1 //===--- LiteralSupport.h ---------------------------------------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file defines the NumericLiteralParser, CharLiteralParser, and
10 // StringLiteralParser interfaces.
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #ifndef LLVM_CLANG_LEX_LITERALSUPPORT_H
15 #define LLVM_CLANG_LEX_LITERALSUPPORT_H
16 
17 #include "clang/Basic/CharInfo.h"
18 #include "clang/Basic/LLVM.h"
19 #include "clang/Basic/TokenKinds.h"
20 #include "llvm/ADT/APFloat.h"
21 #include "llvm/ADT/ArrayRef.h"
22 #include "llvm/ADT/SmallString.h"
23 #include "llvm/ADT/StringRef.h"
24 #include "llvm/Support/DataTypes.h"
25 
26 namespace clang {
27 
28 class DiagnosticsEngine;
29 class Preprocessor;
30 class Token;
31 class SourceLocation;
32 class TargetInfo;
33 class SourceManager;
34 class LangOptions;
35 
36 /// Copy characters from Input to Buf, expanding any UCNs.
37 void expandUCNs(SmallVectorImpl<char> &Buf, StringRef Input);
38 
39 /// NumericLiteralParser - This performs strict semantic analysis of the content
40 /// of a ppnumber, classifying it as either integer, floating, or erroneous,
41 /// determines the radix of the value and can convert it to a useful value.
42 class NumericLiteralParser {
43   const SourceManager &SM;
44   const LangOptions &LangOpts;
45   DiagnosticsEngine &Diags;
46 
47   const char *const ThisTokBegin;
48   const char *const ThisTokEnd;
49   const char *DigitsBegin, *SuffixBegin; // markers
50   const char *s; // cursor
51 
52   unsigned radix;
53 
54   bool saw_exponent, saw_period, saw_ud_suffix, saw_fixed_point_suffix;
55 
56   SmallString<32> UDSuffixBuf;
57 
58 public:
59   NumericLiteralParser(StringRef TokSpelling, SourceLocation TokLoc,
60                        const SourceManager &SM, const LangOptions &LangOpts,
61                        const TargetInfo &Target, DiagnosticsEngine &Diags);
62   bool hadError : 1;
63   bool isUnsigned : 1;
64   bool isLong : 1;          // This is *not* set for long long.
65   bool isLongLong : 1;
66   bool isSizeT : 1;         // 1z, 1uz (C++2b)
67   bool isHalf : 1;          // 1.0h
68   bool isFloat : 1;         // 1.0f
69   bool isImaginary : 1;     // 1.0i
70   bool isFloat16 : 1;       // 1.0f16
71   bool isFloat128 : 1;      // 1.0q
72   uint8_t MicrosoftInteger; // Microsoft suffix extension i8, i16, i32, or i64.
73 
74   bool isFract : 1;         // 1.0hr/r/lr/uhr/ur/ulr
75   bool isAccum : 1;         // 1.0hk/k/lk/uhk/uk/ulk
76 
77   bool isFixedPointLiteral() const {
78     return (saw_period || saw_exponent) && saw_fixed_point_suffix;
79   }
80 
81   bool isIntegerLiteral() const {
82     return !saw_period && !saw_exponent && !isFixedPointLiteral();
83   }
84   bool isFloatingLiteral() const {
85     return (saw_period || saw_exponent) && !isFixedPointLiteral();
86   }
87 
88   bool hasUDSuffix() const {
89     return saw_ud_suffix;
90   }
91   StringRef getUDSuffix() const {
92     assert(saw_ud_suffix);
93     return UDSuffixBuf;
94   }
95   unsigned getUDSuffixOffset() const {
96     assert(saw_ud_suffix);
97     return SuffixBegin - ThisTokBegin;
98   }
99 
100   static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix);
101 
102   unsigned getRadix() const { return radix; }
103 
104   /// GetIntegerValue - Convert this numeric literal value to an APInt that
105   /// matches Val's input width.  If there is an overflow (i.e., if the unsigned
106   /// value read is larger than the APInt's bits will hold), set Val to the low
107   /// bits of the result and return true.  Otherwise, return false.
