1 //===--- clang/Basic/CharInfo.h - Classifying ASCII Characters --*- 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 #ifndef LLVM_CLANG_BASIC_CHARINFO_H
10 #define LLVM_CLANG_BASIC_CHARINFO_H
11 
12 #include "clang/Basic/LLVM.h"
13 #include "llvm/ADT/StringRef.h"
14 #include "llvm/Support/Compiler.h"
15 #include "llvm/Support/DataTypes.h"
16 
17 namespace clang {
18 namespace charinfo {
19   extern const uint16_t InfoTable[256];
20 
21   enum {
22     CHAR_HORZ_WS  = 0x0001,  // '\t', '\f', '\v'.  Note, no '\0'
23     CHAR_VERT_WS  = 0x0002,  // '\r', '\n'
24     CHAR_SPACE    = 0x0004,  // ' '
25     CHAR_DIGIT    = 0x0008,  // 0-9
26     CHAR_XLETTER  = 0x0010,  // a-f,A-F
27     CHAR_UPPER    = 0x0020,  // A-Z
28     CHAR_LOWER    = 0x0040,  // a-z
29     CHAR_UNDER    = 0x0080,  // _
30     CHAR_PERIOD   = 0x0100,  // .
31     CHAR_RAWDEL   = 0x0200,  // {}[]#<>%:;?*+-/^&|~!=,"'
32     CHAR_PUNCT    = 0x0400   // `$@()
33   };
34 
35   enum {
36     CHAR_XUPPER = CHAR_XLETTER | CHAR_UPPER,
37     CHAR_XLOWER = CHAR_XLETTER | CHAR_LOWER
38   };
39 } // end namespace charinfo
40 
41 /// Returns true if this is an ASCII character.
isASCII(char c)42 LLVM_READNONE inline bool isASCII(char c) {
43   return static_cast<unsigned char>(c) <= 127;
44 }
45 
46 /// Returns true if this is a valid first character of a C identifier,
47 /// which is [a-zA-Z_].
48 LLVM_READONLY inline bool isIdentifierHead(unsigned char c,
49                                            bool AllowDollar = false) {
50   using namespace charinfo;
51   if (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_UNDER))
52     return true;
53   return AllowDollar && c == '$';
54 }
55 
56 /// Returns true if this is a body character of a C identifier,
57 /// which is [a-zA-Z0-9_].
58 LLVM_READONLY inline bool isIdentifierBody(unsigned char c,
59                                            bool AllowDollar = false) {
60   using namespace charinfo;
61   if (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_DIGIT|CHAR_UNDER))
62     return true;
63   return AllowDollar && c == '$';
64 }
65 
66 /// Returns true if this character is horizontal ASCII whitespace:
67 /// ' ', '\\t', '\\f', '\\v'.
68 ///
69 /// Note that this returns false for '\\0'.
isHorizontalWhitespace(unsigned char c)70 LLVM_READONLY inline bool isHorizontalWhitespace(unsigned char c) {
71   using namespace charinfo;
72   return (InfoTable[c] & (CHAR_HORZ_WS|CHAR_SPACE)) != 0;
73 }
74 
75 /// Returns true if this character is vertical ASCII whitespace: '\\n', '\\r'.
76 ///
77 /// Note that this returns false for '\\0'.
isVerticalWhitespace(unsigned char c)78 LLVM_READONLY inline bool isVerticalWhitespace(unsigned char c) {
79   using namespace charinfo;
80   return (InfoTable[c] & CHAR_VERT_WS) != 0;
81 }
82 
83 /// Return true if this character is horizontal or vertical ASCII whitespace:
84 /// ' ', '\\t', '\\f', '\\v', '\\n', '\\r'.
85 ///
86 /// Note that this returns false for '\\0'.
