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 *strncpy16(char16_t *, const char16_t *, size_t);
32 char16_t *strstr16(const char16_t*, const char16_t*);
33 
34 // Version of comparison that supports embedded NULs.
35 // This is different than strncmp() because we don't stop
36 // at a nul character and consider the strings to be different
37 // if the lengths are different (thus we need to supply the
38 // lengths of both strings).  This can also be used when
39 // your string is not nul-terminated as it will have the
40 // equivalent result as strcmp16 (unlike strncmp16).
41 int strzcmp16(const char16_t *s1, size_t n1, const char16_t *s2, size_t n2);
42 
43 // Version of strzcmp16 for comparing strings in different endianness.
44 int strzcmp16_h_n(const char16_t *s1H, size_t n1, const char16_t *s2N, size_t n2);
45 
46 // Standard string functions on char32_t strings.
47 size_t strlen32(const char32_t *);
48 size_t strnlen32(const char32_t *, size_t);
49 
50 /**
51  * Measure the length of a UTF-32 string in UTF-8. If the string is invalid
52  * such as containing a surrogate character, -1 will be returned.
53  */
54 ssize_t utf32_to_utf8_length(const char32_t *src, size_t src_len);
55 
56 /**
57  * Stores a UTF-8 string converted from "src" in "dst", if "dst_length" is not
58  * large enough to store the string, the part of the "src" string is stored
59  * into "dst" as much as possible. See the examples for more detail.
60  * Returns the size actually used for storing the string.
61  * dst" is not nul-terminated when dst_len is fully used (like strncpy).
62  *
63  * \code
64  * Example 1
65  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
66  * "src_len" == 2
67  * "dst_len" >= 7
68  * ->
69  * Returned value == 6
70  * "dst" becomes \xE3\x81\x82\xE3\x81\x84\0
71  * (note that "dst" is nul-terminated)
72  *
73  * Example 2
74  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
75  * "src_len" == 2
76  * "dst_len" == 5
77  * ->
78  * Returned value == 3
79  * "dst" becomes \xE3\x81\x82\0
80  * (note that "dst" is nul-terminated, but \u3044 is not stored in "dst"
81  * since "dst" does not have enough size to store the character)
82  *
83  * Example 3
84  * "src" == \u3042\u3044 (\xE3\x81\x82\xE3\x81\x84)
85  * "src_len" == 2
86  * "dst_len" == 6
87  * ->
88  * Returned value == 6
89  * "dst" becomes \xE3\x81\x82\xE3\x81\x84
90  * (note that "dst" is NOT nul-terminated, like strncpy)
91  * \endcode
92  */
93 void utf32_to_utf8(const char32_t* src, size_t src_len, char* dst, size_t dst_len);
94 
95 /**
96  * Returns the unicode value at "index".
97  * Returns -1 when the index is invalid (equals to or more than "src_len").
98  * If returned value is positive, it is able to be converted to char32_t, which
99  * is unsigned. Then, if "next_index" is not NULL, the next index to be used is
100  * stored in "next_index". "next_index" can be NULL.
101  */
102 int32_t utf32_from_utf8_at(const char *src, size_t src_len, size_t index, size_t *next_index);
103 
104 
105 /**
106  * Returns the UTF-8 length of UTF-16 string "src".
107  */
108 ssize_t utf16_to_utf8_length(const char16_t *src, size_t src_len);
109 
110 /**
111  * Converts a UTF-16 string to UTF-8. The destination buffer must be large
112  * enough to fit the UTF-16 as measured by utf16_to_utf8_length with an added
113  * NUL terminator.
114  */
115 void utf16_to_utf8(const char16_t* src, size_t src_len, char* dst, size_t dst_len);
116 
117 /**
118  * Returns the length of "src" when "src" is valid UTF-8 string.
119  * Returns 0 if src is NULL or 0-length string. Returns -1 when the source
120  * is an invalid string.
121  *
122  * This function should be used to determine whether "src" is valid UTF-8
123  * characters with valid unicode codepoints. "src" must be nul-terminated.
124  *
125  * If you are going to use other utf8_to_... functions defined in this header
126  * with string which may not be valid UTF-8 with valid codepoint (form 0 to
127  * 0x10FFFF), you should use this function before calling others, since the
128  * other functions do not check whether the string is valid UTF-8 or not.
129  *
130  * If you do not care whether "src" is valid UTF-8 or not, you should use
131  * strlen() as usual, which should be much faster.
132  */
133 ssize_t utf8_length(const char *src);
134 
135 /**
136  * Measure the length of a UTF-32 string.
137  */
138 size_t utf8_to_utf32_length(const char *src, size_t src_len);
139 
140 /**
141  * Stores a UTF-32 string converted from "src" in "dst". "dst" must be large
142  * enough to store the entire converted string as measured by
143  * utf8_to_utf32_length plus space for a NUL terminator.
144  */
145 void utf8_to_utf32(const char* src, size_t src_len, char32_t* dst);
146 
147 /**
148  * Returns the UTF-16 length of UTF-8 string "src". Returns -1 in case
149  * it's invalid utf8. No buffer over-read occurs because of bound checks. Using overreadIsFatal you
150  * can ask to log a message and fail in case the invalid utf8 could have caused an override if no
151  * bound checks were used (otherwise -1 is returned).
152  */
153 ssize_t utf8_to_utf16_length(const uint8_t* src, size_t srcLen, bool overreadIsFatal = false);
154 
155 /**
156  * Convert UTF-8 to UTF-16 including surrogate pairs.
157  * Returns a pointer to the end of the string (where a NUL terminator might go
158  * if you wanted to add one). At most dstLen characters are written; it won't emit half a surrogate
159  * pair. If dstLen == 0 nothing is written and dst is returned. If dstLen > SSIZE_MAX it aborts
160  * (this being probably a negative number returned as an error and casted to unsigned).
161  */
162 char16_t* utf8_to_utf16_no_null_terminator(
163         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
164 
165 /**
166  * Convert UTF-8 to UTF-16 including surrogate pairs. At most dstLen - 1
167  * characters are written; it won't emit half a surrogate pair; and a NUL terminator is appended
168  * after. dstLen - 1 can be measured beforehand using utf8_to_utf16_length. Aborts if dstLen == 0
169  * (at least one character is needed for the NUL terminator) or dstLen > SSIZE_MAX (the latter
170  * case being likely a negative number returned as an error and casted to unsigned) . Returns a
171  * pointer to the NUL terminator.
172  */
173 char16_t *utf8_to_utf16(
174         const uint8_t* src, size_t srcLen, char16_t* dst, size_t dstLen);
175 
176 }
177 
178 #endif
179