108   bool GetIntegerValue(llvm::APInt &Val);
109 
110   /// GetFloatValue - Convert this numeric literal to a floating value, using
111   /// the specified APFloat fltSemantics (specifying float, double, etc).
112   /// The optional bool isExact (passed-by-reference) has its value
113   /// set to true if the returned APFloat can represent the number in the
114   /// literal exactly, and false otherwise.
115   llvm::APFloat::opStatus GetFloatValue(llvm::APFloat &Result);
116 
117   /// GetFixedPointValue - Convert this numeric literal value into a
118   /// scaled integer that represents this value. Returns true if an overflow
119   /// occurred when calculating the integral part of the scaled integer or
120   /// calculating the digit sequence of the exponent.
121   bool GetFixedPointValue(llvm::APInt &StoreVal, unsigned Scale);
122 
123 private:
124 
125   void ParseNumberStartingWithZero(SourceLocation TokLoc);
126   void ParseDecimalOrOctalCommon(SourceLocation TokLoc);
127 
128   static bool isDigitSeparator(char C) { return C == '\''; }
129 
130   /// Determine whether the sequence of characters [Start, End) contains
131   /// any real digits (not digit separators).
132   bool containsDigits(const char *Start, const char *End) {
133     return Start != End && (Start + 1 != End || !isDigitSeparator(Start[0]));
134   }
135 
136   enum CheckSeparatorKind { CSK_BeforeDigits, CSK_AfterDigits };
137 
138   /// Ensure that we don't have a digit separator here.
139   void checkSeparator(SourceLocation TokLoc, const char *Pos,
140                       CheckSeparatorKind IsAfterDigits);
141 
142   /// SkipHexDigits - Read and skip over any hex digits, up to End.
143   /// Return a pointer to the first non-hex digit or End.
144   const char *SkipHexDigits(const char *ptr) {
145     while (ptr != ThisTokEnd && (isHexDigit(*ptr) || isDigitSeparator(*ptr)))
146       ptr++;
147     return ptr;
148   }
149 
150   /// SkipOctalDigits - Read and skip over any octal digits, up to End.
151   /// Return a pointer to the first non-hex digit or End.
152   const char *SkipOctalDigits(const char *ptr) {
153     while (ptr != ThisTokEnd &&
154            ((*ptr >= '0' && *ptr <= '7') || isDigitSeparator(*ptr)))
155       ptr++;
156     return ptr;
157   }
158 
159   /// SkipDigits - Read and skip over any digits, up to End.
160   /// Return a pointer to the first non-hex digit or End.
161   const char *SkipDigits(const char *ptr) {
162     while (ptr != ThisTokEnd && (isDigit(*ptr) || isDigitSeparator(*ptr)))
163       ptr++;
164     return ptr;
165   }
166 
167   /// SkipBinaryDigits - Read and skip over any binary digits, up to End.
168   /// Return a pointer to the first non-binary digit or End.
169   const char *SkipBinaryDigits(const char *ptr) {
170     while (ptr != ThisTokEnd &&
171            (*ptr == '0' || *ptr == '1' || isDigitSeparator(*ptr)))
172       ptr++;
173     return ptr;
174   }
175 
176 };
177 
178 /// CharLiteralParser - Perform interpretation and semantic analysis of a
179 /// character literal.
180 class CharLiteralParser {
181   uint64_t Value;
182   tok::TokenKind Kind;
183   bool IsMultiChar;
184   bool HadError;
185   SmallString<32> UDSuffixBuf;
186   unsigned UDSuffixOffset;
187 public:
188   CharLiteralParser(const char *begin, const char *end,
189                     SourceLocation Loc, Preprocessor &PP,
190                     tok::TokenKind kind);
191 
192   bool hadError() const { return HadError; }
193   bool isAscii() const { return Kind == tok::char_constant; }
194   bool isWide() const { return Kind == tok::wide_char_constant; }
195   bool isUTF8() const { return Kind == tok::utf8_char_constant; }
196   bool isUTF16() const { return Kind == tok::utf16_char_constant; }
197   bool isUTF32() const { return Kind == tok::utf32_char_constant; }
198   bool isMultiChar() const { return IsMultiChar; }
199   uint64_t getValue() const { return Value; }
200   StringRef getUDSuffix() const { return UDSuffixBuf; }
201   unsigned getUDSuffixOffset() const {
202     assert(!UDSuffixBuf.empty() && "no ud-suffix");
203     return UDSuffixOffset;
204   }
205 };
206 
207 /// StringLiteralParser - This decodes string escape characters and performs
208 /// wide string analysis and Translation Phase #6 (concatenation of string
209 /// literals) (C99 5.1.1.2p1).
