svn_utf_private.h revision 299742
1/** 2 * @copyright 3 * ==================================================================== 4 * Licensed to the Apache Software Foundation (ASF) under one 5 * or more contributor license agreements. See the NOTICE file 6 * distributed with this work for additional information 7 * regarding copyright ownership. The ASF licenses this file 8 * to you under the Apache License, Version 2.0 (the 9 * "License"); you may not use this file except in compliance 10 * with the License. You may obtain a copy of the License at 11 * 12 * http://www.apache.org/licenses/LICENSE-2.0 13 * 14 * Unless required by applicable law or agreed to in writing, 15 * software distributed under the License is distributed on an 16 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY 17 * KIND, either express or implied. See the License for the 18 * specific language governing permissions and limitations 19 * under the License. 20 * ==================================================================== 21 * @endcopyright 22 * 23 * @file svn_utf_private.h 24 * @brief UTF validation and normalization routines 25 */ 26 27#ifndef SVN_UTF_PRIVATE_H 28#define SVN_UTF_PRIVATE_H 29 30#include <apr.h> 31#include <apr_pools.h> 32 33#include "svn_types.h" 34#include "svn_string.h" 35#include "svn_string_private.h" 36 37#ifdef __cplusplus 38extern "C" { 39#endif /* __cplusplus */ 40 41 42/* Return TRUE if the string SRC of length LEN is a valid UTF-8 encoding 43 * according to the rules laid down by the Unicode 4.0 standard, FALSE 44 * otherwise. This function is faster than svn_utf__last_valid(). 45 */ 46svn_boolean_t 47svn_utf__is_valid(const char *src, apr_size_t len); 48 49/* As for svn_utf__is_valid but SRC is NULL terminated. */ 50svn_boolean_t 51svn_utf__cstring_is_valid(const char *src); 52 53/* Return a pointer to the first character after the last valid UTF-8 54 * potentially multi-byte character in the string SRC of length LEN. 55 * Validity of bytes from SRC to SRC+LEN-1, inclusively, is checked. 56 * If SRC is a valid UTF-8, the return value will point to the byte SRC+LEN, 57 * otherwise it will point to the start of the first invalid character. 58 * In either case all the characters between SRC and the return pointer - 1, 59 * inclusively, are valid UTF-8. 60 * 61 * See also svn_utf__is_valid(). 62 */ 63const char * 64svn_utf__last_valid(const char *src, apr_size_t len); 65 66/* As for svn_utf__last_valid but uses a different implementation without 67 lookup tables. It avoids the table memory use (about 400 bytes) but the 68 function is longer (about 200 bytes extra) and likely to be slower when 69 the string is valid. If the string is invalid this function may be 70 faster since it returns immediately rather than continuing to the end of 71 the string. The main reason this function exists is to test the table 72 driven implementation. */ 73const char * 74svn_utf__last_valid2(const char *src, apr_size_t len); 75 76/* Copy LENGTH bytes of SRC, converting characters as follows: 77 - Pass characters from the ASCII subset to the result 78 - Strip all combining marks from the string 79 - Represent other valid Unicode chars as {U+XXXX} 80 - Replace invalid Unicode chars with {U?XXXX} 81 - Represent chars that are not valid UTF-8 as ?\XX 82 - Replace codes outside the Unicode range with a sequence of ?\XX 83 - Represent the null byte as \0 84 Allocate the result in POOL. */ 85const char * 86svn_utf__fuzzy_escape(const char *src, apr_size_t length, apr_pool_t *pool); 87 88const char * 89svn_utf__cstring_from_utf8_fuzzy(const char *src, 90 apr_pool_t *pool, 91 svn_error_t *(*convert_from_utf8) 92 (const char **, 93 const char *, 94 apr_pool_t *)); 95 96 97#if defined(WIN32) 98/* On Windows: Convert the UTF-8 string SRC to UTF-16. 99 If PREFIX is not NULL, prepend it to the converted result. 100 The result, if not empty, will be allocated in RESULT_POOL. */ 101svn_error_t * 102svn_utf__win32_utf8_to_utf16(const WCHAR **result, 103 const char *src, 104 const WCHAR *prefix, 105 apr_pool_t *result_pool); 106 107/* On Windows: Convert the UTF-16 string SRC to UTF-8. 108 If PREFIX is not NULL, prepend it to the converted result. 109 The result, if not empty, will be allocated in RESULT_POOL. */ 110svn_error_t * 111svn_utf__win32_utf16_to_utf8(const char **result, 112 const WCHAR *src, 113 const char *prefix, 114 apr_pool_t *result_pool); 115#endif /* WIN32*/ 116 117 118/* A constant used for many length parameters in the utf8proc wrappers 119 * to indicate that the length of a string is unknonw. */ 120#define SVN_UTF__UNKNOWN_LENGTH ((apr_size_t) -1) 121 122 123/* Compare two UTF-8 strings, ignoring normalization, using buffers 124 * BUF1 and BUF2 for temporary storage. If either of LEN1 or LEN2 is 125 * SVN_UTF__UNKNOWN_LENGTH, assume the associated string is 126 * null-terminated; otherwise, consider the string only up to the 127 * given length. 