URL.subproj/CFURL.h

   1 /*
   2  * Copyright (c) 2005 Apple Computer, Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_HEADER_START@
   5  *
   6  * This file contains Original Code and/or Modifications of Original Code
   7  * as defined in and that are subject to the Apple Public Source License
   8  * Version 2.0 (the 'License'). You may not use this file except in
   9  * compliance with the License. Please obtain a copy of the License at
  10  * http://www.opensource.apple.com/apsl/ and read it before using this
  11  * file.
  12  *
  13  * The Original Code and all software distributed under the License are
  14  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  15  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  16  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  17  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  18  * Please see the License for the specific language governing rights and
  19  * limitations under the License.
  20  *
  21  * @APPLE_LICENSE_HEADER_END@
  22  */
  23 /*      CFURL.h
  24         Copyright (c) 1998-2005, Apple, Inc. All rights reserved.
  25 */
  26
  27 #if !defined(__COREFOUNDATION_CFURL__)
  28 #define __COREFOUNDATION_CFURL__ 1
  29
  30 #include <CoreFoundation/CFBase.h>
  31 #include <CoreFoundation/CFData.h>
  32 #include <CoreFoundation/CFString.h>
  33
  34 #if defined(__cplusplus)
  35 extern "C" {
  36 #endif
  37
  38 typedef enum {
  39     kCFURLPOSIXPathStyle = 0,
  40     kCFURLHFSPathStyle,
  41     kCFURLWindowsPathStyle
  42 } CFURLPathStyle;
  43
  44 typedef const struct __CFURL * CFURLRef;
  45
  46 /* CFURLs are composed of two fundamental pieces - their string, and a */
  47 /* (possibly NULL) base URL.  A relative URL is one in which the string */
  48 /* by itself does not fully specify the URL (for instance "myDir/image.tiff"); */
  49 /* an absolute URL is one in which the string does fully specify the URL */
  50 /* ("file://localhost/myDir/image.tiff").  Absolute URLs always have NULL */
  51 /* base URLs; however, it is possible for a URL to have a NULL base, and still */
  52 /* not be absolute.  Such a URL has only a relative string, and cannot be */
  53 /* resolved.  Two CFURLs are considered equal if and only if their strings */
  54 /* are equal and their bases are equal.  In other words, */
  55 /* "file://localhost/myDir/image.tiff" is NOT equal to the URL with relative */
  56 /* string "myDir/image.tiff" and base URL "file://localhost/".  Clients that */
  57 /* need these less strict form of equality should convert all URLs to their */
  58 /* absolute form via CFURLCopyAbsoluteURL(), then compare the absolute forms. */
  59
  60 CF_EXPORT
  61 CFTypeID CFURLGetTypeID(void);
  62
  63 /* encoding will be used both to interpret the bytes of URLBytes, and to */
  64 /* interpret any percent-escapes within the bytes. */
  65 CF_EXPORT
  66 CFURLRef CFURLCreateWithBytes(CFAllocatorRef allocator, const UInt8 *URLBytes, CFIndex length, CFStringEncoding encoding, CFURLRef baseURL);
  67
  68 /* Escapes any character that is not 7-bit ASCII with the byte-code */
  69 /* for the given encoding.  If escapeWhitespace is true, whitespace */
  70 /* characters (' ', '\t', '\r', '\n') will be escaped also (desirable */
  71 /* if embedding the URL into a larger text stream like HTML) */
  72 CF_EXPORT
  73 CFDataRef CFURLCreateData(CFAllocatorRef allocator, CFURLRef url, CFStringEncoding encoding, Boolean escapeWhitespace);
  74
  75 /* Any escape sequences in URLString will be interpreted via UTF-8. */
