I don't think we know of any tests that really leak anymore
[python.git] / Lib / posixpath.py
blob9eac6bc13ebc697798b37aa70bdc4386e5f6b799
1 """Common operations on Posix pathnames.
3 Instead of importing this module directly, import os and refer to
4 this module as os.path. The "os.path" name is an alias for this
5 module on Posix systems; on other systems (e.g. Mac, Windows),
6 os.path provides the same operations in a manner specific to that
7 platform, and is an alias to another module (e.g. macpath, ntpath).
9 Some of this can actually be useful on non-Posix systems too, e.g.
10 for manipulation of the pathname component of URLs.
11 """
13 import os
14 import stat
16 __all__ = ["normcase","isabs","join","splitdrive","split","splitext",
17 "basename","dirname","commonprefix","getsize","getmtime",
18 "getatime","getctime","islink","exists","lexists","isdir","isfile",
19 "ismount","walk","expanduser","expandvars","normpath","abspath",
20 "samefile","sameopenfile","samestat",
21 "curdir","pardir","sep","pathsep","defpath","altsep","extsep",
22 "devnull","realpath","supports_unicode_filenames"]
24 # strings representing various path-related bits and pieces
25 curdir = '.'
26 pardir = '..'
27 extsep = '.'
28 sep = '/'
29 pathsep = ':'
30 defpath = ':/bin:/usr/bin'
31 altsep = None
32 devnull = '/dev/null'
34 # Normalize the case of a pathname. Trivial in Posix, string.lower on Mac.
35 # On MS-DOS this may also turn slashes into backslashes; however, other
36 # normalizations (such as optimizing '../' away) are not allowed
37 # (another function should be defined to do that).
39 def normcase(s):
40 """Normalize case of pathname. Has no effect under Posix"""
41 return s
44 # Return whether a path is absolute.
45 # Trivial in Posix, harder on the Mac or MS-DOS.
47 def isabs(s):
48 """Test whether a path is absolute"""
49 return s.startswith('/')
52 # Join pathnames.
53 # Ignore the previous parts if a part is absolute.
54 # Insert a '/' unless the first part is empty or already ends in '/'.
56 def join(a, *p):
57 """Join two or more pathname components, inserting '/' as needed"""
58 path = a
59 for b in p:
60 if b.startswith('/'):
61 path = b
62 elif path == '' or path.endswith('/'):
63 path += b
64 else:
65 path += '/' + b
66 return path
69 # Split a path in head (everything up to the last '/') and tail (the
70 # rest). If the path ends in '/', tail will be empty. If there is no
71 # '/' in the path, head will be empty.
72 # Trailing '/'es are stripped from head unless it is the root.
74 def split(p):
75 """Split a pathname. Returns tuple "(head, tail)" where "tail" is
76 everything after the final slash. Either part may be empty."""
77 i = p.rfind('/') + 1
78 head, tail = p[:i], p[i:]
79 if head and head != '/'*len(head):
80 head = head.rstrip('/')
81 return head, tail
84 # Split a path in root and extension.
85 # The extension is everything starting at the last dot in the last
86 # pathname component; the root is everything before that.
87 # It is always true that root + ext == p.
89 def splitext(p):
90 """Split the extension from a pathname. Extension is everything from the
91 last dot to the end. Returns "(root, ext)", either part may be empty."""
92 i = p.rfind('.')
93 if i<=p.rfind('/'):
94 return p, ''
95 else:
96 return p[:i], p[i:]
99 # Split a pathname into a drive specification and the rest of the
100 # path. Useful on DOS/Windows/NT; on Unix, the drive is always empty.
102 def splitdrive(p):
103 """Split a pathname into drive and path. On Posix, drive is always
104 empty."""
105 return '', p
108 # Return the tail (basename) part of a path.
110 def basename(p):
111 """Returns the final component of a pathname"""
112 return split(p)[1]
115 # Return the head (dirname) part of a path.
117 def dirname(p):
118 """Returns the directory component of a pathname"""
119 return split(p)[0]
122 # Return the longest prefix of all list elements.
124 def commonprefix(m):
125 "Given a list of pathnames, returns the longest common leading component"
126 if not m: return ''
127 s1 = min(m)
128 s2 = max(m)
129 n = min(len(s1), len(s2))
130 for i in xrange(n):
131 if s1[i] != s2[i]:
132 return s1[:i]
133 return s1[:n]
135 # Get size, mtime, atime of files.
137 def getsize(filename):
138 """Return the size of a file, reported by os.stat()."""
139 return os.stat(filename).st_size
141 def getmtime(filename):
142 """Return the last modification time of a file, reported by os.stat()."""
