Home | History | Annotate | Download | only in ADT 
      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/StringRef.h"
     19 
     20 namespace llvm {
     21 template<typename T> class SmallVectorImpl;
     22 
     23 /// hexdigit - Return the hexadecimal character for the
     24 /// given number \arg X (which should be less than 16).
     25 static 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 /// utohex_buffer - Emit the specified number into the buffer specified by
     31 /// BufferEnd, returning a pointer to the start of the string.  This can be used
     32 /// like this: (note that the buffer must be large enough to handle any number):
     33 ///    char Buffer[40];
     34 ///    printf("0x%s", utohex_buffer(X, Buffer+40));
     35 ///
     36 /// This should only be used with unsigned types.
     37 ///
     38 template<typename IntTy>
     39 static inline char *utohex_buffer(IntTy X, char *BufferEnd) {
     40   char *BufPtr = BufferEnd;
     41   *--BufPtr = 0;      // Null terminate buffer.
     42   if (X == 0) {
     43     *--BufPtr = '0';  // Handle special case.
     44     return BufPtr;
     45   }
     46 
     47   while (X) {
     48     unsigned char Mod = static_cast<unsigned char>(X) & 15;
     49     *--BufPtr = hexdigit(Mod);
     50     X >>= 4;
     51   }
     52   return BufPtr;
     53 }
     54 
     55 static inline std::string utohexstr(uint64_t X) {
     56   char Buffer[17];
     57   return utohex_buffer(X, Buffer+17);
     58 }
     59 
     60 static inline std::string utostr_32(uint32_t X, bool isNeg = false) {
     61   char Buffer[11];
     62   char *BufPtr = Buffer+11;
     63 
     64   if (X == 0) *--BufPtr = '0';  // Handle special case...
     65 
     66   while (X) {
     67     *--BufPtr = '0' + char(X % 10);
     68     X /= 10;
     69   }
     70 
     71   if (isNeg) *--BufPtr = '-';   // Add negative sign...
     72 
     73   return std::string(BufPtr, Buffer+11);
     74 }
     75 
     76 static inline std::string utostr(uint64_t X, bool isNeg = false) {
     77   char Buffer[21];
     78   char *BufPtr = Buffer+21;
     79 
     80   if (X == 0) *--BufPtr = '0';  // Handle special case...
     81 
     82   while (X) {
     83     *--BufPtr = '0' + char(X % 10);
     84     X /= 10;
     85   }
     86 
     87   if (isNeg) *--BufPtr = '-';   // Add negative sign...
     88   return std::string(BufPtr, Buffer+21);
     89 }
     90 
     91 
     92 static inline std::string itostr(int64_t X) {
     93   if (X < 0)
     94     return utostr(static_cast<uint64_t>(-X), true);
     95   else
     96     return utostr(static_cast<uint64_t>(X));
     97 }
     98 
     99 /// StrInStrNoCase - Portable version of strcasestr.  Locates the first
    100 /// occurrence of string 's1' in string 's2', ignoring case.  Returns
    101 /// the offset of s2 in s1 or npos if s2 cannot be found.
    102 StringRef::size_type StrInStrNoCase(StringRef s1, StringRef s2);
    103 
    104 /// getToken - This function extracts one token from source, ignoring any
    105 /// leading characters that appear in the Delimiters string, and ending the
    106 /// token at any of the characters that appear in the Delimiters string.  If
    107 /// there are no tokens in the source string, an empty string is returned.
    108 /// The function returns a pair containing the extracted token and the
    109 /// remaining tail string.
    110 std::pair<StringRef, StringRef> getToken(StringRef Source,
    111                                          StringRef Delimiters = " \t\n\v\f\r");
    112 
    113 /// SplitString - Split up the specified string according to the specified
    114 /// delimiters, appending the result fragments to the output list.
    115 void SplitString(StringRef Source,
    116                  SmallVectorImpl<StringRef> &OutFragments,
    117                  StringRef Delimiters = " \t\n\v\f\r");
    118 
    119 /// HashString - Hash function for strings.
    120 ///
    121 /// This is the Bernstein hash function.
    122 //
    123 // FIXME: Investigate whether a modified bernstein hash function performs
    124 // better: http://eternallyconfuzzled.com/tuts/algorithms/jsw_tut_hashing.aspx
    125 //   X*33+c -> X*33^c
    126 static inline unsigned HashString(StringRef Str, unsigned Result = 0) {
    127   for (unsigned i = 0, e = Str.size(); i != e; ++i)
    128     Result = Result * 33 + (unsigned char)Str[i];
    129   return Result;
    130 }
    131 
    132 } // End llvm namespace
    133 
    134 #endif
    135