1 //===--- UnicodeCharRanges.h - Types and functions for character ranges ---===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 #ifndef LLVM_SUPPORT_UNICODECHARRANGES_H
9 #define LLVM_SUPPORT_UNICODECHARRANGES_H
10 
11 #include "llvm/ADT/ArrayRef.h"
12 #include "llvm/Support/Compiler.h"
13 #include "llvm/Support/Debug.h"
14 #include "llvm/Support/raw_ostream.h"
15 #include <algorithm>
16 
17 #define DEBUG_TYPE "unicode"
18 
19 namespace llvm {
20 namespace sys {
21 
22 /// Represents a closed range of Unicode code points [Lower, Upper].
23 struct UnicodeCharRange {
24   uint32_t Lower;
25   uint32_t Upper;
26 };
27 
28 inline bool operator<(uint32_t Value, UnicodeCharRange Range) {
29   return Value < Range.Lower;
30 }
31 inline bool operator<(UnicodeCharRange Range, uint32_t Value) {
32   return Range.Upper < Value;
33 }
34 
35 /// Holds a reference to an ordered array of UnicodeCharRange and allows
36 /// to quickly check if a code point is contained in the set represented by this
37 /// array.
38 class UnicodeCharSet {
39 public:
40   typedef ArrayRef<UnicodeCharRange> CharRanges;
41 
42   /// Constructs a UnicodeCharSet instance from an array of
43   /// UnicodeCharRanges.
44   ///
45   /// Array pointed by \p Ranges should have the lifetime at least as long as
46   /// the UnicodeCharSet instance, and should not change. Array is validated by
47   /// the constructor, so it makes sense to create as few UnicodeCharSet
48   /// instances per each array of ranges, as possible.
49 #ifdef NDEBUG
50 
51   // FIXME: This could use constexpr + static_assert. This way we
52   // may get rid of NDEBUG in this header. Unfortunately there are some
53   // problems to get this working with MSVC 2013. Change this when
54   // the support for MSVC 2013 is dropped.
55   constexpr UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) {}
56 #else
57   UnicodeCharSet(CharRanges Ranges) : Ranges(Ranges) {
58     assert(rangesAreValid());
59   }
60 #endif
61 
62   /// Returns true if the character set contains the Unicode code point
63   /// \p C.
64   bool contains(uint32_t C) const {
65     return std::binary_search(Ranges.begin(), Ranges.end(), C);
66   }
67 
68 private:
69   /// Returns true if each of the ranges is a proper closed range
70   /// [min, max], and if the ranges themselves are ordered and non-overlapping.
71   bool rangesAreValid() const {
72     uint32_t Prev = 0;
73     for (CharRanges::const_iterator I = Ranges.begin(), E = Ranges.end();
74          I != E; ++I) {
75       if (I != Ranges.begin() && Prev >= I->Lower) {
76         LLVM_DEBUG(dbgs() << "Upper bound 0x");
77         LLVM_DEBUG(dbgs().write_hex(Prev));
78         LLVM_DEBUG(dbgs() << " should be less than succeeding lower bound 0x");
79         LLVM_DEBUG(dbgs().write_hex(I->Lower) << "\n");
80         return false;
81       }
82       if (I->Upper < I->Lower) {
83         LLVM_DEBUG(dbgs() << "Upper bound 0x");
84         LLVM_DEBUG(dbgs().write_hex(I->Lower));
85         LLVM_DEBUG(dbgs() << " should not be less than lower bound 0x");
86         LLVM_DEBUG(dbgs().write_hex(I->Upper) << "\n");
87         return false;
88       }
89       Prev = I->Upper;
90     }
91 
92     return true;
93   }
94 
95   const CharRanges Ranges;
96 };
97 
98 } // namespace sys
99 } // namespace llvm
100 
101 #undef DEBUG_TYPE // "unicode"
102 
103 #endif // LLVM_SUPPORT_UNICODECHARRANGES_H
104