143 return os.stat(filename).st_mtime
145 def getatime(filename):
146 """Return the last access time of a file, reported by os.stat()."""
147 return os.stat(filename).st_atime
149 def getctime(filename):
150 """Return the metadata change time of a file, reported by os.stat()."""
151 return os.stat(filename).st_ctime
153 # Is a path a symbolic link?
154 # This will always return false on systems where os.lstat doesn't exist.
156 def islink(path):
157 """Test whether a path is a symbolic link"""
158 try:
159 st = os.lstat(path)
160 except (os.error, AttributeError):
161 return False
162 return stat.S_ISLNK(st.st_mode)
165 # Does a path exist?
166 # This is false for dangling symbolic links.
168 def exists(path):
169 """Test whether a path exists. Returns False for broken symbolic links"""
170 try:
171 st = os.stat(path)
172 except os.error:
173 return False
174 return True
177 # Being true for dangling symbolic links is also useful.
179 def lexists(path):
180 """Test whether a path exists. Returns True for broken symbolic links"""
181 try:
182 st = os.lstat(path)
183 except os.error:
184 return False
185 return True
188 # Is a path a directory?
189 # This follows symbolic links, so both islink() and isdir() can be true
190 # for the same path.
192 def isdir(path):
193 """Test whether a path is a directory"""
194 try:
195 st = os.stat(path)
196 except os.error:
197 return False
198 return stat.S_ISDIR(st.st_mode)
201 # Is a path a regular file?
202 # This follows symbolic links, so both islink() and isfile() can be true
203 # for the same path.
205 def isfile(path):
206 """Test whether a path is a regular file"""
207 try:
208 st = os.stat(path)
209 except os.error:
210 return False
211 return stat.S_ISREG(st.st_mode)
214 # Are two filenames really pointing to the same file?
216 def samefile(f1, f2):
217 """Test whether two pathnames reference the same actual file"""
218 s1 = os.stat(f1)
219 s2 = os.stat(f2)
220 return samestat(s1, s2)
223 # Are two open files really referencing the same file?
224 # (Not necessarily the same file descriptor!)
226 def sameopenfile(fp1, fp2):
227 """Test whether two open file objects reference the same file"""
228 s1 = os.fstat(fp1)
229 s2 = os.fstat(fp2)
230 return samestat(s1, s2)
233 # Are two stat buffers (obtained from stat, fstat or lstat)
234 # describing the same file?
236 def samestat(s1, s2):
237 """Test whether two stat buffers reference the same file"""
238 return s1.st_ino == s2.st_ino and \
239 s1.st_dev == s2.st_dev
242 # Is a path a mount point?
243 # (Does this work for all UNIXes? Is it even guaranteed to work by Posix?)
245 def ismount(path):
246 """Test whether a path is a mount point"""
247 try:
248 s1 = os.stat(path)
249 s2 = os.stat(join(path, '..'))
250 except os.error:
251 return False # It doesn't exist -- so not a mount point :-)
252 dev1 = s1.st_dev
253 dev2 = s2.st_dev
254 if dev1 != dev2:
255 return True # path/.. on a different device as path
256 ino1 = s1.st_ino
257 ino2 = s2.st_ino
258 if ino1 == ino2:
259 return True # path/.. is the same i-node as path
260 return False
263 # Directory tree walk.
264 # For each directory under top (including top itself, but excluding
265 # '.' and '..'), func(arg, dirname, filenames) is called, where
266 # dirname is the name of the directory and filenames is the list
267 # of files (and subdirectories etc.) in the directory.
268 # The func may modify the filenames list, to implement a filter,
269 # or to impose a different order of visiting.
271 def walk(top, func, arg):
272 """Directory tree walk with callback function.
274 For each directory in the directory tree rooted at top (including top
275 itself, but excluding '.' and '..'), call func(arg, dirname, fnames).
276 dirname is the name of the directory, and fnames a list of the names of
277 the files and subdirectories in dirname (excluding '.' and '..'). func
278 may modify the fnames list in-place (e.g. via del or slice assignment),
279 and walk will only recurse into the subdirectories whose names remain in
280 fnames; this can be used to implement a filter, or to impose a specific
281 order of visiting. No semantics are defined for, or required of, arg,
282 beyond that arg is always passed to func. It can be used, e.g., to pass
283 a filename pattern, or a mutable object designed to accumulate
284 statistics. Passing None for arg is common."""
286 try:
287 names = os.listdir(top)
288 except os.error:
289 return
290 func(arg, top, names)
291 for name in names:
292 name = join(top, name)
293 try:
294 st = os.lstat(name)
295 except os.error:
296 continue
297 if stat.S_ISDIR(st.st_mode):
298 walk(name, func, arg)
301 # Expand paths beginning with '~' or '~user'.
