1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
|
/*
* Copyright (C) 2011, 2012 Apple Inc. All rights reserved.
* Copyright (C) 2011 Nokia Corporation and/or its subsidiary(-ies).
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public License
* along with this library; see the file COPYING.LIB. If not, write to
* the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
* Boston, MA 02110-1301, USA.
*
*/
#pragma once
#include <stdint.h>
#include <unicode/utypes.h>
#include <wtf/BitSet.h>
#include <wtf/StdLibExtras.h>
#include <wtf/text/LChar.h>
#if CPU(X86_SSE2)
#include <emmintrin.h>
#endif
namespace WTF {
template<unsigned charactersCount>
inline constexpr BitSet<256> makeLatin1CharacterBitSet(const char (&characters)[charactersCount])
{
static_assert(charactersCount > 0, "Since string literal is null terminated, characterCount is always larger than 0");
BitSet<256> bitmap;
for (unsigned i = 0; i < charactersCount - 1; ++i)
bitmap.set(characters[i]);
return bitmap;
}
inline constexpr BitSet<256> makeLatin1CharacterBitSet(NOESCAPE const Invocable<bool(LChar)> auto& matches)
{
BitSet<256> bitmap;
for (unsigned i = 0; i < bitmap.size(); ++i) {
if (matches(static_cast<LChar>(i)))
bitmap.set(i);
}
return bitmap;
}
template <uintptr_t mask>
inline bool isAlignedTo(const void* pointer)
{
return !(reinterpret_cast<uintptr_t>(pointer) & mask);
}
// Assuming that a pointer is the size of a "machine word", then
// uintptr_t is an integer type that is also a machine word.
typedef uintptr_t MachineWord;
const uintptr_t machineWordAlignmentMask = sizeof(MachineWord) - 1;
inline bool isAlignedToMachineWord(const void* pointer)
{
return isAlignedTo<machineWordAlignmentMask>(pointer);
}
template<typename T> inline T* alignToMachineWord(T* pointer)
{
return reinterpret_cast<T*>(reinterpret_cast<uintptr_t>(pointer) & ~machineWordAlignmentMask);
}
template<size_t size, typename CharacterType> struct NonASCIIMask;
template<> struct NonASCIIMask<4, UChar> {
static inline uint32_t value() { return 0xFF80FF80U; }
};
template<> struct NonASCIIMask<4, LChar> {
static inline uint32_t value() { return 0x80808080U; }
};
template<> struct NonASCIIMask<4, char8_t> {
static inline uint32_t value() { return 0x80808080U; }
};
template<> struct NonASCIIMask<8, UChar> {
static inline uint64_t value() { return 0xFF80FF80FF80FF80ULL; }
};
template<> struct NonASCIIMask<8, LChar> {
static inline uint64_t value() { return 0x8080808080808080ULL; }
};
template<> struct NonASCIIMask<8, char8_t> {
static inline uint64_t value() { return 0x8080808080808080ULL; }
};
template<size_t size, typename CharacterType> struct NonLatin1Mask;
template<> struct NonLatin1Mask<4, UChar> {
static inline uint32_t value() { return 0xFF00FF00U; }
};
template<> struct NonLatin1Mask<8, UChar> {
static inline uint64_t value() { return 0xFF00FF00FF00FF00ULL; }
};
template<typename CharacterType>
inline bool containsOnlyASCII(MachineWord word)
{
return !(word & NonASCIIMask<sizeof(MachineWord), CharacterType>::value());
}
// Note: This function assume the input is likely all ASCII, and
// does not leave early if it is not the case.
template<typename CharacterType>
inline bool charactersAreAllASCII(std::span<const CharacterType> span)
{
MachineWord allCharBits = 0;
// Prologue: align the input.
while (!span.empty() && !isAlignedToMachineWord(span.data()))
allCharBits |= WTF::consume(span);
// Compare the values of CPU word size.
size_t sizeAfterAlignedEnd = std::to_address(span.end()) - alignToMachineWord(std::to_address(span.end()));
const size_t loopIncrement = sizeof(MachineWord) / sizeof(CharacterType);
while (span.size() > sizeAfterAlignedEnd)
allCharBits |= reinterpretCastSpanStartTo<const MachineWord>(consumeSpan(span, loopIncrement));
// Process the remaining bytes.
while (!span.empty())
allCharBits |= WTF::consume(span);
MachineWord nonASCIIBitMask = NonASCIIMask<sizeof(MachineWord), CharacterType>::value();
return !(allCharBits & nonASCIIBitMask);
}
// Note: This function assume the input is likely all Latin1, and
// does not leave early if it is not the case.
template<typename CharacterType>
inline bool charactersAreAllLatin1(std::span<const CharacterType> span)
{
if constexpr (sizeof(CharacterType) == 1)
return true;
else {
MachineWord allCharBits = 0;
// Prologue: align the input.
while (!span.empty() && !isAlignedToMachineWord(span.data()))
allCharBits |= WTF::consume(span);
// Compare the values of CPU word size.
size_t sizeAfterAlignedEnd = std::to_address(span.end()) - alignToMachineWord(std::to_address(span.end()));
const size_t loopIncrement = sizeof(MachineWord) / sizeof(CharacterType);
while (span.size() > sizeAfterAlignedEnd)
allCharBits |= reinterpretCastSpanStartTo<const MachineWord>(consumeSpan(span, loopIncrement));
// Process the remaining bytes.
while (!span.empty())
allCharBits |= WTF::consume(span);
MachineWord nonLatin1BitMask = NonLatin1Mask<sizeof(MachineWord), CharacterType>::value();
return !(allCharBits & nonLatin1BitMask);
}
}
} // namespace WTF
using WTF::charactersAreAllASCII;
using WTF::makeLatin1CharacterBitSet;
|
