xref: /freebsd/contrib/llvm-project/llvm/include/llvm/Support/Path.h (revision acb1f1269c6f4ff89a0d28ba742f6687e9ef779d)
1 //===- llvm/Support/Path.h - Path Operating System Concept ------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file declares the llvm::sys::path namespace. It is designed after
10 // TR2/boost filesystem (v3), but modified to remove exception handling and the
11 // path class.
12 //
13 //===----------------------------------------------------------------------===//
14 
15 #ifndef LLVM_SUPPORT_PATH_H
16 #define LLVM_SUPPORT_PATH_H
17 
18 #include "llvm/ADT/Twine.h"
19 #include "llvm/ADT/iterator.h"
20 #include "llvm/Support/DataTypes.h"
21 #include <iterator>
22 #include <system_error>
23 
24 namespace llvm {
25 namespace sys {
26 namespace path {
27 
28 enum class Style { windows, posix, native };
29 
30 /// @name Lexical Component Iterator
31 /// @{
32 
33 /// Path iterator.
34 ///
35 /// This is an input iterator that iterates over the individual components in
36 /// \a path. The traversal order is as follows:
37 /// * The root-name element, if present.
38 /// * The root-directory element, if present.
39 /// * Each successive filename element, if present.
40 /// * Dot, if one or more trailing non-root slash characters are present.
41 /// Traversing backwards is possible with \a reverse_iterator
42 ///
43 /// Iteration examples. Each component is separated by ',':
44 /// @code
45 ///   /          => /
46 ///   /foo       => /,foo
47 ///   foo/       => foo,.
48 ///   /foo/bar   => /,foo,bar
49 ///   ../        => ..,.
50 ///   C:\foo\bar => C:,\,foo,bar
51 /// @endcode
52 class const_iterator
53     : public iterator_facade_base<const_iterator, std::input_iterator_tag,
54                                   const StringRef> {
55   StringRef Path;          ///< The entire path.
56   StringRef Component;     ///< The current component. Not necessarily in Path.
57   size_t    Position = 0;  ///< The iterators current position within Path.
58   Style S = Style::native; ///< The path style to use.
59 
60   // An end iterator has Position = Path.size() + 1.
61   friend const_iterator begin(StringRef path, Style style);
62   friend const_iterator end(StringRef path);
63 
64 public:
65   reference operator*() const { return Component; }
66   const_iterator &operator++();    // preincrement
67   bool operator==(const const_iterator &RHS) const;
68 
69   /// Difference in bytes between this and RHS.
70   ptrdiff_t operator-(const const_iterator &RHS) const;
71 };
72 
73 /// Reverse path iterator.
74 ///
75 /// This is an input iterator that iterates over the individual components in
76 /// \a path in reverse order. The traversal order is exactly reversed from that
77 /// of \a const_iterator
78 class reverse_iterator
79     : public iterator_facade_base<reverse_iterator, std::input_iterator_tag,
80                                   const StringRef> {
81   StringRef Path;          ///< The entire path.
82   StringRef Component;     ///< The current component. Not necessarily in Path.
83   size_t    Position = 0;  ///< The iterators current position within Path.
84   Style S = Style::native; ///< The path style to use.
85 
86   friend reverse_iterator rbegin(StringRef path, Style style);
87   friend reverse_iterator rend(StringRef path);
88 
89 public:
90   reference operator*() const { return Component; }
91   reverse_iterator &operator++();    // preincrement
92   bool operator==(const reverse_iterator &RHS) const;
93 
94   /// Difference in bytes between this and RHS.
95   ptrdiff_t operator-(const reverse_iterator &RHS) const;
96 };
97 
98 /// Get begin iterator over \a path.
99 /// @param path Input path.
100 /// @returns Iterator initialized with the first component of \a path.
101 const_iterator begin(StringRef path, Style style = Style::native);
102 
103 /// Get end iterator over \a path.
104 /// @param path Input path.
105 /// @returns Iterator initialized to the end of \a path.
106 const_iterator end(StringRef path);
107 
108 /// Get reverse begin iterator over \a path.
109 /// @param path Input path.
110 /// @returns Iterator initialized with the first reverse component of \a path.
111 reverse_iterator rbegin(StringRef path, Style style = Style::native);
112 
113 /// Get reverse end iterator over \a path.
114 /// @param path Input path.
115 /// @returns Iterator initialized to the reverse end of \a path.
116 reverse_iterator rend(StringRef path);
117 
118 /// @}
119 /// @name Lexical Modifiers
120 /// @{
121 
122 /// Remove the last component from \a path unless it is the root dir.
123 ///
124 /// Similar to the POSIX "dirname" utility.
125 ///
126 /// @code
127 ///   directory/filename.cpp => directory/
128 ///   directory/             => directory
129 ///   filename.cpp           => <empty>
130 ///   /                      => /
131 /// @endcode
132 ///
133 /// @param path A path that is modified to not have a file component.
