StringExtras.h revision 9bd34602fbe945cc2ed81f8f8ac7de59f4679038 
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/Support/DataTypes.h"
18#include "llvm/ADT/APFloat.h"
19#include "llvm/ADT/StringRef.h"
20#include <cctype>
21#include <cstdio>
22#include <string>
23#include <vector>
24
25namespace llvm {
26
27/// hexdigit - Return the (uppercase) hexadecimal character for the
28/// given number \arg X (which should be less than 16).
29static inline char hexdigit(unsigned X) {
30  return X < 10 ? '0' + X : 'A' + X - 10;
31}
32
33/// utohex_buffer - Emit the specified number into the buffer specified by
34/// BufferEnd, returning a pointer to the start of the string.  This can be used
35/// like this: (note that the buffer must be large enough to handle any number):
36///    char Buffer[40];
37///    printf("0x%s", utohex_buffer(X, Buffer+40));
38///
39/// This should only be used with unsigned types.
40///
41template<typename IntTy>
42static inline char *utohex_buffer(IntTy X, char *BufferEnd) {
43  char *BufPtr = BufferEnd;
44  *--BufPtr = 0;      // Null terminate buffer.
45  if (X == 0) {
46    *--BufPtr = '0';  // Handle special case.
47    return BufPtr;
48  }
49
50  while (X) {
51    unsigned char Mod = static_cast<unsigned char>(X) & 15;
52    *--BufPtr = hexdigit(Mod);
53    X >>= 4;
54  }
55  return BufPtr;
56}
57
58static inline std::string utohexstr(uint64_t X) {
59  char Buffer[40];
60  return utohex_buffer(X, Buffer+40);
61}
62
63static inline std::string utostr_32(uint32_t X, bool isNeg = false) {
64  char Buffer[20];
65  char *BufPtr = Buffer+19;
66
67  *BufPtr = 0;                  // Null terminate buffer...
68  if (X == 0) *--BufPtr = '0';  // Handle special case...
69
70  while (X) {
71    *--BufPtr = '0' + char(X % 10);
72    X /= 10;
73  }
74
75  if (isNeg) *--BufPtr = '-';   // Add negative sign...
76
77  return std::string(BufPtr);
78}
79
80static inline std::string utostr(uint64_t X, bool isNeg = false) {
81  if (X == uint32_t(X))
82    return utostr_32(uint32_t(X), isNeg);
83
84  char Buffer[40];
85  char *BufPtr = Buffer+39;
86
87  *BufPtr = 0;                  // Null terminate buffer...
88  if (X == 0) *--BufPtr = '0';  // Handle special case...
89
90  while (X) {
91    *--BufPtr = '0' + char(X % 10);
92    X /= 10;
93  }
94
95  if (isNeg) *--BufPtr = '-';   // Add negative sign...
96  return std::string(BufPtr);
97}
98
99
100static inline std::string itostr(int64_t X) {
101  if (X < 0)
102    return utostr(static_cast<uint64_t>(-X), true);
103  else
104    return utostr(static_cast<uint64_t>(X));
105}
106
107static inline std::string ftostr(double V) {
108  char Buffer[200];
109  sprintf(Buffer, "%20.6e", V);
110  char *B = Buffer;
111  while (*B == ' ') ++B;
112  return B;
113}
114
115static inline std::string ftostr(const APFloat& V) {
116  if (&V.getSemantics() == &APFloat::IEEEdouble)
117    return ftostr(V.convertToDouble());
118  else if (&V.getSemantics() == &APFloat::IEEEsingle)
119    return ftostr((double)V.convertToFloat());
120  return "<unknown format in ftostr>"; // error
121}
122
123static inline std::string LowercaseString(const std::string &S) {
124  std::string result(S);
125  for (unsigned i = 0; i < S.length(); ++i)
126    if (isupper(result[i]))
127      result[i] = char(tolower(result[i]));
128  return result;
129}
130
131static inline std::string UppercaseString(const std::string &S) {
132  std::string result(S);
133  for (unsigned i = 0; i < S.length(); ++i)
134    if (islower(result[i]))
135      result[i] = char(toupper(result[i]));
136  return result;
137}
138
139/// StringsEqualNoCase - Return true if the two strings are equal, ignoring
140/// case.
