1 /*
2  * Copyright (C) 2005 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  */
16 
17 #ifndef ANDROID_UNICODE_H
18 #define ANDROID_UNICODE_H
19 
20 #include <sys/types.h>
21 #include <stdint.h>
22 
23 extern "C" {
24 
25 // Standard string functions on char16_t strings.
26 int strcmp16(const char16_t *, const char16_t *);
27 int strncmp16(const char16_t *s1, const char16_t *s2, size_t n);
28 size_t strlen16(const char16_t *);
29 size_t strnlen16(const char16_t *, size_t);
30 char16_t *strstr16(const char16_t*, const char16_t*);
31 
32 // Version of comparison that supports embedded NULs.
33 // This is different than strncmp() because we don't stop
34 // at a nul character and consider the strings to be different
35 // if the lengths are different (thus we need to supply the
36 // lengths of both strings).  This can also be used when
37 // your string is not nul-terminated as it will have the
38 // equivalent result as strcmp16 (unlike strncmp16).
39 int strzcmp16(const char16_t *s1, size_t n1, const char16_t *s2, size_t n2);
40 
41 /**
42  * Measure the length of a UTF-32 string in UTF-8. If the string is invalid
43  * such as containing a surrogate character, -1 will be returned.
44  */
45 ssize_t utf32_to_utf8_length(const char32_t *src, size_t src_len);
46 
47 /**
48  * Stores a UTF-8 string converted from "src" in "dst", if "dst_length" is not
49  * large enough to store the string, the part of the "src" string is stored
50  * into "dst" as much as possible. See the examples for more detail.
51  * Returns the size actually used for storing the string.
52  * dst" is not nul-terminated when dst_len is fully used (like strncpy).
53  *
54  * \code
55  * Example 1
56  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
57  * "src_len" == 2
58  * "dst_len" >= 7
59  * ->
60  * Returned value == 6
61  * "dst" becomes \xE3\x81\x82\xE3\x81\x84\0
62  * (note that "dst" is nul-terminated)
63  *
64  * Example 2
65  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
66  * "src_len" == 2
67  * "dst_len" == 5
68  * ->
69  * Returned value == 3
70  * "dst" becomes \xE3\x81\x82\0
71  * (note that "dst" is nul-terminated, but \u3044 is not stored in "dst"
72  * since "dst" does not have enough size to store the character)
73  *
74  * Example 3
75  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
76  * "src_len" == 2
77  * "dst_len" == 6
78  * ->
79  * Returned value == 6
80  * "dst" becomes \xE3\x81\x82\xE3\x81\x84
81  * (note that "dst" is NOT nul-terminated, like strncpy)
82  * \endcode
83  */
84 void utf32_to_utf8(const char32_t* src, size_t src_len, char* dst, size_t dst_len);
85 
86 /**
87  * Returns the unicode value at "index".
88  * Returns -1 when the index is invalid (equals to or more than "src_len").
89  * If returned value is positive, it is able to be converted to char32_t, which
90  * is unsigned. Then, if "next_index" is not NULL, the next index to be used is
91  * stored in "next_index". "next_index" can be NULL.
92  */
93 int32_t utf32_from_utf8_at(const char *src, size_t src_len, size_t index, size_t *next_index);
94 
95 
96 /**
97  * Returns the UTF-8 length of UTF-16 string "src".
98  */
99 ssize_t utf16_to_utf8_length(const char16_t *src, size_t src_len);
100 
101 /**
102  * Converts a UTF-16 string to UTF-8. The destination buffer must be large
103  * enough to fit the UTF-16 as measured by utf16_to_utf8_length with an added
104  * NUL terminator.
105  */
106 void utf16_to_utf8(const char16_t* src, size_t src_len, char* dst, size_t dst_len);
107 
108 /**
109  * Returns the UTF-16 length of UTF-8 string "src". Returns -1 in case
110  * it's invalid utf8. No buffer over-read occurs because of bound checks. Using overreadIsFatal you
111  * can ask to log a message and fail in case the invalid utf8 could have caused an override if no
112  * bound checks were used (otherwise -1 is returned).
113  */
114 ssize_t utf8_to_utf16_length(const uint8_t* src, size_t srcLen, bool overreadIsFatal = false);
115 
116 /**
117  * Convert UTF-8 to UTF-16 including surrogate pairs.
118  * Returns a pointer to the end of the string (where a NUL terminator might go
119  * if you wanted to add one). At most dstLen characters are written; it won't emit half a surrogate
120  * pair. If dstLen == 0 nothing is written and dst is returned. If dstLen > SSIZE_MAX it aborts
121  * (this being probably a negative number returned as an error and casted to unsigned).
122  */
123 char16_t* utf8_to_utf16_no_null_terminator(
124         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
125 
126 /**
127  * Convert UTF-8 to UTF-16 including surrogate pairs. At most dstLen - 1
128  * characters are written; it won't emit half a surrogate pair; and a NUL terminator is appended
129  * after. dstLen - 1 can be measured beforehand using utf8_to_utf16_length. Aborts if dstLen == 0
130  * (at least one character is needed for the NUL terminator) or dstLen > SSIZE_MAX (the latter
131  * case being likely a negative number returned as an error and casted to unsigned) . Returns a
132  * pointer to the NUL terminator.
133  */
134 char16_t *utf8_to_utf16(
135         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
136 
137 }
138 
139 #endif
140