  76 CF_EXPORT
  77 CFURLRef CFURLCreateWithString(CFAllocatorRef allocator, CFStringRef URLString, CFURLRef baseURL);
  78
  79 #if MAC_OS_X_VERSION_10_3 <= MAC_OS_X_VERSION_MAX_ALLOWED
  80
  81 /* Create an absolute URL directly, without requiring the extra step */
  82 /* of calling CFURLCopyAbsoluteURL().  If useCompatibilityMode is  */
  83 /* true, the rules historically used on the web are used to resolve */
  84 /* relativeString against baseURL - these rules are generally listed */
  85 /* in the RFC as optional or alternate interpretations.  Otherwise, */
  86 /* the strict rules from the RFC are used.  The major differences are */
  87 /* that in compatibility mode, we are lenient of the scheme appearing */
  88 /* in relative portion, leading "../" components are removed from the */
  89 /* final URL's path, and if the relative portion contains only */
  90 /* resource specifier pieces (query, parameters, and fragment), then */
  91 /* the last path component of the base URL will not be deleted  */
  92 CF_EXPORT
  93 CFURLRef CFURLCreateAbsoluteURLWithBytes(CFAllocatorRef alloc, const UInt8 *relativeURLBytes, CFIndex length, CFStringEncoding encoding, CFURLRef baseURL, Boolean useCompatibilityMode) AVAILABLE_MAC_OS_X_VERSION_10_3_AND_LATER;
  94 #endif
  95
  96 /* filePath should be the URL's path expressed as a path of the type */
  97 /* fsType.  If filePath is not absolute, the resulting URL will be */
  98 /* considered relative to the current working directory (evaluated */
  99 /* at creation time).  isDirectory determines whether filePath is */
 100 /* treated as a directory path when resolving against relative path */
 101 /* components */
 102 CF_EXPORT
 103 CFURLRef CFURLCreateWithFileSystemPath(CFAllocatorRef allocator, CFStringRef filePath, CFURLPathStyle pathStyle, Boolean isDirectory);
 104
 105 CF_EXPORT
 106 CFURLRef CFURLCreateFromFileSystemRepresentation(CFAllocatorRef allocator, const UInt8 *buffer, CFIndex bufLen, Boolean isDirectory);
 107
 108 CF_EXPORT
 109 CFURLRef CFURLCreateWithFileSystemPathRelativeToBase(CFAllocatorRef allocator, CFStringRef filePath, CFURLPathStyle pathStyle, Boolean isDirectory, CFURLRef baseURL);
 110
 111 CF_EXPORT
 112 CFURLRef CFURLCreateFromFileSystemRepresentationRelativeToBase(CFAllocatorRef allocator, const UInt8 *buffer, CFIndex bufLen, Boolean isDirectory, CFURLRef baseURL);
 113
 114 /* Fills buffer with the file system's native representation of */
 115 /* url's path. No more than maxBufLen bytes are written to buffer. */
 116 /* The buffer should be at least the maximum path length for */
 117 /* the file system in question to avoid failures for insufficiently */
 118 /* large buffers.  If resolveAgainstBase is true, the url's relative */
 119 /* portion is resolved against its base before the path is computed. */
 120 /* Returns success or failure. */
 121 CF_EXPORT
 122 Boolean CFURLGetFileSystemRepresentation(CFURLRef url, Boolean resolveAgainstBase, UInt8 *buffer, CFIndex maxBufLen);
 123
 124 /* Creates a new URL by resolving the relative portion of relativeURL against its base. */
 125 CF_EXPORT
 126 CFURLRef CFURLCopyAbsoluteURL(CFURLRef relativeURL);
 127
 128 /* Returns the URL's string. */
 129 CF_EXPORT
 130 CFStringRef CFURLGetString(CFURLRef anURL);
 131
 132 /* Returns the base URL if it exists */
 133 CF_EXPORT
 134 CFURLRef CFURLGetBaseURL(CFURLRef anURL);
 135
 136 /*
 137 All URLs can be broken into two pieces - the scheme (preceding the
 138 first colon) and the resource specifier (following the first colon).