210 class StringLiteralParser {
211   const SourceManager &SM;
212   const LangOptions &Features;
213   const TargetInfo &Target;
214   DiagnosticsEngine *Diags;
215 
216   unsigned MaxTokenLength;
217   unsigned SizeBound;
218   unsigned CharByteWidth;
219   tok::TokenKind Kind;
220   SmallString<512> ResultBuf;
221   char *ResultPtr; // cursor
222   SmallString<32> UDSuffixBuf;
223   unsigned UDSuffixToken;
224   unsigned UDSuffixOffset;
225 public:
226   StringLiteralParser(ArrayRef<Token> StringToks,
227                       Preprocessor &PP, bool Complain = true);
228   StringLiteralParser(ArrayRef<Token> StringToks,
229                       const SourceManager &sm, const LangOptions &features,
230                       const TargetInfo &target,
231                       DiagnosticsEngine *diags = nullptr)
232     : SM(sm), Features(features), Target(target), Diags(diags),
233       MaxTokenLength(0), SizeBound(0), CharByteWidth(0), Kind(tok::unknown),
234       ResultPtr(ResultBuf.data()), hadError(false), Pascal(false) {
235     init(StringToks);
236   }
237 
238 
239   bool hadError;
240   bool Pascal;
241 
242   StringRef GetString() const {
243     return StringRef(ResultBuf.data(), GetStringLength());
244   }
245   unsigned GetStringLength() const { return ResultPtr-ResultBuf.data(); }
246 
247   unsigned GetNumStringChars() const {
248     return GetStringLength() / CharByteWidth;
249   }
250   /// getOffsetOfStringByte - This function returns the offset of the
251   /// specified byte of the string data represented by Token.  This handles
252   /// advancing over escape sequences in the string.
253   ///
254   /// If the Diagnostics pointer is non-null, then this will do semantic
255   /// checking of the string literal and emit errors and warnings.
256   unsigned getOffsetOfStringByte(const Token &TheTok, unsigned ByteNo) const;
257 
258   bool isAscii() const { return Kind == tok::string_literal; }
259   bool isWide() const { return Kind == tok::wide_string_literal; }
260   bool isUTF8() const { return Kind == tok::utf8_string_literal; }
261   bool isUTF16() const { return Kind == tok::utf16_string_literal; }
262   bool isUTF32() const { return Kind == tok::utf32_string_literal; }
263   bool isPascal() const { return Pascal; }
264 
265   StringRef getUDSuffix() const { return UDSuffixBuf; }
266 
267   /// Get the index of a token containing a ud-suffix.
268   unsigned getUDSuffixToken() const {
269     assert(!UDSuffixBuf.empty() && "no ud-suffix");
270     return UDSuffixToken;
271   }
272   /// Get the spelling offset of the first byte of the ud-suffix.
273   unsigned getUDSuffixOffset() const {
274     assert(!UDSuffixBuf.empty() && "no ud-suffix");
275     return UDSuffixOffset;
276   }
277 
278   static bool isValidUDSuffix(const LangOptions &LangOpts, StringRef Suffix);
279 
280 private:
281   void init(ArrayRef<Token> StringToks);
282   bool CopyStringFragment(const Token &Tok, const char *TokBegin,
283                           StringRef Fragment);
284   void DiagnoseLexingError(SourceLocation Loc);
285 };
286 
287 }  // end namespace clang
288 
289 #endif
290