1 //===--- clang/Basic/CharInfo.h - Classifying ASCII Characters --*- C++ -*-===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 #ifndef LLVM_CLANG_BASIC_CHARINFO_H
11 #define LLVM_CLANG_BASIC_CHARINFO_H
13 #include "clang/Basic/LLVM.h"
14 #include "llvm/ADT/StringRef.h"
15 #include "llvm/Support/Compiler.h"
16 #include "llvm/Support/DataTypes.h"
20 extern const uint16_t InfoTable[256];
23 CHAR_HORZ_WS = 0x0001, // '\t', '\f', '\v'. Note, no '\0'
24 CHAR_VERT_WS = 0x0002, // '\r', '\n'
25 CHAR_SPACE = 0x0004, // ' '
26 CHAR_DIGIT = 0x0008, // 0-9
27 CHAR_XLETTER = 0x0010, // a-f,A-F
28 CHAR_UPPER = 0x0020, // A-Z
29 CHAR_LOWER = 0x0040, // a-z
30 CHAR_UNDER = 0x0080, // _
31 CHAR_PERIOD = 0x0100, // .
32 CHAR_RAWDEL = 0x0200, // {}[]#<>%:;?*+-/^&|~!=,"'
33 CHAR_PUNCT = 0x0400 // `$@()
37 CHAR_XUPPER = CHAR_XLETTER | CHAR_UPPER,
38 CHAR_XLOWER = CHAR_XLETTER | CHAR_LOWER
40 } // end namespace charinfo
42 /// Returns true if this is an ASCII character.
43 LLVM_READNONE inline bool isASCII(char c) {
44 return static_cast<unsigned char>(c) <= 127;
47 /// Returns true if this is a valid first character of a C identifier,
48 /// which is [a-zA-Z_].
49 LLVM_READONLY inline bool isIdentifierHead(unsigned char c,
50 bool AllowDollar = false) {
51 using namespace charinfo;
52 if (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_UNDER))
54 return AllowDollar && c == '$';
57 /// Returns true if this is a body character of a C identifier,
58 /// which is [a-zA-Z0-9_].
59 LLVM_READONLY inline bool isIdentifierBody(unsigned char c,
60 bool AllowDollar = false) {
61 using namespace charinfo;
62 if (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_DIGIT|CHAR_UNDER))
64 return AllowDollar && c == '$';
67 /// Returns true if this character is horizontal ASCII whitespace:
68 /// ' ', '\\t', '\\f', '\\v'.
70 /// Note that this returns false for '\\0'.
71 LLVM_READONLY inline bool isHorizontalWhitespace(unsigned char c) {
72 using namespace charinfo;
73 return (InfoTable[c] & (CHAR_HORZ_WS|CHAR_SPACE)) != 0;
76 /// Returns true if this character is vertical ASCII whitespace: '\\n', '\\r'.
78 /// Note that this returns false for '\\0'.
79 LLVM_READONLY inline bool isVerticalWhitespace(unsigned char c) {
80 using namespace charinfo;
81 return (InfoTable[c] & CHAR_VERT_WS) != 0;
84 /// Return true if this character is horizontal or vertical ASCII whitespace:
85 /// ' ', '\\t', '\\f', '\\v', '\\n', '\\r'.
87 /// Note that this returns false for '\\0'.
88 LLVM_READONLY inline bool isWhitespace(unsigned char c) {
89 using namespace charinfo;
90 return (InfoTable[c] & (CHAR_HORZ_WS|CHAR_VERT_WS|CHAR_SPACE)) != 0;
93 /// Return true if this character is an ASCII digit: [0-9]
94 LLVM_READONLY inline bool isDigit(unsigned char c) {
95 using namespace charinfo;
96 return (InfoTable[c] & CHAR_DIGIT) != 0;
99 /// Return true if this character is a lowercase ASCII letter: [a-z]
100 LLVM_READONLY inline bool isLowercase(unsigned char c) {
101 using namespace charinfo;
102 return (InfoTable[c] & CHAR_LOWER) != 0;
105 /// Return true if this character is an uppercase ASCII letter: [A-Z]
106 LLVM_READONLY inline bool isUppercase(unsigned char c) {
107 using namespace charinfo;
108 return (InfoTable[c] & CHAR_UPPER) != 0;
111 /// Return true if this character is an ASCII letter: [a-zA-Z]
112 LLVM_READONLY inline bool isLetter(unsigned char c) {
113 using namespace charinfo;
114 return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER)) != 0;
117 /// Return true if this character is an ASCII letter or digit: [a-zA-Z0-9]
118 LLVM_READONLY inline bool isAlphanumeric(unsigned char c) {
119 using namespace charinfo;
120 return (InfoTable[c] & (CHAR_DIGIT|CHAR_UPPER|CHAR_LOWER)) != 0;
123 /// Return true if this character is an ASCII hex digit: [0-9a-fA-F]
124 LLVM_READONLY inline bool isHexDigit(unsigned char c) {
125 using namespace charinfo;
126 return (InfoTable[c] & (CHAR_DIGIT|CHAR_XLETTER)) != 0;
129 /// Return true if this character is an ASCII punctuation character.
131 /// Note that '_' is both a punctuation character and an identifier character!
132 LLVM_READONLY inline bool isPunctuation(unsigned char c) {
133 using namespace charinfo;
134 return (InfoTable[c] & (CHAR_UNDER|CHAR_PERIOD|CHAR_RAWDEL|CHAR_PUNCT)) != 0;
137 /// Return true if this character is an ASCII printable character; that is, a
138 /// character that should take exactly one column to print in a fixed-width
140 LLVM_READONLY inline bool isPrintable(unsigned char c) {
141 using namespace charinfo;
142 return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_PERIOD|CHAR_PUNCT|
143 CHAR_DIGIT|CHAR_UNDER|CHAR_RAWDEL|CHAR_SPACE)) != 0;
146 /// Return true if this is the body character of a C preprocessing number,
147 /// which is [a-zA-Z0-9_.].
148 LLVM_READONLY inline bool isPreprocessingNumberBody(unsigned char c) {
149 using namespace charinfo;
150 return (InfoTable[c] &
151 (CHAR_UPPER|CHAR_LOWER|CHAR_DIGIT|CHAR_UNDER|CHAR_PERIOD)) != 0;
154 /// Return true if this is the body character of a C++ raw string delimiter.
155 LLVM_READONLY inline bool isRawStringDelimBody(unsigned char c) {
156 using namespace charinfo;
157 return (InfoTable[c] & (CHAR_UPPER|CHAR_LOWER|CHAR_PERIOD|
158 CHAR_DIGIT|CHAR_UNDER|CHAR_RAWDEL)) != 0;
162 /// Converts the given ASCII character to its lowercase equivalent.
164 /// If the character is not an uppercase character, it is returned as is.
165 LLVM_READONLY inline char toLowercase(char c) {
167 return c + 'a' - 'A';
171 /// Converts the given ASCII character to its uppercase equivalent.
173 /// If the character is not a lowercase character, it is returned as is.
174 LLVM_READONLY inline char toUppercase(char c) {
176 return c + 'A' - 'a';
181 /// Return true if this is a valid ASCII identifier.
183 /// Note that this is a very simple check; it does not accept UCNs as valid
184 /// identifier characters.
185 LLVM_READONLY inline bool isValidIdentifier(StringRef S,
186 bool AllowDollar = false) {
187 if (S.empty() || !isIdentifierHead(S[0], AllowDollar))
190 for (StringRef::iterator I = S.begin(), E = S.end(); I != E; ++I)
191 if (!isIdentifierBody(*I, AllowDollar))
197 } // end namespace clang