1// Copyright (c) 2012 The Chromium Authors. All rights reserved. 2// Use of this source code is governed by a BSD-style license that can be 3// found in the LICENSE file. 4 5// FilePath is a container for pathnames stored in a platform's native string 6// type, providing containers for manipulation in according with the 7// platform's conventions for pathnames. It supports the following path 8// types: 9// 10// POSIX Windows 11// --------------- ---------------------------------- 12// Fundamental type char[] wchar_t[] 13// Encoding unspecified* UTF-16 14// Separator / \, tolerant of / 15// Drive letters no case-insensitive A-Z followed by : 16// Alternate root // (surprise!) \\, for UNC paths 17// 18// * The encoding need not be specified on POSIX systems, although some 19// POSIX-compliant systems do specify an encoding. Mac OS X uses UTF-8. 20// Chrome OS also uses UTF-8. 21// Linux does not specify an encoding, but in practice, the locale's 22// character set may be used. 23// 24// For more arcane bits of path trivia, see below. 25// 26// FilePath objects are intended to be used anywhere paths are. An 27// application may pass FilePath objects around internally, masking the 28// underlying differences between systems, only differing in implementation 29// where interfacing directly with the system. For example, a single 30// OpenFile(const FilePath &) function may be made available, allowing all 31// callers to operate without regard to the underlying implementation. On 32// POSIX-like platforms, OpenFile might wrap fopen, and on Windows, it might 33// wrap _wfopen_s, perhaps both by calling file_path.value().c_str(). This 34// allows each platform to pass pathnames around without requiring conversions 35// between encodings, which has an impact on performance, but more imporantly, 36// has an impact on correctness on platforms that do not have well-defined 37// encodings for pathnames. 38// 39// Several methods are available to perform common operations on a FilePath 40// object, such as determining the parent directory (DirName), isolating the 41// final path component (BaseName), and appending a relative pathname string 42// to an existing FilePath object (Append). These methods are highly 43// recommended over attempting to split and concatenate strings directly. 44// These methods are based purely on string manipulation and knowledge of 45// platform-specific pathname conventions, and do not consult the filesystem 46// at all, making them safe to use without fear of blocking on I/O operations. 47// These methods do not function as mutators but instead return distinct 48// instances of FilePath objects, and are therefore safe to use on const 49// objects. The objects themselves are safe to share between threads. 50// 51// To aid in initialization of FilePath objects from string literals, a 52// FILE_PATH_LITERAL macro is provided, which accounts for the difference 53// between char[]-based pathnames on POSIX systems and wchar_t[]-based 54// pathnames on Windows. 55// 56// Paths can't contain NULs as a precaution agaist premature truncation. 57// 58// Because a FilePath object should not be instantiated at the global scope, 59// instead, use a FilePath::CharType[] and initialize it with 60// FILE_PATH_LITERAL. At runtime, a FilePath object can be created from the 61// character array. Example: 62// 63// | const FilePath::CharType kLogFileName[] = FILE_PATH_LITERAL("log.txt"); 64// | 65// | void Function() { 66// | FilePath log_file_path(kLogFileName); 67// | [...] 68// | } 69// 70// WARNING: FilePaths should ALWAYS be displayed with LTR directionality, even 71// when the UI language is RTL. This means you always need to pass filepaths 72// through base::i18n::WrapPathWithLTRFormatting() before displaying it in the 73// RTL UI. 74// 75// This is a very common source of bugs, please try to keep this in mind. 76// 77// ARCANE BITS OF PATH TRIVIA 78// 79// - A double leading slash is actually part of the POSIX standard. Systems 80// are allowed to treat // as an alternate root, as Windows does for UNC 81// (network share) paths. Most POSIX systems don't do anything special 82// with two leading slashes, but FilePath handles this case properly 83// in case it ever comes across such a system. FilePath needs this support 84// for Windows UNC paths, anyway. 