1 //===--- UnicodeCharRanges.h - Types and functions for character ranges ---===// 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 #ifndef LLVM_SUPPORT_UNICODECHARRANGES_H 10 #define LLVM_SUPPORT_UNICODECHARRANGES_H 11 12 #include "llvm/ADT/ArrayRef.h" 13 #include "llvm/ADT/SmallPtrSet.h" 14 #include "llvm/Support/Compiler.h" 15 #include "llvm/Support/Debug.h" 16 #include "llvm/Support/Mutex.h" 17 #include "llvm/Support/MutexGuard.h" 18 #include "llvm/Support/raw_ostream.h" 19 #include <algorithm> 20 21 namespace llvm { 22 namespace sys { 23 24 #define DEBUG_TYPE "unicode" 25 26 /// \brief Represents a closed range of Unicode code points [Lower, Upper]. 27 struct UnicodeCharRange { 28 uint32_t Lower; 29 uint32_t Upper; 30 }; 31 32 inline bool operator<(uint32_t Value, UnicodeCharRange Range) { 33 return Value < Range.Lower; 34 } 35 inline bool operator<(UnicodeCharRange Range, uint32_t Value) { 36 return Range.Upper < Value; 37 } 38 39 /// \brief Holds a reference to an ordered array of UnicodeCharRange and allows 40 /// to quickly check if a code point is contained in the set represented by this 41 /// array. 42 class UnicodeCharSet { 43 public: 44 typedef ArrayRef<UnicodeCharRange> CharRanges; 45 46 /// \brief Constructs a UnicodeCharSet instance from an array of 47 /// UnicodeCharRanges. 48 /// 49 /// Array pointed by \p Ranges should have the lifetime at least as long as 50 /// the UnicodeCharSet instance, and should not change. Array is validated by 51 /// the constructor, so it makes sense to create as few UnicodeCharSet 52 /// instances per each array of ranges, as possible. 53 #ifdef NDEBUG 54 LLVM_CONSTEXPR UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) {} 55 #else 56 UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) { 57 assert(rangesAreValid()); 58 } 59 #endif 60 61 /// \brief Returns true if the character set contains the Unicode code point 62 /// \p C. 63 bool contains(uint32_t C) const { 64 return std::binary_search(Ranges.begin(), Ranges.end(), C); 65 } 66 67 private: 68 /// \brief Returns true if each of the ranges is a proper closed range 69 /// [min, max], and if the ranges themselves are ordered and non-overlapping. 70 bool rangesAreValid() const { 71 uint32_t Prev = 0; 72 for (CharRanges::const_iterator I = Ranges.begin(), E = Ranges.end(); 73 I != E; ++I) { 74 if (I != Ranges.begin() && Prev >= I->Lower) { 75 DEBUG(dbgs() << "Upper bound 0x"); 76 DEBUG(dbgs().write_hex(Prev)); 77 DEBUG(dbgs() << " should be less than succeeding lower bound 0x"); 78 DEBUG(dbgs().write_hex(I->Lower) << "\n"); 79 return false; 80 } 81 if (I->Upper < I->Lower) { 82 DEBUG(dbgs() << "Upper bound 0x"); 83 DEBUG(dbgs().write_hex(I->Lower)); 84 DEBUG(dbgs() << " should not be less than lower bound 0x"); 85 DEBUG(dbgs().write_hex(I->Upper) << "\n"); 86 return false; 87 } 88 Prev = I->Upper; 89 } 90 91 return true; 92 } 93 94 const CharRanges Ranges; 95 }; 96 97 #undef DEBUG_TYPE // "unicode" 98 99 } // namespace sys 100 } // namespace llvm 101 102 103 #endif // LLVM_SUPPORT_UNICODECHARRANGES_H 104