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 *strcpy16(char16_t *, const char16_t *);
31 char16_t *strstr16(const char16_t*, const char16_t*);
32 
33 // Version of comparison that supports embedded NULs.
34 // This is different than strncmp() because we don't stop
35 // at a nul character and consider the strings to be different
36 // if the lengths are different (thus we need to supply the
37 // lengths of both strings).  This can also be used when
38 // your string is not nul-terminated as it will have the
39 // equivalent result as strcmp16 (unlike strncmp16).
40 int strzcmp16(const char16_t *s1, size_t n1, const char16_t *s2, size_t n2);
41 
42 // Standard string functions on char32_t strings.
43 size_t strlen32(const char32_t *);
44 size_t strnlen32(const char32_t *, size_t);
45 
46 /**
47  * Measure the length of a UTF-32 string in UTF-8. If the string is invalid
48  * such as containing a surrogate character, -1 will be returned.
49  */
50 ssize_t utf32_to_utf8_length(const char32_t *src, size_t src_len);
51 
52 /**
53  * Stores a UTF-8 string converted from "src" in "dst", if "dst_length" is not
54  * large enough to store the string, the part of the "src" string is stored
55  * into "dst" as much as possible. See the examples for more detail.
56  * Returns the size actually used for storing the string.
57  * dst" is not nul-terminated when dst_len is fully used (like strncpy).
58  *
59  * \code
60  * Example 1
61  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
62  * "src_len" == 2
63  * "dst_len" >= 7
64  * ->
65  * Returned value == 6
66  * "dst" becomes \xE3\x81\x82\xE3\x81\x84\0
67  * (note that "dst" is nul-terminated)
68  *
69  * Example 2
70  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
71  * "src_len" == 2
72  * "dst_len" == 5
73  * ->
74  * Returned value == 3
75  * "dst" becomes \xE3\x81\x82\0
76  * (note that "dst" is nul-terminated, but \u3044 is not stored in "dst"
77  * since "dst" does not have enough size to store the character)
78  *
79  * Example 3
80  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
81  * "src_len" == 2
82  * "dst_len" == 6
83  * ->
84  * Returned value == 6
85  * "dst" becomes \xE3\x81\x82\xE3\x81\x84
86  * (note that "dst" is NOT nul-terminated, like strncpy)
87  * \endcode
88  */
89 void utf32_to_utf8(const char32_t* src, size_t src_len, char* dst, size_t dst_len);
90 
91 /**
92  * Returns the unicode value at "index".
93  * Returns -1 when the index is invalid (equals to or more than "src_len").
94  * If returned value is positive, it is able to be converted to char32_t, which
95  * is unsigned. Then, if "next_index" is not NULL, the next index to be used is
96  * stored in "next_index". "next_index" can be NULL.
97  */
98 int32_t utf32_from_utf8_at(const char *src, size_t src_len, size_t index, size_t *next_index);
99 
100 
101 /**
102  * Returns the UTF-8 length of UTF-16 string "src".
103  */
104 ssize_t utf16_to_utf8_length(const char16_t *src, size_t src_len);
105 
106 /**
107  * Converts a UTF-16 string to UTF-8. The destination buffer must be large
108  * enough to fit the UTF-16 as measured by utf16_to_utf8_length with an added
109  * NUL terminator.
110  */
111 void utf16_to_utf8(const char16_t* src, size_t src_len, char* dst, size_t dst_len);
112 
113 /**
114  * Returns the length of "src" when "src" is valid UTF-8 string.
115  * Returns 0 if src is NULL or 0-length string. Returns -1 when the source
116  * is an invalid string.
117  *
118  * This function should be used to determine whether "src" is valid UTF-8
119  * characters with valid unicode codepoints. "src" must be nul-terminated.
120  *
121  * If you are going to use other utf8_to_... functions defined in this header
122  * with string which may not be valid UTF-8 with valid codepoint (form 0 to
123  * 0x10FFFF), you should use this function before calling others, since the
124  * other functions do not check whether the string is valid UTF-8 or not.
125  *
126  * If you do not care whether "src" is valid UTF-8 or not, you should use
127  * strlen() as usual, which should be much faster.
128  */
129 ssize_t utf8_length(const char *src);
130 
131 /**
132  * Returns the UTF-16 length of UTF-8 string "src". Returns -1 in case
133  * it's invalid utf8. No buffer over-read occurs because of bound checks. Using overreadIsFatal you
134  * can ask to log a message and fail in case the invalid utf8 could have caused an override if no
135  * bound checks were used (otherwise -1 is returned).
136  */
137 ssize_t utf8_to_utf16_length(const uint8_t* src, size_t srcLen, bool overreadIsFatal = false);
138 
139 /**
140  * Convert UTF-8 to UTF-16 including surrogate pairs.
141  * Returns a pointer to the end of the string (where a NUL terminator might go
142  * if you wanted to add one). At most dstLen characters are written; it won't emit half a surrogate
143  * pair. If dstLen == 0 nothing is written and dst is returned. If dstLen > SSIZE_MAX it aborts
144  * (this being probably a negative number returned as an error and casted to unsigned).
145  */
146 char16_t* utf8_to_utf16_no_null_terminator(
147         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
148 
149 /**
150  * Convert UTF-8 to UTF-16 including surrogate pairs. At most dstLen - 1
151  * characters are written; it won't emit half a surrogate pair; and a NUL terminator is appended
152  * after. dstLen - 1 can be measured beforehand using utf8_to_utf16_length. Aborts if dstLen == 0
153  * (at least one character is needed for the NUL terminator) or dstLen > SSIZE_MAX (the latter
154  * case being likely a negative number returned as an error and casted to unsigned) . Returns a
155  * pointer to the NUL terminator.
156  */
157 char16_t *utf8_to_utf16(
158         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
159 
160 }
161 
162 #endif
163