134 void remove_filename(SmallVectorImpl<char> &path, Style style = Style::native);
135 
136 /// Replace the file extension of \a path with \a extension.
137 ///
138 /// @code
139 ///   ./filename.cpp => ./filename.extension
140 ///   ./filename     => ./filename.extension
141 ///   ./             => ./.extension
142 /// @endcode
143 ///
144 /// @param path A path that has its extension replaced with \a extension.
145 /// @param extension The extension to be added. It may be empty. It may also
146 ///                  optionally start with a '.', if it does not, one will be
147 ///                  prepended.
148 void replace_extension(SmallVectorImpl<char> &path, const Twine &extension,
149                        Style style = Style::native);
150 
151 /// Replace matching path prefix with another path.
152 ///
153 /// @code
154 ///   /foo, /old, /new => /foo
155 ///   /old, /old, /new => /new
156 ///   /old, /old/, /new => /old
157 ///   /old/foo, /old, /new => /new/foo
158 ///   /old/foo, /old/, /new => /new/foo
159 ///   /old/foo, /old/, /new/ => /new/foo
160 ///   /oldfoo, /old, /new => /oldfoo
161 ///   /foo, <empty>, /new => /new/foo
162 ///   /foo, <empty>, new => new/foo
163 ///   /old/foo, /old, <empty> => /foo
164 /// @endcode
165 ///
166 /// @param Path If \a Path starts with \a OldPrefix modify to instead
167 ///        start with \a NewPrefix.
168 /// @param OldPrefix The path prefix to strip from \a Path.
169 /// @param NewPrefix The path prefix to replace \a NewPrefix with.
170 /// @param style The style used to match the prefix. Exact match using
171 /// Posix style, case/separator insensitive match for Windows style.
172 /// @result true if \a Path begins with OldPrefix
173 bool replace_path_prefix(SmallVectorImpl<char> &Path, StringRef OldPrefix,
174                          StringRef NewPrefix,
175                          Style style = Style::native);
176 
177 /// Append to path.
178 ///
179 /// @code
180 ///   /foo  + bar/f => /foo/bar/f
181 ///   /foo/ + bar/f => /foo/bar/f
182 ///   foo   + bar/f => foo/bar/f
183 /// @endcode
184 ///
185 /// @param path Set to \a path + \a component.
186 /// @param a The component to be appended to \a path.
187 void append(SmallVectorImpl<char> &path, const Twine &a,
188                                          const Twine &b = "",
189                                          const Twine &c = "",
190                                          const Twine &d = "");
191 
192 void append(SmallVectorImpl<char> &path, Style style, const Twine &a,
193             const Twine &b = "", const Twine &c = "", const Twine &d = "");
194 
195 /// Append to path.
196 ///
197 /// @code
198 ///   /foo  + [bar,f] => /foo/bar/f
199 ///   /foo/ + [bar,f] => /foo/bar/f
200 ///   foo   + [bar,f] => foo/bar/f
201 /// @endcode
202 ///
203 /// @param path Set to \a path + [\a begin, \a end).
204 /// @param begin Start of components to append.
205 /// @param end One past the end of components to append.
206 void append(SmallVectorImpl<char> &path, const_iterator begin,
207             const_iterator end, Style style = Style::native);
208 
209 /// @}
210 /// @name Transforms (or some other better name)
211 /// @{
212 
213 /// Convert path to the native form. This is used to give paths to users and
214 /// operating system calls in the platform's normal way. For example, on Windows
215 /// all '/' are converted to '\'.
216 ///
217 /// @param path A path that is transformed to native format.
218 /// @param result Holds the result of the transformation.
219 void native(const Twine &path, SmallVectorImpl<char> &result,
220             Style style = Style::native);
221 
222 /// Convert path to the native form in place. This is used to give paths to
223 /// users and operating system calls in the platform's normal way. For example,
224 /// on Windows all '/' are converted to '\'.
225 ///
226 /// @param path A path that is transformed to native format.
227 void native(SmallVectorImpl<char> &path, Style style = Style::native);
228 
229 /// Replaces backslashes with slashes if Windows.
230 ///
231 /// @param path processed path
232 /// @result The result of replacing backslashes with forward slashes if Windows.
233 /// On Unix, this function is a no-op because backslashes are valid path
234 /// chracters.
235 std::string convert_to_slash(StringRef path, Style style = Style::native);
236 
237 /// @}
238 /// @name Lexical Observers
239 /// @{
240 
241 /// Get root name.
242 ///
243 /// @code
244 ///   //net/hello => //net
245 ///   c:/hello    => c: (on Windows, on other platforms nothing)
246 ///   /hello      => <empty>
247 /// @endcode
248 ///
249 /// @param path Input path.