isWhitespace(unsigned char c)87 LLVM_READONLY inline bool isWhitespace(unsigned char c) {
88   using namespace charinfo;
89   return (InfoTable[c] & (CHAR_HORZ_WS|CHAR_VERT_WS|CHAR_SPACE)) != 0;
90 }
91 
92 /// Return true if this character is an ASCII digit: [0-9]
isDigit(unsigned char c)93 LLVM_READONLY inline bool isDigit(unsigned char c) {
94   using namespace charinfo;
95   return (InfoTable[c] & CHAR_DIGIT) != 0;
96 }
97 
98 /// Return true if this character is a lowercase ASCII letter: [a-z]
isLowercase(unsigned char c)99 LLVM_READONLY inline bool isLowercase(unsigned char c) {
100   using namespace charinfo;
101   return (InfoTable[c] & CHAR_LOWER) != 0;
102 }
103 
104 /// Return true if this character is an uppercase ASCII letter: [A-Z]
isUppercase(unsigned char c)105 LLVM_READONLY inline bool isUppercase(unsigned char c) {
106   using namespace charinfo;
107   return (InfoTable[c] & CHAR_UPPER) != 0;
108 }
109 
110 /// Return true if this character is an ASCII letter: [a-zA-Z]
isLetter(unsigned char c)111 LLVM_READONLY inline bool isLetter(unsigned char c) {
112   using namespace charinfo;
113   return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER)) != 0;
114 }
115 
116 /// Return true if this character is an ASCII letter or digit: [a-zA-Z0-9]
isAlphanumeric(unsigned char c)117 LLVM_READONLY inline bool isAlphanumeric(unsigned char c) {
118   using namespace charinfo;
119   return (InfoTable[c] & (CHAR_DIGIT|CHAR_UPPER|CHAR_LOWER)) != 0;
120 }
121 
122 /// Return true if this character is an ASCII hex digit: [0-9a-fA-F]
isHexDigit(unsigned char c)123 LLVM_READONLY inline bool isHexDigit(unsigned char c) {
124   using namespace charinfo;
125   return (InfoTable[c] & (CHAR_DIGIT|CHAR_XLETTER)) != 0;
126 }
127 
128 /// Return true if this character is an ASCII punctuation character.
129 ///
130 /// Note that '_' is both a punctuation character and an identifier character!
isPunctuation(unsigned char c)131 LLVM_READONLY inline bool isPunctuation(unsigned char c) {
132   using namespace charinfo;
133   return (InfoTable[c] & (CHAR_UNDER|CHAR_PERIOD|CHAR_RAWDEL|CHAR_PUNCT)) != 0;
134 }
135 
136 /// Return true if this character is an ASCII printable character; that is, a
137 /// character that should take exactly one column to print in a fixed-width
138 /// terminal.
isPrintable(unsigned char c)139 LLVM_READONLY inline bool isPrintable(unsigned char c) {
140   using namespace charinfo;
141   return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_PERIOD|CHAR_PUNCT|
142                           CHAR_DIGIT|CHAR_UNDER|CHAR_RAWDEL|CHAR_SPACE)) != 0;
143 }
144 
145 /// Return true if this is the body character of a C preprocessing number,
146 /// which is [a-zA-Z0-9_.].
isPreprocessingNumberBody(unsigned char c)147 LLVM_READONLY inline bool isPreprocessingNumberBody(unsigned char c) {
148   using namespace charinfo;
149   return (InfoTable[c] &
150           (CHAR_UPPER|CHAR_LOWER|CHAR_DIGIT|CHAR_UNDER|CHAR_PERIOD)) != 0;
151 }
152 
153 /// Return true if this is the body character of a C++ raw string delimiter.
isRawStringDelimBody(unsigned char c)154 LLVM_READONLY inline bool isRawStringDelimBody(unsigned char c) {
155   using namespace charinfo;
156   return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_PERIOD|
157                           CHAR_DIGIT|CHAR_UNDER|CHAR_RAWDEL)) != 0;
158 }
159 
160 
161 /// Converts the given ASCII character to its lowercase equivalent.
162 ///
163 /// If the character is not an uppercase character, it is returned as is.
toLowercase(char c)164 LLVM_READONLY inline char toLowercase(char c) {
165   if (isUppercase(c))
166     return c + 'a' - 'A';
167   return c;
168 }
169 
170 /// Converts the given ASCII character to its uppercase equivalent.
171 ///
172 /// If the character is not a lowercase character, it is returned as is.
toUppercase(char c)173 LLVM_READONLY inline char toUppercase(char c) {
174   if (isLowercase(c))
175     return c + 'A' - 'a';
176   return c;
177 }
178 
179 
180 /// Return true if this is a valid ASCII identifier.
181 ///
182 /// Note that this is a very simple check; it does not accept UCNs as valid
183 /// identifier characters.
184 LLVM_READONLY inline bool isValidIdentifier(StringRef S,
185                                             bool AllowDollar = false) {
186   if (S.empty() || !isIdentifierHead(S[0], AllowDollar))
187     return false;
188 
189   for (StringRef::iterator I = S.begin(), E = S.end(); I != E; ++I)
190     if (!isIdentifierBody(*I, AllowDollar))
191       return false;
192 
193   return true;
194 }
195 
196 } // end namespace clang
197 
198 #endif
199