302 # '~' means $HOME; '~user' means that user's home directory.
303 # If the path doesn't begin with '~', or if the user or $HOME is unknown,
304 # the path is returned unchanged (leaving error reporting to whatever
305 # function is called with the expanded path as argument).
306 # See also module 'glob' for expansion of *, ? and [...] in pathnames.
307 # (A function should also be defined to do full *sh-style environment
308 # variable expansion.)
310 def expanduser(path):
311 """Expand ~ and ~user constructions. If user or $HOME is unknown,
312 do nothing."""
313 if not path.startswith('~'):
314 return path
315 i = path.find('/', 1)
316 if i < 0:
317 i = len(path)
318 if i == 1:
319 if 'HOME' not in os.environ:
320 import pwd
321 userhome = pwd.getpwuid(os.getuid()).pw_dir
322 else:
323 userhome = os.environ['HOME']
324 else:
325 import pwd
326 try:
327 pwent = pwd.getpwnam(path[1:i])
328 except KeyError:
329 return path
330 userhome = pwent.pw_dir
331 if userhome.endswith('/'):
332 i += 1
333 return userhome + path[i:]
336 # Expand paths containing shell variable substitutions.
337 # This expands the forms $variable and ${variable} only.
338 # Non-existent variables are left unchanged.
340 _varprog = None
342 def expandvars(path):
343 """Expand shell variables of form $var and ${var}. Unknown variables
344 are left unchanged."""
345 global _varprog
346 if '$' not in path:
347 return path
348 if not _varprog:
349 import re
350 _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
351 i = 0
352 while True:
353 m = _varprog.search(path, i)
354 if not m:
355 break
356 i, j = m.span(0)
357 name = m.group(1)
358 if name.startswith('{') and name.endswith('}'):
359 name = name[1:-1]
360 if name in os.environ:
361 tail = path[j:]
362 path = path[:i] + os.environ[name]
363 i = len(path)
364 path += tail
365 else:
366 i = j
367 return path
370 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
371 # It should be understood that this may change the meaning of the path
372 # if it contains symbolic links!
374 def normpath(path):
375 """Normalize path, eliminating double slashes, etc."""
376 if path == '':
377 return '.'
378 initial_slashes = path.startswith('/')
379 # POSIX allows one or two initial slashes, but treats three or more
380 # as single slash.
381 if (initial_slashes and
382 path.startswith('//') and not path.startswith('///')):
383 initial_slashes = 2
384 comps = path.split('/')
385 new_comps = []
386 for comp in comps:
387 if comp in ('', '.'):
388 continue
389 if (comp != '..' or (not initial_slashes and not new_comps) or
390 (new_comps and new_comps[-1] == '..')):
391 new_comps.append(comp)
392 elif new_comps:
393 new_comps.pop()
394 comps = new_comps
395 path = '/'.join(comps)
396 if initial_slashes:
397 path = '/'*initial_slashes + path
398 return path or '.'
401 def abspath(path):
402 """Return an absolute path."""
403 if not isabs(path):
404 path = join(os.getcwd(), path)
405 return normpath(path)
408 # Return a canonical path (i.e. the absolute location of a file on the
409 # filesystem).
411 def realpath(filename):
412 """Return the canonical path of the specified filename, eliminating any
413 symbolic links encountered in the path."""
414 if isabs(filename):
415 bits = ['/'] + filename.split('/')[1:]
416 else:
417 bits = [''] + filename.split('/')
419 for i in range(2, len(bits)+1):
420 component = join(*bits[0:i])
421 # Resolve symbolic links.
422 if islink(component):
423 resolved = _resolve_link(component)
424 if resolved is None:
425 # Infinite loop -- return original component + rest of the path
426 return abspath(join(*([component] + bits[i:])))
427 else:
428 newpath = join(*([resolved] + bits[i:]))
429 return realpath(newpath)
431 return abspath(filename)
434 def _resolve_link(path):
435 """Internal helper function. Takes a path and follows symlinks
436 until we either arrive at something that isn't a symlink, or
437 encounter a path we've seen before (meaning that there's a loop).
439 paths_seen = []
440 while islink(path):
441 if path in paths_seen:
442 # Already seen this path, so we must have a symlink loop
443 return None
444 paths_seen.append(path)
445 # Resolve where the link points to
446 resolved = os.readlink(path)
447 if not isabs(resolved):
448 dir = dirname(path)
449 path = normpath(join(dir, resolved))
450 else:
451 path = normpath(resolved)
452 return path
454 supports_unicode_filenames = False