 139 Most URLs are also "standard" URLs conforming to RFC 1808 (available
 140 from www.w3c.org).  This category includes URLs of the file, http,
 141 https, and ftp schemes, to name a few.  Standard URLs start the
 142 resource specifier with two slashes ("//"), and can be broken into
 143 four distinct pieces - the scheme, the net location, the path, and
 144 further resource specifiers (typically an optional parameter, query,
 145 and/or fragment).  The net location appears immediately following
 146 the two slashes and goes up to the next slash; it's format is
 147 scheme-specific, but is usually composed of some or all of a username,
 148 password, host name, and port.  The path is a series of path components
 149 separated by slashes; if the net location is present, the path always
 150 begins with a slash.  Standard URLs can be relative to another URL,
 151 in which case at least the scheme and possibly other pieces as well
 152 come from the base URL (see RFC 1808 for precise details when resolving
 153 a relative URL against its base).  The full URL is therefore
 154
 155 <scheme> "://" <net location> <path, always starting with slash> <add'l resource specifiers>
 156
 157 If a given CFURL can be decomposed (that is, conforms to RFC 1808), you
 158 can ask for each of the four basic pieces (scheme, net location, path,
 159 and resource specifer) separately, as well as for its base URL.  The
 160 basic pieces are returned with any percent escape sequences still in
 161 place (although note that the scheme may not legally include any
 162 percent escapes); this is to allow the caller to distinguish between
 163 percent sequences that may have syntactic meaning if replaced by the
 164 character being escaped (for instance, a '/' in a path component).
 165 Since only the individual schemes know which characters are
 166 syntactically significant, CFURL cannot safely replace any percent
 167 escape sequences.  However, you can use
 168 CFURLCreateStringByReplacingPercentEscapes() to create a new string with
 169 the percent escapes removed; see below.
 170
 171 If a given CFURL can not be decomposed, you can ask for its scheme and its
 172 resource specifier; asking it for its net location or path will return NULL.
 173
 174 To get more refined information about the components of a decomposable
 175 CFURL, you may ask for more specific pieces of the URL, expressed with
 176 the percent escapes removed.  The available functions are CFURLCopyHostName(),
 177 CFURLGetPortNumber() (returns an Int32), CFURLCopyUserName(),
 178 CFURLCopyPassword(), CFURLCopyQuery(), CFURLCopyParameters(), and
 179 CFURLCopyFragment().  Because the parameters, query, and fragment of an
 180 URL may contain scheme-specific syntaxes, these methods take a second
 181 argument, giving a list of characters which should NOT be replaced if
 182 percent escaped.  For instance, the ftp parameter syntax gives simple
 183 key-value pairs as "<key>=<value>;"  Clearly if a key or value includes
 184 either '=' or ';', it must be escaped to avoid corrupting the meaning of
 185 the parameters, so the caller may request the parameter string as
 186
 187 CFStringRef myParams = CFURLCopyParameters(ftpURL, CFSTR("=;%"));
 188
 189 requesting that all percent escape sequences be replaced by the represented
 190 characters, except for escaped '=', '%' or ';' characters.  Pass the empty
 191 string (CFSTR("")) to request that all percent escapes be replaced, or NULL
 192 to request that none be.