85// References: 86// The Open Group Base Specifications Issue 7, sections 3.266 ("Pathname") 87// and 4.12 ("Pathname Resolution"), available at: 88// http://www.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_266 89// http://www.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap04.html#tag_04_12 90// 91// - Windows treats c:\\ the same way it treats \\. This was intended to 92// allow older applications that require drive letters to support UNC paths 93// like \\server\share\path, by permitting c:\\server\share\path as an 94// equivalent. Since the OS treats these paths specially, FilePath needs 95// to do the same. Since Windows can use either / or \ as the separator, 96// FilePath treats c://, c:\\, //, and \\ all equivalently. 97// Reference: 98// The Old New Thing, "Why is a drive letter permitted in front of UNC 99// paths (sometimes)?", available at: 100// http://blogs.msdn.com/oldnewthing/archive/2005/11/22/495740.aspx 101 102#ifndef BASE_FILES_FILE_PATH_H_ 103#define BASE_FILES_FILE_PATH_H_ 104 105#include <stddef.h> 106#include <string> 107#include <vector> 108 109#include "base/base_export.h" 110#include "base/compiler_specific.h" 111#include "base/containers/hash_tables.h" 112#include "base/strings/string16.h" 113#include "base/strings/string_piece.h" // For implicit conversions. 114#include "build/build_config.h" 115 116// Windows-style drive letter support and pathname separator characters can be 117// enabled and disabled independently, to aid testing. These #defines are 118// here so that the same setting can be used in both the implementation and 119// in the unit test. 120#if defined(OS_WIN) 121#define FILE_PATH_USES_DRIVE_LETTERS 122#define FILE_PATH_USES_WIN_SEPARATORS 123#endif // OS_WIN 124 125class Pickle; 126class PickleIterator; 127 128namespace base { 129 130// An abstraction to isolate users from the differences between native 131// pathnames on different platforms. 132class BASE_EXPORT FilePath { 133 public: 134#if defined(OS_POSIX) 135 // On most platforms, native pathnames are char arrays, and the encoding 136 // may or may not be specified. On Mac OS X, native pathnames are encoded 137 // in UTF-8. 138 typedef std::string StringType; 139#elif defined(OS_WIN) 140 // On Windows, for Unicode-aware applications, native pathnames are wchar_t 141 // arrays encoded in UTF-16. 142 typedef std::wstring StringType; 143#endif // OS_WIN 144 145 typedef StringType::value_type CharType; 146 147 // Null-terminated array of separators used to separate components in 148 // hierarchical paths. Each character in this array is a valid separator, 149 // but kSeparators[0] is treated as the canonical separator and will be used 150 // when composing pathnames. 151 static const CharType kSeparators[]; 152 153 // arraysize(kSeparators). 154 static const size_t kSeparatorsLength; 155 156 // A special path component meaning "this directory." 157 static const CharType kCurrentDirectory[]; 158 159 // A special path component meaning "the parent directory." 160 static const CharType kParentDirectory[]; 161 162 // The character used to identify a file extension. 163 static const CharType kExtensionSeparator; 164 165 FilePath(); 166 FilePath(const FilePath& that); 167 explicit FilePath(const StringType& path); 168 ~FilePath(); 169 FilePath& operator=(const FilePath& that); 170 171 bool operator==(const FilePath& that) const; 172 173 bool operator!=(const FilePath& that) const; 174 175 // Required for some STL containers and operations 176 bool operator<(const FilePath& that) const { 177 return path_ < that.path_; 178 } 179 180 const StringType& value() const { return path_; } 181 182 bool empty() const { return path_.empty(); } 183 184 void clear() { path_.clear(); } 185 186 // Returns true if |character| is in kSeparators. 187 static bool IsSeparator(CharType character); 188 189 // Returns a vector of all of the components of the provided path. It is 190 // equivalent to calling DirName().value() on the path's root component, 191 // and BaseName().value() on each child component. 