250 /// @result The root name of \a path if it has one, otherwise "".
251 StringRef root_name(StringRef path, Style style = Style::native);
252 
253 /// Get root directory.
254 ///
255 /// @code
256 ///   /goo/hello => /
257 ///   c:/hello   => /
258 ///   d/file.txt => <empty>
259 /// @endcode
260 ///
261 /// @param path Input path.
262 /// @result The root directory of \a path if it has one, otherwise
263 ///               "".
264 StringRef root_directory(StringRef path, Style style = Style::native);
265 
266 /// Get root path.
267 ///
268 /// Equivalent to root_name + root_directory.
269 ///
270 /// @param path Input path.
271 /// @result The root path of \a path if it has one, otherwise "".
272 StringRef root_path(StringRef path, Style style = Style::native);
273 
274 /// Get relative path.
275 ///
276 /// @code
277 ///   C:\hello\world => hello\world
278 ///   foo/bar        => foo/bar
279 ///   /foo/bar       => foo/bar
280 /// @endcode
281 ///
282 /// @param path Input path.
283 /// @result The path starting after root_path if one exists, otherwise "".
284 StringRef relative_path(StringRef path, Style style = Style::native);
285 
286 /// Get parent path.
287 ///
288 /// @code
289 ///   /          => <empty>
290 ///   /foo       => /
291 ///   foo/../bar => foo/..
292 /// @endcode
293 ///
294 /// @param path Input path.
295 /// @result The parent path of \a path if one exists, otherwise "".
296 StringRef parent_path(StringRef path, Style style = Style::native);
297 
298 /// Get filename.
299 ///
300 /// @code
301 ///   /foo.txt    => foo.txt
302 ///   .          => .
303 ///   ..         => ..
304 ///   /          => /
305 /// @endcode
306 ///
307 /// @param path Input path.
308 /// @result The filename part of \a path. This is defined as the last component
309 ///         of \a path. Similar to the POSIX "basename" utility.
310 StringRef filename(StringRef path, Style style = Style::native);
311 
312 /// Get stem.
313 ///
314 /// If filename contains a dot but not solely one or two dots, result is the
315 /// substring of filename ending at (but not including) the last dot. Otherwise
316 /// it is filename.
317 ///
318 /// @code
319 ///   /foo/bar.txt => bar
320 ///   /foo/bar     => bar
321 ///   /foo/.txt    => <empty>
322 ///   /foo/.       => .
323 ///   /foo/..      => ..
324 /// @endcode
325 ///
326 /// @param path Input path.
327 /// @result The stem of \a path.
328 StringRef stem(StringRef path, Style style = Style::native);
329 
330 /// Get extension.
331 ///
332 /// If filename contains a dot but not solely one or two dots, result is the
333 /// substring of filename starting at (and including) the last dot, and ending
334 /// at the end of \a path. Otherwise "".
335 ///
336 /// @code
337 ///   /foo/bar.txt => .txt
338 ///   /foo/bar     => <empty>
339 ///   /foo/.txt    => .txt
340 /// @endcode
341 ///
342 /// @param path Input path.
343 /// @result The extension of \a path.
344 StringRef extension(StringRef path, Style style = Style::native);
345 
346 /// Check whether the given char is a path separator on the host OS.
347 ///
348 /// @param value a character
349 /// @result true if \a value is a path separator character on the host OS
350 bool is_separator(char value, Style style = Style::native);
351 
352 /// Return the preferred separator for this platform.
353 ///
354 /// @result StringRef of the preferred separator, null-terminated.
355 StringRef get_separator(Style style = Style::native);
356 
357 /// Get the typical temporary directory for the system, e.g.,
358 /// "/var/tmp" or "C:/TEMP"
359 ///
360 /// @param erasedOnReboot Whether to favor a path that is erased on reboot
361 /// rather than one that potentially persists longer. This parameter will be
362 /// ignored if the user or system has set the typical environment variable
363 /// (e.g., TEMP on Windows, TMPDIR on *nix) to specify a temporary directory.
364 ///
365 /// @param result Holds the resulting path name.
366 void system_temp_directory(bool erasedOnReboot, SmallVectorImpl<char> &result);
367 
368 /// Get the user's home directory.
369 ///
370 /// @param result Holds the resulting path name.
371 /// @result True if a home directory is set, false otherwise.
372 bool home_directory(SmallVectorImpl<char> &result);
373 
374 /// Get the directory where packages should read user-specific configurations.
375 /// e.g. $XDG_CONFIG_HOME.
376 ///
377 /// @param result Holds the resulting path name.
378 /// @result True if the appropriate path was determined, it need not exist.
379 bool user_config_directory(SmallVectorImpl<char> &result);
380 
381 /// Get the directory where installed packages should put their
382 /// machine-local cache, e.g. $XDG_CACHE_HOME.
