1//===-- llvm/ADT/StringExtras.h - Useful string functions -------*- C++ -*-===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file contains some functions that are useful when dealing with strings.
11//
12//===----------------------------------------------------------------------===//
13
14#ifndef LLVM_ADT_STRINGEXTRAS_H
15#define LLVM_ADT_STRINGEXTRAS_H
16
17#include "llvm/ADT/StringRef.h"
18#include "llvm/Support/DataTypes.h"
19
20namespace llvm {
21template<typename T> class SmallVectorImpl;
22
23/// hexdigit - Return the hexadecimal character for the
24/// given number \p X (which should be less than 16).
25static inline char hexdigit(unsigned X, bool LowerCase = false) {
26  const char HexChar = LowerCase ? 'a' : 'A';
27  return X < 10 ? '0' + X : HexChar + X - 10;
28}
29
30/// Interpret the given character \p C as a hexadecimal digit and return its
31/// value.
32///
33/// If \p C is not a valid hex digit, -1U is returned.
34static inline unsigned hexDigitValue(char C) {
35  if (C >= '0' && C <= '9') return C-'0';
36  if (C >= 'a' && C <= 'f') return C-'a'+10U;
37  if (C >= 'A' && C <= 'F') return C-'A'+10U;
38  return -1U;
39}
40
41/// utohex_buffer - Emit the specified number into the buffer specified by
42/// BufferEnd, returning a pointer to the start of the string.  This can be used
43/// like this: (note that the buffer must be large enough to handle any number):
44///    char Buffer[40];
45///    printf("0x%s", utohex_buffer(X, Buffer+40));
46///
47/// This should only be used with unsigned types.
48///
49template<typename IntTy>
50static inline char *utohex_buffer(IntTy X, char *BufferEnd) {
51  char *BufPtr = BufferEnd;
52  *--BufPtr = 0;      // Null terminate buffer.
53  if (X == 0) {
54    *--BufPtr = '0';  // Handle special case.
55    return BufPtr;
56  }
57
58  while (X) {
59    unsigned char Mod = static_cast<unsigned char>(X) & 15;
60    *--BufPtr = hexdigit(Mod);
61    X >>= 4;
62  }
63  return BufPtr;
64}
65
66static inline std::string utohexstr(uint64_t X) {
67  char Buffer[17];
68  return utohex_buffer(X, Buffer+17);
69}
70
71static inline std::string utostr_32(uint32_t X, bool isNeg = false) {
72  char Buffer[11];
73  char *BufPtr = Buffer+11;
74
75  if (X == 0) *--BufPtr = '0';  // Handle special case...
76
77  while (X) {
78    *--BufPtr = '0' + char(X % 10);
79    X /= 10;
80  }
81
82  if (isNeg) *--BufPtr = '-';   // Add negative sign...
83
84  return std::string(BufPtr, Buffer+11);
85}
86
87static inline std::string utostr(uint64_t X, bool isNeg = false) {
88  char Buffer[21];
89  char *BufPtr = Buffer+21;
90
91  if (X == 0) *--BufPtr = '0';  // Handle special case...
92
93  while (X) {
94    *--BufPtr = '0' + char(X % 10);
95    X /= 10;
96  }
97
98  if (isNeg) *--BufPtr = '-';   // Add negative sign...
99  return std::string(BufPtr, Buffer+21);
100}
101
102
103static inline std::string itostr(int64_t X) {
104  if (X < 0)
105    return utostr(static_cast<uint64_t>(-X), true);
106  else
107    return utostr(static_cast<uint64_t>(X));
108}
109
110/// StrInStrNoCase - Portable version of strcasestr.  Locates the first
111/// occurrence of string 's1' in string 's2', ignoring case.  Returns
112/// the offset of s2 in s1 or npos if s2 cannot be found.
113StringRef::size_type StrInStrNoCase(StringRef s1, StringRef s2);
114
115/// getToken - This function extracts one token from source, ignoring any
116/// leading characters that appear in the Delimiters string, and ending the
117/// token at any of the characters that appear in the Delimiters string.  If
118/// there are no tokens in the source string, an empty string is returned.
119/// The function returns a pair containing the extracted token and the
120/// remaining tail string.
121std::pair<StringRef, StringRef> getToken(StringRef Source,
122                                         StringRef Delimiters = " \t\n\v\f\r");
123
124/// SplitString - Split up the specified string according to the specified
125/// delimiters, appending the result fragments to the output list.
126void SplitString(StringRef Source,
127                 SmallVectorImpl<StringRef> &OutFragments,
128                 StringRef Delimiters = " \t\n\v\f\r");
129
130/// HashString - Hash function for strings.
131///
132/// This is the Bernstein hash function.
133//
134// FIXME: Investigate whether a modified bernstein hash function performs
135// better: http://eternallyconfuzzled.com/tuts/algorithms/jsw_tut_hashing.aspx
136//   X*33+c -> X*33^c
137static inline unsigned HashString(StringRef Str, unsigned Result = 0) {
138  for (unsigned i = 0, e = Str.size(); i != e; ++i)
139    Result = Result * 33 + (unsigned char)Str[i];
140  return Result;
141}
142
143/// Returns the English suffix for an ordinal integer (-st, -nd, -rd, -th).
144static inline StringRef getOrdinalSuffix(unsigned Val) {
145  // It is critically important that we do this perfectly for
146  // user-written sequences with over 100 elements.
147  switch (Val % 100) {
148  case 11:
149  case 12:
150  case 13:
151    return "th";
152  default:
153    switch (Val % 10) {
154      case 1: return "st";
155      case 2: return "nd";
156      case 3: return "rd";
157      default: return "th";
158    }
159  }
160}
161
162} // End llvm namespace
163
164#endif
165