192 void GetComponents(std::vector<FilePath::StringType>* components) const; 193 194 // Returns true if this FilePath is a strict parent of the |child|. Absolute 195 // and relative paths are accepted i.e. is /foo parent to /foo/bar and 196 // is foo parent to foo/bar. Does not convert paths to absolute, follow 197 // symlinks or directory navigation (e.g. ".."). A path is *NOT* its own 198 // parent. 199 bool IsParent(const FilePath& child) const; 200 201 // If IsParent(child) holds, appends to path (if non-NULL) the 202 // relative path to child and returns true. For example, if parent 203 // holds "/Users/johndoe/Library/Application Support", child holds 204 // "/Users/johndoe/Library/Application Support/Google/Chrome/Default", and 205 // *path holds "/Users/johndoe/Library/Caches", then after 206 // parent.AppendRelativePath(child, path) is called *path will hold 207 // "/Users/johndoe/Library/Caches/Google/Chrome/Default". Otherwise, 208 // returns false. 209 bool AppendRelativePath(const FilePath& child, FilePath* path) const; 210 211 // Returns a FilePath corresponding to the directory containing the path 212 // named by this object, stripping away the file component. If this object 213 // only contains one component, returns a FilePath identifying 214 // kCurrentDirectory. If this object already refers to the root directory, 215 // returns a FilePath identifying the root directory. 216 FilePath DirName() const WARN_UNUSED_RESULT; 217 218 // Returns a FilePath corresponding to the last path component of this 219 // object, either a file or a directory. If this object already refers to 220 // the root directory, returns a FilePath identifying the root directory; 221 // this is the only situation in which BaseName will return an absolute path. 222 FilePath BaseName() const WARN_UNUSED_RESULT; 223 224 // Returns ".jpg" for path "C:\pics\jojo.jpg", or an empty string if 225 // the file has no extension. If non-empty, Extension() will always start 226 // with precisely one ".". The following code should always work regardless 227 // of the value of path. 228 // new_path = path.RemoveExtension().value().append(path.Extension()); 229 // ASSERT(new_path == path.value()); 230 // NOTE: this is different from the original file_util implementation which 231 // returned the extension without a leading "." ("jpg" instead of ".jpg") 232 StringType Extension() const; 233 234 // Returns "C:\pics\jojo" for path "C:\pics\jojo.jpg" 235 // NOTE: this is slightly different from the similar file_util implementation 236 // which returned simply 'jojo'. 237 FilePath RemoveExtension() const WARN_UNUSED_RESULT; 238 239 // Inserts |suffix| after the file name portion of |path| but before the 240 // extension. Returns "" if BaseName() == "." or "..". 241 // Examples: 242 // path == "C:\pics\jojo.jpg" suffix == " (1)", returns "C:\pics\jojo (1).jpg" 243 // path == "jojo.jpg" suffix == " (1)", returns "jojo (1).jpg" 244 // path == "C:\pics\jojo" suffix == " (1)", returns "C:\pics\jojo (1)" 245 // path == "C:\pics.old\jojo" suffix == " (1)", returns "C:\pics.old\jojo (1)" 246 FilePath InsertBeforeExtension( 247 const StringType& suffix) const WARN_UNUSED_RESULT; 248 FilePath InsertBeforeExtensionASCII( 249 const base::StringPiece& suffix) const WARN_UNUSED_RESULT; 250 251 // Adds |extension| to |file_name|. Returns the current FilePath if 252 // |extension| is empty. Returns "" if BaseName() == "." or "..". 253 FilePath AddExtension( 254 const StringType& extension) const WARN_UNUSED_RESULT; 255 256 // Replaces the extension of |file_name| with |extension|. If |file_name| 257 // does not have an extension, then |extension| is added. If |extension| is 258 // empty, then the extension is removed from |file_name|. 259 // Returns "" if BaseName() == "." or "..". 260 FilePath ReplaceExtension( 261 const StringType& extension) const WARN_UNUSED_RESULT; 262 263 // Returns true if the file path matches the specified extension. The test is 264 // case insensitive. Don't forget the leading period if appropriate. 265 bool MatchesExtension(const StringType& extension) const; 266 267 // Returns a FilePath by appending a separator and the supplied path 268 // component to this object's path. Append takes care to avoid adding 269 // excessive separators if this object's path already ends with a separator. 