141static inline bool StringsEqualNoCase(const std::string &LHS,
142                                      const std::string &RHS) {
143  if (LHS.size() != RHS.size()) return false;
144  for (unsigned i = 0, e = static_cast<unsigned>(LHS.size()); i != e; ++i)
145    if (tolower(LHS[i]) != tolower(RHS[i])) return false;
146  return true;
147}
148
149/// StringsEqualNoCase - Return true if the two strings are equal, ignoring
150/// case.
151static inline bool StringsEqualNoCase(const std::string &LHS,
152                                      const char *RHS) {
153  for (unsigned i = 0, e = static_cast<unsigned>(LHS.size()); i != e; ++i) {
154    if (RHS[i] == 0) return false;  // RHS too short.
155    if (tolower(LHS[i]) != tolower(RHS[i])) return false;
156  }
157  return RHS[LHS.size()] == 0;  // Not too long?
158}
159
160/// StringsEqualNoCase - Return true if the two null-terminated C strings are
161///  equal, ignoring
162
163static inline bool StringsEqualNoCase(const char *LHS, const char *RHS,
164                                      unsigned len) {
165
166  for (unsigned i = 0; i < len; ++i) {
167    if (tolower(LHS[i]) != tolower(RHS[i]))
168      return false;
169
170    // If RHS[i] == 0 then LHS[i] == 0 or otherwise we would have returned
171    // at the previous branch as tolower('\0') == '\0'.
172    if (RHS[i] == 0)
173      return true;
174  }
175
176  return true;
177}
178
179/// CStrInCStrNoCase - Portable version of strcasestr.  Locates the first
180///  occurance of c-string 's2' in string 's1', ignoring case.  Returns
181///  NULL if 's2' cannot be found.
182static inline const char* CStrInCStrNoCase(const char *s1, const char *s2) {
183
184  // Are either strings NULL or empty?
185  if (!s1 || !s2 || s1[0] == '\0' || s2[0] == '\0')
186    return 0;
187
188  if (s1 == s2)
189    return s1;
190
191  const char *I1=s1, *I2=s2;
192
193  while (*I1 != '\0' && *I2 != '\0' )
194    if (tolower(*I1) != tolower(*I2)) { // No match.  Start over.
195      ++s1; I1 = s1; I2 = s2;
196    }
197    else { // Character match.  Advance to the next character.
198      ++I1; ++I2;
199    }
200
201  // If we exhausted all of the characters in 's2', then 's2' appears in 's1'.
202  return *I2 == '\0' ? s1 : 0;
203}
204
205/// getToken - This function extracts one token from source, ignoring any
206/// leading characters that appear in the Delimiters string, and ending the
207/// token at any of the characters that appear in the Delimiters string.  If
208/// there are no tokens in the source string, an empty string is returned.
209/// The Source source string is updated in place to remove the returned string
210/// and any delimiter prefix from it.
211std::string getToken(std::string &Source,
212                     const char *Delimiters = " \t\n\v\f\r");
213
214/// SplitString - Split up the specified string according to the specified
215/// delimiters, appending the result fragments to the output list.
216void SplitString(const std::string &Source,
217                 std::vector<std::string> &OutFragments,
218                 const char *Delimiters = " \t\n\v\f\r");
219
220/// HashString - Hash funtion for strings.
221///
222/// This is the Bernstein hash function.
223//
224// FIXME: Investigate whether a modified bernstein hash function performs
225// better: http://eternallyconfuzzled.com/tuts/algorithms/jsw_tut_hashing.aspx
226//   X*33+c -> X*33^c
227static inline unsigned HashString(StringRef Str, unsigned Result = 0) {
228  for (unsigned i = 0, e = Str.size(); i != e; ++i)
229    Result = Result * 33 + Str[i];
230  return Result;
231}
232
233} // End llvm namespace
234
235#endif
236