 193 */
 194
 195 /* Returns true if anURL conforms to RFC 1808 */
 196 CF_EXPORT
 197 Boolean CFURLCanBeDecomposed(CFURLRef anURL);
 198
 199 /* The next several methods leave any percent escape sequences intact */
 200
 201 CF_EXPORT
 202 CFStringRef CFURLCopyScheme(CFURLRef anURL);
 203
 204 /* NULL if CFURLCanBeDecomposed(anURL) is false */
 205 CF_EXPORT
 206 CFStringRef CFURLCopyNetLocation(CFURLRef anURL);
 207
 208 /* NULL if CFURLCanBeDecomposed(anURL) is false; also does not resolve the URL */
 209 /* against its base.  See also CFURLCopyAbsoluteURL().  Note that, strictly */
 210 /* speaking, any leading '/' is not considered part of the URL's path, although */
 211 /* its presence or absence determines whether the path is absolute. */
 212 /* CFURLCopyPath()'s return value includes any leading slash (giving the path */
 213 /* the normal POSIX appearance); CFURLCopyStrictPath()'s return value omits any */
 214 /* leading slash, and uses isAbsolute to report whether the URL's path is absolute. */
 215
 216 /* CFURLCopyFileSystemPath() returns the URL's path as a file system path for the */
 217 /* given path style.  All percent escape sequences are replaced.  The URL is not */
 218 /* resolved against its base before computing the path. */
 219 CF_EXPORT
 220 CFStringRef CFURLCopyPath(CFURLRef anURL);
 221
 222 CF_EXPORT
 223 CFStringRef CFURLCopyStrictPath(CFURLRef anURL, Boolean *isAbsolute);
 224
 225 CF_EXPORT
 226 CFStringRef CFURLCopyFileSystemPath(CFURLRef anURL, CFURLPathStyle pathStyle);
 227
 228 /* Returns whether anURL's path represents a directory */
 229 /* (true returned) or a simple file (false returned) */
 230 CF_EXPORT
 231 Boolean CFURLHasDirectoryPath(CFURLRef anURL);
 232
 233 /* Any additional resource specifiers after the path.  For URLs */
 234 /* that cannot be decomposed, this is everything except the scheme itself. */
 235 CF_EXPORT
 236 CFStringRef CFURLCopyResourceSpecifier(CFURLRef anURL);
 237
 238 CF_EXPORT
 239 CFStringRef CFURLCopyHostName(CFURLRef anURL);
 240
 241 CF_EXPORT
 242 SInt32 CFURLGetPortNumber(CFURLRef anURL); /* Returns -1 if no port number is specified */
 243
 244 CF_EXPORT
 245 CFStringRef CFURLCopyUserName(CFURLRef anURL);
 246
 247 CF_EXPORT
 248 CFStringRef CFURLCopyPassword(CFURLRef anURL);
 249
 250 /* These remove all percent escape sequences except those for */
 251 /* characters in charactersToLeaveEscaped.  If charactersToLeaveEscaped */
 252 /* is empty (""), all percent escape sequences are replaced by their */
 253 /* corresponding characters.  If charactersToLeaveEscaped is NULL, */
 254 /* then no escape sequences are removed at all */
 255 CF_EXPORT
 256 CFStringRef CFURLCopyParameterString(CFURLRef anURL, CFStringRef charactersToLeaveEscaped);
 257
 258 CF_EXPORT
 259 CFStringRef CFURLCopyQueryString(CFURLRef anURL, CFStringRef charactersToLeaveEscaped);
 260
 261 CF_EXPORT
 262 CFStringRef CFURLCopyFragment(CFURLRef anURL, CFStringRef charactersToLeaveEscaped);
 263
 264 CF_EXPORT
 265 CFStringRef CFURLCopyLastPathComponent(CFURLRef url);
 266
 267 CF_EXPORT
 268 CFStringRef CFURLCopyPathExtension(CFURLRef url);
 269
 270 /* These functions all treat the base URL of the supplied url as */
 271 /* invariant.  In other words, the URL returned will always have */
 272 /* the same base as the URL supplied as an argument. */
 273
 274 CF_EXPORT
 275 CFURLRef CFURLCreateCopyAppendingPathComponent(CFAllocatorRef allocator, CFURLRef url, CFStringRef pathComponent, Boolean isDirectory);
 276
 277 CF_EXPORT
 278 CFURLRef CFURLCreateCopyDeletingLastPathComponent(CFAllocatorRef allocator, CFURLRef url);