270 // If this object's path is kCurrentDirectory, a new FilePath corresponding 271 // only to |component| is returned. |component| must be a relative path; 272 // it is an error to pass an absolute path. 273 FilePath Append(const StringType& component) const WARN_UNUSED_RESULT; 274 FilePath Append(const FilePath& component) const WARN_UNUSED_RESULT; 275 276 // Although Windows StringType is std::wstring, since the encoding it uses for 277 // paths is well defined, it can handle ASCII path components as well. 278 // Mac uses UTF8, and since ASCII is a subset of that, it works there as well. 279 // On Linux, although it can use any 8-bit encoding for paths, we assume that 280 // ASCII is a valid subset, regardless of the encoding, since many operating 281 // system paths will always be ASCII. 282 FilePath AppendASCII(const base::StringPiece& component) 283 const WARN_UNUSED_RESULT; 284 285 // Returns true if this FilePath contains an absolute path. On Windows, an 286 // absolute path begins with either a drive letter specification followed by 287 // a separator character, or with two separator characters. On POSIX 288 // platforms, an absolute path begins with a separator character. 289 bool IsAbsolute() const; 290 291 // Returns true if the patch ends with a path separator character. 292 bool EndsWithSeparator() const WARN_UNUSED_RESULT; 293 294 // Returns a copy of this FilePath that ends with a trailing separator. If 295 // the input path is empty, an empty FilePath will be returned. 296 FilePath AsEndingWithSeparator() const WARN_UNUSED_RESULT; 297 298 // Returns a copy of this FilePath that does not end with a trailing 299 // separator. 300 FilePath StripTrailingSeparators() const WARN_UNUSED_RESULT; 301 302 // Returns true if this FilePath contains any attempt to reference a parent 303 // directory (i.e. has a path component that is ".." 304 bool ReferencesParent() const; 305 306 // Return a Unicode human-readable version of this path. 307 // Warning: you can *not*, in general, go from a display name back to a real 308 // path. Only use this when displaying paths to users, not just when you 309 // want to stuff a string16 into some other API. 310 string16 LossyDisplayName() const; 311 312 // Return the path as ASCII, or the empty string if the path is not ASCII. 313 // This should only be used for cases where the FilePath is representing a 314 // known-ASCII filename. 315 std::string MaybeAsASCII() const; 316 317 // Return the path as UTF-8. 318 // 319 // This function is *unsafe* as there is no way to tell what encoding is 320 // used in file names on POSIX systems other than Mac and Chrome OS, 321 // although UTF-8 is practically used everywhere these days. To mitigate 322 // the encoding issue, this function internally calls 323 // SysNativeMBToWide() on POSIX systems other than Mac and Chrome OS, 324 // per assumption that the current locale's encoding is used in file 325 // names, but this isn't a perfect solution. 326 // 327 // Once it becomes safe to to stop caring about non-UTF-8 file names, 328 // the SysNativeMBToWide() hack will be removed from the code, along 329 // with "Unsafe" in the function name. 330 std::string AsUTF8Unsafe() const; 331 332 // Similar to AsUTF8Unsafe, but returns UTF-16 instead. 333 string16 AsUTF16Unsafe() const; 334 335 // Older Chromium code assumes that paths are always wstrings. 336 // This function converts wstrings to FilePaths, and is 337 // useful to smooth porting that old code to the FilePath API. 338 // It has "Hack" its name so people feel bad about using it. 339 // http://code.google.com/p/chromium/issues/detail?id=24672 340 // 341 // If you are trying to be a good citizen and remove these, ask yourself: 342 // - Am I interacting with other Chrome code that deals with files? Then 343 // try to convert the API into using FilePath. 344 // - Am I interacting with OS-native calls? Then use value() to get at an 345 // OS-native string format. 346 // - Am I using well-known file names, like "config.ini"? Then use the 347 // ASCII functions (we require paths to always be supersets of ASCII). 348 // - Am I displaying a string to the user in some UI? Then use the 349 // LossyDisplayName() function, but keep in mind that you can't 350 // ever use the result of that again as a path. 351 static FilePath FromWStringHack(const std::wstring& wstring); 352 353 // Returns a FilePath object from a path name in UTF-8. This function 354 // should only be used for cases where you are sure that the input 355 // string is UTF-8. 