStart, Fragment.end()-NextStart);
1617 Dummy.reserve(Fragment.size() * CharByteWidth);
1618 char *Ptr = Dummy.data();
1620 while (!ConvertUTF8toWide(CharByteWidth, NextFragment, Ptr, ErrorPtrTmp)) {
1621 const char *ErrorPtr =
reinterpret_cast<const char *
>(ErrorPtrTmp);
1622 NextStart =
resyncUTF8(ErrorPtr, Fragment.end());
1624 ErrorPtr, NextStart);
1625 NextFragment = StringRef(NextStart, Fragment.end()-NextStart);
1628 return !NoErrorOnBadEncoding;
1631 void StringLiteralParser::DiagnoseLexingError(
SourceLocation Loc) {
1634 Diags->
Report(Loc, diag::err_lexing_string);
1641 unsigned ByteNo)
const {
1646 bool StringInvalid =
false;
1647 const char *SpellingPtr = &SpellingBuffer[0];
1653 const char *SpellingStart = SpellingPtr;
1654 const char *SpellingEnd = SpellingPtr+TokLen;
1657 if (SpellingPtr[0] ==
'u' && SpellingPtr[1] ==
'8')
1660 assert(SpellingPtr[0] !=
'L' && SpellingPtr[0] !=
'u' &&
1661 SpellingPtr[0] !=
'U' &&
"Doesn't handle wide or utf strings yet");
1664 if (SpellingPtr[0] ==
'R') {
1665 assert(SpellingPtr[1] ==
'"' &&
"Should be a raw string literal!");
1668 while (*SpellingPtr !=
'(') {
1670 assert(SpellingPtr < SpellingEnd &&
"Missing ( for raw string literal");
1674 return SpellingPtr - SpellingStart + ByteNo;
1678 assert(SpellingPtr[0] ==
'"' &&
"Should be a string literal!");
1683 assert(SpellingPtr < SpellingEnd &&
"Didn't find byte offset!");
1686 if (*SpellingPtr !=
'\\') {
1693 bool HadError =
false;
1694 if (SpellingPtr[1] ==
'u' || SpellingPtr[1] ==
'U') {
1695 const char *EscapePtr = SpellingPtr;
1697 1, Features, HadError);
1700 SpellingPtr = EscapePtr;
1707 CharByteWidth*8, Diags, Features);
1710 assert(!HadError &&
"This method isn't valid on erroneous strings");
1713 return SpellingPtr-SpellingStart;
SourceManager & getSourceManager() const
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...
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 LLVM_READONLY bool isDigit(unsigned char c)
Return true if this character is an ASCII digit: [0-9].
unsigned getChar16Width() const
getChar16Width/Align - Return the size of 'char16_t' for this target, in bits.
StringLiteralParser(ArrayRef< Token > StringToks, Preprocessor &PP, bool Complain=true)
const SourceManager & getManager() const
std::unique_ptr< llvm::MemoryBuffer > Buffer
DiagnosticBuilder Report(SourceLocation Loc, unsigned DiagID)
Issue the message to the client.
unsigned getChar32Width() const
getChar32Width/Align - Return the size of 'char32_t' for this target, in bits.
static 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_.
static bool alwaysFitsInto64Bits(unsigned Radix, unsigned NumDigits)
unsigned getOffsetOfStringByte(const Token &TheTok, unsigned ByteNo) const
getOffsetOfStringByte - This function returns the offset of the specified byte of the string data rep...
const LangOptions & getLangOpts() const
Token - This structure provides full information about a lexed token.
Keeps track of the various options that can be enabled, which controls the dialect of C or C++ that i...
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...
Concrete class used by the front-end to report problems and issues.
static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix)
Determine whether a suffix is a valid ud-suffix.
NumericLiteralParser(StringRef TokSpelling, SourceLocation TokLoc, Preprocessor &PP)
integer-constant: [C99 6.4.4.1] decimal-constant integer-suffix octal-constant integer-suffix hexadec...
SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart, unsigned Char) const
Given a location that specifies the start of a token, return a new location that specifies a characte...
const TargetInfo & getTargetInfo() const
static SourceLocation AdvanceToTokenCharacter(SourceLocation TokStart, unsigned Character, const SourceManager &SM, const LangOptions &LangOpts)
AdvanceToTokenCharacter - If the current SourceLocation specifies a location at the start of a token...
detail::InMemoryDirectory::const_iterator I
DiagnosticBuilder Diag(SourceLocation Loc, unsigned DiagID) const
Forwarding function for diagnostics.
unsigned getWCharWidth() const
getWCharWidth/Align - Return the size of 'wchar_t' for this target, in bits.
static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix)
Determine whether a suffix is a valid ud-suffix.
A little helper class used to produce diagnostics.
bool isFloatingLiteral() const
Exposes information about the current target.
CharLiteralParser(const char *begin, const char *end, SourceLocation Loc, Preprocessor &PP, tok::TokenKind kind)
Defines the clang::LangOptions interface.
Represents a character-granular source range.
Defines the clang::Preprocessor interface.
static const char * resyncUTF8(const char *Err, const char *End)
char __ovld __cnfn min(char x, char y)
Returns y if y < x, otherwise it returns x.
SourceLocation getLocation() const
Return a source location identifier for the specified offset in the current file. ...
static void appendCodePoint(unsigned Codepoint, llvm::SmallVectorImpl< char > &Str)
static CharSourceRange getCharRange(SourceRange R)
bool GetIntegerValue(llvm::APInt &Val)
GetIntegerValue - Convert this numeric literal value to an APInt that matches Val's input width...
Encodes a location in the source.
llvm::APFloat::opStatus GetFloatValue(llvm::APFloat &Result)
GetFloatValue - Convert this numeric literal to a floating value, using the specified APFloat fltSema...
static unsigned ProcessCharEscape(const char *ThisTokBegin, const char *&ThisTokBuf, const char *ThisTokEnd, bool &HadError, FullSourceLoc Loc, unsigned CharWidth, DiagnosticsEngine *Diags, const LangOptions &Features)
ProcessCharEscape - Parse a standard C escape sequence, which can occur in either a character or a st...
TokenKind
Provides a simple uniform namespace for tokens from all C languages.
void expandUCNs(SmallVectorImpl< char > &Buf, StringRef Input)
Copy characters from Input to Buf, expanding any UCNs.
DiagnosticsEngine & getDiagnostics() const
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 LLVM_READONLY bool isPrintable(unsigned char c)
Return true if this character is an ASCII printable character; that is, a character that should take ...
unsigned getCharWidth() const
detail::InMemoryDirectory::const_iterator E
unsigned getIntWidth() const
getIntWidth/Align - Return the size of 'signed int' and 'unsigned int' for this target, in bits.
Defines the clang::SourceLocation class and associated facilities.
BoundNodesTreeBuilder *const Builder
static unsigned getCharWidth(tok::TokenKind kind, const TargetInfo &Target)
unsigned kind
All of the diagnostics that can be emitted by the frontend.
Defines the clang::TargetInfo interface.
unsigned GetStringLength() const
A SourceLocation and its associated SourceManager.
unsigned getLength() const
A trivial tuple used to represent a source range.
unsigned GetNumStringChars() const
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)
Engages in a tight little dance with the lexer to efficiently preprocess tokens.
static LLVM_READONLY bool isHexDigit(unsigned char c)
Return true if this character is an ASCII hex digit: [0-9a-fA-F].