 279
 280 CF_EXPORT
 281 CFURLRef CFURLCreateCopyAppendingPathExtension(CFAllocatorRef allocator, CFURLRef url, CFStringRef extension);
 282
 283 CF_EXPORT
 284 CFURLRef CFURLCreateCopyDeletingPathExtension(CFAllocatorRef allocator, CFURLRef url);
 285
 286 #if MAC_OS_X_VERSION_10_3 <= MAC_OS_X_VERSION_MAX_ALLOWED
 287 /* Fills buffer with the bytes for url, returning the number of bytes */
 288 /* filled.  If buffer is of insufficient size, returns -1 and no bytes */
 289 /* are placed in buffer.  If buffer is NULL, the needed length is */
 290 /* computed and returned.  The returned bytes are the original bytes */
 291 /* from which the URL was created; if the URL was created from a */
 292 /* string, the bytes will be the bytes of the string encoded via UTF-8  */
 293 CF_EXPORT
 294 CFIndex CFURLGetBytes(CFURLRef url, UInt8 *buffer, CFIndex bufferLength) AVAILABLE_MAC_OS_X_VERSION_10_3_AND_LATER;
 295
 296 typedef enum {
 297         kCFURLComponentScheme = 1,
 298         kCFURLComponentNetLocation = 2,
 299         kCFURLComponentPath = 3,
 300         kCFURLComponentResourceSpecifier = 4,
 301
 302         kCFURLComponentUser = 5,
 303         kCFURLComponentPassword = 6,
 304         kCFURLComponentUserInfo = 7,
 305         kCFURLComponentHost = 8,
 306         kCFURLComponentPort = 9,
 307         kCFURLComponentParameterString = 10,
 308         kCFURLComponentQuery = 11,
 309         kCFURLComponentFragment = 12
 310 } CFURLComponentType;
 311
 312 /*
 313 Gets the  range of the requested component in the bytes of url, as
 314 returned by CFURLGetBytes().  This range is only good for use in the
 315 bytes returned by CFURLGetBytes!
 316
 317 If non-NULL, rangeIncludingSeparators gives the range of component
 318 including the sequences that separate component from the previous and
 319 next components.  If there is no previous or next component, that end of
 320 rangeIncludingSeparators will match the range of the component itself.
 321 If url does not contain the given component type, (kCFNotFound, 0) is
 322 returned, and rangeIncludingSeparators is set to the location where the
 323 component would be inserted.  Some examples -
 324
 325 For the URL http://www.apple.com/hotnews/
 326
 327 Component           returned range      rangeIncludingSeparators
 328 scheme              (0, 4)              (0, 7)
 329 net location        (7, 13)             (4, 16)
 330 path                (20, 9)             (20, 9)
 331 resource specifier  (kCFNotFound, 0)    (29, 0)
 332 user                (kCFNotFound, 0)    (7, 0)
 333 password            (kCFNotFound, 0)    (7, 0)
 334 user info           (kCFNotFound, 0)    (7, 0)
 335 host                (7, 13)             (4, 16)
 336 port                (kCFNotFound, 0)    (20, 0)
 337 parameter           (kCFNotFound, 0)    (29, 0)
 338 query               (kCFNotFound, 0)    (29, 0)
 339 fragment            (kCFNotFound, 0)    (29, 0)
 340
 341
 342 For the URL ./relPath/file.html#fragment
 343
 344 Component           returned range      rangeIncludingSeparators
 345 scheme              (kCFNotFound, 0)    (0, 0)
 346 net location        (kCFNotFound, 0)    (0, 0)
 347 path                (0, 19)             (0, 20)
 348 resource specifier  (20, 8)             (19, 9)
 349 user                (kCFNotFound, 0)    (0, 0)
 350 password            (kCFNotFound, 0)    (0, 0)
 351 user info           (kCFNotFound, 0)    (0, 0)
 352 host                (kCFNotFound, 0)    (0, 0)
 353 port                (kCFNotFound, 0)    (0, 0)