383 ///
384 /// @param result Holds the resulting path name.
385 /// @result True if the appropriate path was determined, it need not exist.
386 bool cache_directory(SmallVectorImpl<char> &result);
387 
388 /// Has root name?
389 ///
390 /// root_name != ""
391 ///
392 /// @param path Input path.
393 /// @result True if the path has a root name, false otherwise.
394 bool has_root_name(const Twine &path, Style style = Style::native);
395 
396 /// Has root directory?
397 ///
398 /// root_directory != ""
399 ///
400 /// @param path Input path.
401 /// @result True if the path has a root directory, false otherwise.
402 bool has_root_directory(const Twine &path, Style style = Style::native);
403 
404 /// Has root path?
405 ///
406 /// root_path != ""
407 ///
408 /// @param path Input path.
409 /// @result True if the path has a root path, false otherwise.
410 bool has_root_path(const Twine &path, Style style = Style::native);
411 
412 /// Has relative path?
413 ///
414 /// relative_path != ""
415 ///
416 /// @param path Input path.
417 /// @result True if the path has a relative path, false otherwise.
418 bool has_relative_path(const Twine &path, Style style = Style::native);
419 
420 /// Has parent path?
421 ///
422 /// parent_path != ""
423 ///
424 /// @param path Input path.
425 /// @result True if the path has a parent path, false otherwise.
426 bool has_parent_path(const Twine &path, Style style = Style::native);
427 
428 /// Has filename?
429 ///
430 /// filename != ""
431 ///
432 /// @param path Input path.
433 /// @result True if the path has a filename, false otherwise.
434 bool has_filename(const Twine &path, Style style = Style::native);
435 
436 /// Has stem?
437 ///
438 /// stem != ""
439 ///
440 /// @param path Input path.
441 /// @result True if the path has a stem, false otherwise.
442 bool has_stem(const Twine &path, Style style = Style::native);
443 
444 /// Has extension?
445 ///
446 /// extension != ""
447 ///
448 /// @param path Input path.
449 /// @result True if the path has a extension, false otherwise.
450 bool has_extension(const Twine &path, Style style = Style::native);
451 
452 /// Is path absolute?
453 ///
454 /// According to cppreference.com, C++17 states: "An absolute path is a path
455 /// that unambiguously identifies the location of a file without reference to
456 /// an additional starting location."
457 ///
458 /// In other words, the rules are:
459 /// 1) POSIX style paths with nonempty root directory are absolute.
460 /// 2) Windows style paths with nonempty root name and root directory are
461 ///    absolute.
462 /// 3) No other paths are absolute.
463 ///
464 /// \see has_root_name
465 /// \see has_root_directory
466 ///
467 /// @param path Input path.
468 /// @result True if the path is absolute, false if it is not.
469 bool is_absolute(const Twine &path, Style style = Style::native);
470 
471 /// Is path absolute using GNU rules?
472 ///
473 /// GNU rules are:
474 /// 1) Paths starting with a path separator are absolute.
475 /// 2) Windows style paths are also absolute if they start with a character
476 ///    followed by ':'.
477 /// 3) No other paths are absolute.
478 ///
479 /// On Windows style the path "C:\Users\Default" has "C:" as root name and "\"
480 /// as root directory.
481 ///
482 /// Hence "C:" on Windows is absolute under GNU rules and not absolute under
483 /// C++17 because it has no root directory. Likewise "/" and "\" on Windows are
484 /// absolute under GNU and are not absolute under C++17 due to empty root name.
485 ///
486 /// \see has_root_name
487 /// \see has_root_directory
488 ///
489 /// @param path Input path.
490 /// @param style The style of \p path (e.g. Windows or POSIX). "native" style
491 /// means to derive the style from the host.
492 /// @result True if the path is absolute following GNU rules, false if it is
493 /// not.
494 bool is_absolute_gnu(const Twine &path, Style style = Style::native);
495 
496 /// Is path relative?
497 ///
498 /// @param path Input path.
499 /// @result True if the path is relative, false if it is not.
500 bool is_relative(const Twine &path, Style style = Style::native);
501 
502 /// Remove redundant leading "./" pieces and consecutive separators.
503 ///
504 /// @param path Input path.
505 /// @result The cleaned-up \a path.
506 StringRef remove_leading_dotslash(StringRef path, Style style = Style::native);
507 
508 /// In-place remove any './' and optionally '../' components from a path.
509 ///
510 /// @param path processed path
511 /// @param remove_dot_dot specify if '../' (except for leading "../") should be
512 /// removed
513 /// @result True if path was changed
514 bool remove_dots(SmallVectorImpl<char> &path, bool remove_dot_dot = false,
515                  Style style = Style::native);
516 
517 } // end namespace path
518 } // end namespace sys
519 } // end namespace llvm
520 
521 #endif
522