blob: 8b430a4b7324a3288073c0574fc8fa129485d680 (
plain) (
blame)
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
|
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
**********************************************************************
* Copyright (C) 2005-2014, International Business Machines
* Corporation and others. All Rights Reserved.
**********************************************************************
*/
#include "unicode/utypes.h"
#if !UCONFIG_NO_CONVERSION
#include "csrutf8.h"
#include "csmatch.h"
U_NAMESPACE_BEGIN
CharsetRecog_UTF8::~CharsetRecog_UTF8()
{
// nothing to do
}
const char *CharsetRecog_UTF8::getName() const
{
return "UTF-8";
}
UBool CharsetRecog_UTF8::match(InputText* input, CharsetMatch *results) const {
bool hasBOM = FALSE;
int32_t numValid = 0;
int32_t numInvalid = 0;
const uint8_t *inputBytes = input->fRawInput;
int32_t i;
int32_t trailBytes = 0;
int32_t confidence;
if (input->fRawLength >= 3 &&
inputBytes[0] == 0xEF && inputBytes[1] == 0xBB && inputBytes[2] == 0xBF) {
hasBOM = TRUE;
}
// Scan for multi-byte sequences
for (i=0; i < input->fRawLength; i += 1) {
int32_t b = inputBytes[i];
if ((b & 0x80) == 0) {
continue; // ASCII
}
// Hi bit on char found. Figure out how long the sequence should be
if ((b & 0x0E0) == 0x0C0) {
trailBytes = 1;
} else if ((b & 0x0F0) == 0x0E0) {
trailBytes = 2;
} else if ((b & 0x0F8) == 0xF0) {
trailBytes = 3;
} else {
numInvalid += 1;
continue;
}
// Verify that we've got the right number of trail bytes in the sequence
for (;;) {
i += 1;
if (i >= input->fRawLength) {
break;
}
b = inputBytes[i];
if ((b & 0xC0) != 0x080) {
numInvalid += 1;
break;
}
if (--trailBytes == 0) {
numValid += 1;
break;
}
}
}
// Cook up some sort of confidence score, based on presence of a BOM
// and the existence of valid and/or invalid multi-byte sequences.
confidence = 0;
if (hasBOM && numInvalid == 0) {
confidence = 100;
} else if (hasBOM && numValid > numInvalid*10) {
confidence = 80;
} else if (numValid > 3 && numInvalid == 0) {
confidence = 100;
} else if (numValid > 0 && numInvalid == 0) {
confidence = 80;
} else if (numValid == 0 && numInvalid == 0) {
// Plain ASCII. Confidence must be > 10, it's more likely than UTF-16, which
// accepts ASCII with confidence = 10.
confidence = 15;
} else if (numValid > numInvalid*10) {
// Probably corruput utf-8 data. Valid sequences aren't likely by chance.
confidence = 25;
}
results->set(input, this, confidence);
return (confidence > 0);
}
U_NAMESPACE_END
#endif
|
