blob: 05b847b806bac68825489fa1da08dff1cd45a5e5 [file] [log] [blame]
Elliott Hughes2faa5f12012-01-30 14:42:07 -08001/*
2 * Copyright (C) 2011 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
Elliott Hughes814e4032011-08-23 12:07:56 -070016
17#include "utf.h"
18
Elliott Hughes07ed66b2012-12-12 18:34:25 -080019#include "base/logging.h"
Ian Rogers2dd0e2c2013-01-24 12:42:14 -080020#include "mirror/array.h"
Ian Rogers4f6ad8a2013-03-18 15:27:28 -070021#include "mirror/object-inl.h"
Ian Rogersa6724902013-09-23 09:23:37 -070022#include "utf-inl.h"
Elliott Hughesb465ab02011-08-24 11:21:21 -070023
Elliott Hughes814e4032011-08-23 12:07:56 -070024namespace art {
25
26size_t CountModifiedUtf8Chars(const char* utf8) {
27 size_t len = 0;
28 int ic;
29 while ((ic = *utf8++) != '\0') {
30 len++;
31 if ((ic & 0x80) == 0) {
32 // one-byte encoding
33 continue;
34 }
35 // two- or three-byte encoding
36 utf8++;
37 if ((ic & 0x20) == 0) {
38 // two-byte encoding
39 continue;
40 }
41 // three-byte encoding
42 utf8++;
43 }
44 return len;
45}
46
47void ConvertModifiedUtf8ToUtf16(uint16_t* utf16_data_out, const char* utf8_data_in) {
48 while (*utf8_data_in != '\0') {
49 *utf16_data_out++ = GetUtf16FromUtf8(&utf8_data_in);
50 }
51}
52
Elliott Hughesb465ab02011-08-24 11:21:21 -070053void ConvertUtf16ToModifiedUtf8(char* utf8_out, const uint16_t* utf16_in, size_t char_count) {
54 while (char_count--) {
55 uint16_t ch = *utf16_in++;
56 if (ch > 0 && ch <= 0x7f) {
57 *utf8_out++ = ch;
58 } else {
59 if (ch > 0x07ff) {
60 *utf8_out++ = (ch >> 12) | 0xe0;
61 *utf8_out++ = ((ch >> 6) & 0x3f) | 0x80;
62 *utf8_out++ = (ch & 0x3f) | 0x80;
63 } else /*(ch > 0x7f || ch == 0)*/ {
64 *utf8_out++ = (ch >> 6) | 0xc0;
65 *utf8_out++ = (ch & 0x3f) | 0x80;
66 }
67 }
68 }
69}
70
Ian Rogersef7d42f2014-01-06 12:55:46 -080071int32_t ComputeUtf16Hash(mirror::CharArray* chars, int32_t offset,
Ian Rogers0cfe1fb2011-08-26 03:29:44 -070072 size_t char_count) {
Ian Rogers8f41dc32014-10-30 15:16:16 -070073 uint32_t hash = 0;
Ian Rogers0cfe1fb2011-08-26 03:29:44 -070074 for (size_t i = 0; i < char_count; i++) {
75 hash = hash * 31 + chars->Get(offset + i);
76 }
Ian Rogers8f41dc32014-10-30 15:16:16 -070077 return static_cast<int32_t>(hash);
Ian Rogers0cfe1fb2011-08-26 03:29:44 -070078}
79
Elliott Hughes814e4032011-08-23 12:07:56 -070080int32_t ComputeUtf16Hash(const uint16_t* chars, size_t char_count) {
Ian Rogers8f41dc32014-10-30 15:16:16 -070081 uint32_t hash = 0;
Elliott Hughes814e4032011-08-23 12:07:56 -070082 while (char_count--) {
83 hash = hash * 31 + *chars++;
84 }
Ian Rogers8f41dc32014-10-30 15:16:16 -070085 return static_cast<int32_t>(hash);
Elliott Hughes814e4032011-08-23 12:07:56 -070086}
87
Mathieu Chartiere7c9a8c2014-11-06 16:35:45 -080088size_t ComputeModifiedUtf8Hash(const char* chars) {
89 size_t hash = 0;
Ian Rogers68b56852014-08-29 20:19:11 -070090 while (*chars != '\0') {
Mathieu Chartiere7c9a8c2014-11-06 16:35:45 -080091 hash = hash * 31 + *chars++;
Ian Rogers68b56852014-08-29 20:19:11 -070092 }
Ian Rogers8f41dc32014-10-30 15:16:16 -070093 return static_cast<int32_t>(hash);
Ian Rogers68b56852014-08-29 20:19:11 -070094}
95
Ian Rogers637c65b2013-05-31 11:46:00 -070096int CompareModifiedUtf8ToUtf16AsCodePointValues(const char* utf8_1, const uint16_t* utf8_2) {
97 for (;;) {
98 if (*utf8_1 == '\0') {
99 return (*utf8_2 == '\0') ? 0 : -1;
100 } else if (*utf8_2 == '\0') {
101 return 1;
102 }
103
104 int c1 = GetUtf16FromUtf8(&utf8_1);
105 int c2 = *utf8_2;
106
107 if (c1 != c2) {
108 return c1 > c2 ? 1 : -1;
109 }
110 }
111}
112
Elliott Hughes814e4032011-08-23 12:07:56 -0700113size_t CountUtf8Bytes(const uint16_t* chars, size_t char_count) {
114 size_t result = 0;
115 while (char_count--) {
116 uint16_t ch = *chars++;
117 if (ch > 0 && ch <= 0x7f) {
118 ++result;
119 } else {
120 if (ch > 0x7ff) {
121 result += 3;
122 } else {
123 result += 2;
124 }
125 }
126 }
127 return result;
128}
129
130} // namespace art