 354 parameter           (kCFNotFound, 0)    (19, 0)
 355 query               (kCFNotFound, 0)    (19, 0)
 356 fragment            (20, 8)             (19, 9)
 357
 358
 359 For the URL scheme://user:pass@host:1/path/path2/file.html;params?query#fragment
 360
 361 Component           returned range      rangeIncludingSeparators
 362 scheme              (0, 6)              (0, 9)
 363 net location        (9, 16)             (6, 19)
 364 path                (25, 21)            (25, 22)
 365 resource specifier  (47, 21)            (46, 22)
 366 user                (9, 4)              (6, 8)
 367 password            (14, 4)             (13, 6)
 368 user info           (9, 9)              (6, 13)
 369 host                (19, 4)             (18, 6)
 370 port                (24, 1)             (23, 2)
 371 parameter           (47, 6)             (46, 8)
 372 query               (54, 5)             (53, 7)
 373 fragment            (60, 8)             (59, 9)
 374 */
 375 CF_EXPORT
 376 CFRange CFURLGetByteRangeForComponent(CFURLRef url, CFURLComponentType component, CFRange *rangeIncludingSeparators) AVAILABLE_MAC_OS_X_VERSION_10_3_AND_LATER;
 377 #endif
 378
 379 /* Returns a string with any percent escape sequences that do NOT */
 380 /* correspond to characters in charactersToLeaveEscaped with their */
 381 /* equivalent.  Returns NULL on failure (if an invalid percent sequence */
 382 /* is encountered), or the original string (retained) if no characters */
 383 /* need to be replaced. Pass NULL to request that no percent escapes be */
 384 /* replaced, or the empty string (CFSTR("")) to request that all percent */
 385 /* escapes be replaced.  Uses UTF8 to interpret percent escapes. */
 386 CF_EXPORT
 387 CFStringRef CFURLCreateStringByReplacingPercentEscapes(CFAllocatorRef allocator, CFStringRef originalString, CFStringRef charactersToLeaveEscaped);
 388
 389 #if MAC_OS_X_VERSION_10_3 <= MAC_OS_X_VERSION_MAX_ALLOWED
 390 /* As above, but allows you to specify the encoding to use when interpreting percent escapes */
 391 CF_EXPORT
 392 CFStringRef CFURLCreateStringByReplacingPercentEscapesUsingEncoding(CFAllocatorRef allocator, CFStringRef origString, CFStringRef charsToLeaveEscaped, CFStringEncoding encoding) AVAILABLE_MAC_OS_X_VERSION_10_3_AND_LATER;
 393 #endif
 394
 395 /* Creates a copy or originalString, replacing certain characters with */
 396 /* the equivalent percent escape sequence based on the encoding specified. */
 397 /* If the originalString does not need to be modified (no percent escape */
 398 /* sequences are missing), may retain and return originalString. */
 399 /* If you are uncertain of the correct encoding, you should use UTF-8, */
 400 /* which is the encoding designated by RFC 2396 as the correct encoding */
 401 /* for use in URLs.  The characters so escaped are all characters that */
 402 /* are not legal URL characters (based on RFC 2396), plus any characters */
 403 /* in legalURLCharactersToBeEscaped, less any characters in */
 404 /* charactersToLeaveUnescaped.  To simply correct any non-URL characters */
 405 /* in an otherwise correct URL string, do: */
 406
 407 /* newString = CFURLCreateStringByAddingPercentEscapes(NULL, origString, NULL, NULL, kCFStringEncodingUTF8); */
 408 CF_EXPORT
 409 CFStringRef CFURLCreateStringByAddingPercentEscapes(CFAllocatorRef allocator, CFStringRef originalString, CFStringRef charactersToLeaveUnescaped, CFStringRef legalURLCharactersToBeEscaped, CFStringEncoding encoding);
 410
 411
 412 #if defined(__cplusplus)
 413 }
 414 #endif
 415
 416 #endif /* !__COREFOUNDATION_CFURL__ */
 417