356 // 357 // Like AsUTF8Unsafe(), this function is unsafe. This function 358 // internally calls SysWideToNativeMB() on POSIX systems other than Mac 359 // and Chrome OS, to mitigate the encoding issue. See the comment at 360 // AsUTF8Unsafe() for details. 361 static FilePath FromUTF8Unsafe(const std::string& utf8); 362 363 // Similar to FromUTF8Unsafe, but accepts UTF-16 instead. 364 static FilePath FromUTF16Unsafe(const string16& utf16); 365 366 void WriteToPickle(Pickle* pickle) const; 367 bool ReadFromPickle(PickleIterator* iter); 368 369 // Normalize all path separators to backslash on Windows 370 // (if FILE_PATH_USES_WIN_SEPARATORS is true), or do nothing on POSIX systems. 371 FilePath NormalizePathSeparators() const; 372 373 // Compare two strings in the same way the file system does. 374 // Note that these always ignore case, even on file systems that are case- 375 // sensitive. If case-sensitive comparison is ever needed, add corresponding 376 // methods here. 377 // The methods are written as a static method so that they can also be used 378 // on parts of a file path, e.g., just the extension. 379 // CompareIgnoreCase() returns -1, 0 or 1 for less-than, equal-to and 380 // greater-than respectively. 381 static int CompareIgnoreCase(const StringType& string1, 382 const StringType& string2); 383 static bool CompareEqualIgnoreCase(const StringType& string1, 384 const StringType& string2) { 385 return CompareIgnoreCase(string1, string2) == 0; 386 } 387 static bool CompareLessIgnoreCase(const StringType& string1, 388 const StringType& string2) { 389 return CompareIgnoreCase(string1, string2) < 0; 390 } 391 392#if defined(OS_MACOSX) 393 // Returns the string in the special canonical decomposed form as defined for 394 // HFS, which is close to, but not quite, decomposition form D. See 395 // http://developer.apple.com/mac/library/technotes/tn/tn1150.html#UnicodeSubtleties 396 // for further comments. 397 // Returns the epmty string if the conversion failed. 398 static StringType GetHFSDecomposedForm(const FilePath::StringType& string); 399 400 // Special UTF-8 version of FastUnicodeCompare. Cf: 401 // http://developer.apple.com/mac/library/technotes/tn/tn1150.html#StringComparisonAlgorithm 402 // IMPORTANT: The input strings must be in the special HFS decomposed form! 403 // (cf. above GetHFSDecomposedForm method) 404 static int HFSFastUnicodeCompare(const StringType& string1, 405 const StringType& string2); 406#endif 407 408 private: 409 // Remove trailing separators from this object. If the path is absolute, it 410 // will never be stripped any more than to refer to the absolute root 411 // directory, so "////" will become "/", not "". A leading pair of 412 // separators is never stripped, to support alternate roots. This is used to 413 // support UNC paths on Windows. 414 void StripTrailingSeparatorsInternal(); 415 416 StringType path_; 417}; 418 419} // namespace base 420 421// This is required by googletest to print a readable output on test failures. 422BASE_EXPORT extern void PrintTo(const base::FilePath& path, std::ostream* out); 423 424// Macros for string literal initialization of FilePath::CharType[], and for 425// using a FilePath::CharType[] in a printf-style format string. 426#if defined(OS_POSIX) 427#define FILE_PATH_LITERAL(x) x 428#define PRFilePath "s" 429#define PRFilePathLiteral "%s" 430#elif defined(OS_WIN) 431#define FILE_PATH_LITERAL(x) L ## x 432#define PRFilePath "ls" 433#define PRFilePathLiteral L"%ls" 434#endif // OS_WIN 435 436// Provide a hash function so that hash_sets and maps can contain FilePath 437// objects. 438namespace BASE_HASH_NAMESPACE { 439#if defined(COMPILER_GCC) 440 441template<> 442struct hash<base::FilePath> { 443 size_t operator()(const base::FilePath& f) const { 444 return hash<base::FilePath::StringType>()(f.value()); 445 } 446}; 447 448#elif defined(COMPILER_MSVC) 449 450inline size_t hash_value(const base::FilePath& f) { 451 return hash_value(f.value()); 452} 453 454#endif // COMPILER 455 456} // namespace BASE_HASH_NAMESPACE 457 458#endif // BASE_FILES_FILE_PATH_H_ 459