128 * 129 * Return compare value in *RESULT. 130 */ 131svn_error_t * 132svn_utf__normcmp(int *result, 133 const char *str1, apr_size_t len1, 134 const char *str2, apr_size_t len2, 135 svn_membuf_t *buf1, svn_membuf_t *buf2); 136 137/* Normalize the UTF-8 string STR to form C, using BUF for temporary 138 * storage. If LEN is SVN_UTF__UNKNOWN_LENGTH, assume STR is 139 * null-terminated; otherwise, consider the string only up to the 140 * given length. 141 * 142 * Return the normalized string in *RESULT, which shares storage with 143 * BUF and is valid only until the next time BUF is modified. 144 * 145 * A returned error may indicate that STRING contains invalid UTF-8 or 146 * invalid Unicode codepoints. 147 */ 148svn_error_t* 149svn_utf__normalize(const char **result, 150 const char *str, apr_size_t len, 151 svn_membuf_t *buf); 152 153/* Check if STRING is a valid, NFC-normalized UTF-8 string. Note that 154 * a FALSE return value may indicate that STRING is not valid UTF-8 at 155 * all. 156 * 157 * Use SCRATCH_POOL for temporary allocations. 158 */ 159svn_boolean_t 160svn_utf__is_normalized(const char *string, apr_pool_t *scratch_pool); 161 162/* Encode an UCS-4 string to UTF-8, placing the result into BUFFER. 163 * While utf8proc does have a similar function, it does more checking 164 * and processing than we want here; this function does not attempt 165 * any normalizations but just encodes the individual code points. 166 * The encoded string will always be NUL-terminated. 167 * 168 * Return the length of the result (excluding the NUL terminator) in 169 * *result_length. 170 * 171 * A returned error indicates that a codepoint is invalid. 172 */ 173svn_error_t * 174svn_utf__encode_ucs4_string(svn_membuf_t *buffer, 175 const apr_int32_t *ucs4str, 176 apr_size_t length, 177 apr_size_t *result_length); 178 179/* Pattern matching similar to the the SQLite LIKE and GLOB 180 * operators. PATTERN, KEY and ESCAPE must all point to UTF-8 181 * strings. Furthermore, ESCAPE, if provided, must be a character from 182 * the ASCII subset. 183 * 184 * If any of PATTERN_LEN, STRING_LEN or ESCAPE_LEN are 185 * SVN_UTF__UNKNOWN_LENGTH, assume the associated string is 186 * null-terminated; otherwise, consider the string only up to the 187 * given length. 188 * 189 * Use buffers PATTERN_BUF, STRING_BUF and TEMP_BUF for temporary storage. 190 * 191 * If SQL_LIKE is true, interpret PATTERN as a pattern used by the SQL 192 * LIKE operator and notice ESCAPE. Otherwise it's a Unix fileglob 193 * pattern, and ESCAPE must be NULL. 194 * 195 * Set *MATCH to the result of the comparison. 196*/ 197svn_error_t * 198svn_utf__glob(svn_boolean_t *match, 199 const char *pattern, apr_size_t pattern_len, 200 const char *string, apr_size_t string_len, 201 const char *escape, apr_size_t escape_len, 202 svn_boolean_t sql_like, 203 svn_membuf_t *pattern_buf, 204 svn_membuf_t *string_buf, 205 svn_membuf_t *temp_buf); 206 207/* Return the compiled version of the wrapped utf8proc library. */ 208const char * 209svn_utf__utf8proc_compiled_version(void); 210 211/* Return the runtime version of the wrapped utf8proc library. */ 212const char * 213svn_utf__utf8proc_runtime_version(void); 214 215/* Convert an UTF-16 (or UCS-2) string to UTF-8, returning the pointer 216 * in RESULT. If BIG_ENDIAN is set, then UTF16STR is big-endian; 217 * otherwise, it's little-endian. 218 * 219 * If UTF16LEN is SVN_UTF__UNKNOWN_LENGTH, then UTF16STR must be 220 * terminated with a zero; otherwise, it is the number of 16-bit codes 221 * to convert, and the source string may contain NUL values. 222 * 223 * Allocate RESULT in RESULT_POOL and use SCRATCH_POOL for 224 * intermediate allocation. 225 * 226 * This function combines UTF-16 surrogate pairs into single code 227 * points, but will leave single lead or trail surrogates unchanged. 228 */ 229svn_error_t * 230svn_utf__utf16_to_utf8(const svn_string_t **result, 231 const apr_uint16_t *utf16str, 232 apr_size_t utf16len, 233 svn_boolean_t big_endian, 234 apr_pool_t *result_pool, 235 apr_pool_t *scratch_pool); 236 237/* Convert an UTF-32 string to UTF-8, returning the pointer in 238 * RESULT. If BIG_ENDIAN is set, then UTF32STR is big-endian; 239 * otherwise, it's little-endian. 240 * 241 * If UTF32LEN is SVN_UTF__UNKNOWN_LENGTH, then UTF32STR must be 242 * terminated with a zero; otherwise, it is the number of 32-bit codes 243 * to convert, and the source string may contain NUL values. 244 * 245 * Allocate RESULT in RESULT_POOL and use SCRATCH_POOL for 246 * intermediate allocation. 247 */ 248svn_error_t * 249svn_utf__utf32_to_utf8(const svn_string_t **result, 250 const apr_int32_t *utf32str, 251 apr_size_t utf32len, 252 svn_boolean_t big_endian, 253 apr_pool_t *result_pool, 254 apr_pool_t *scratch_pool); 255 256 257#ifdef __cplusplus 258} 259#endif /* __cplusplus */ 260 261#endif /* SVN_UTF_PRIVATE_H */ 262