This commit was manufactured by cvs2svn to create tag 'r22a4-fork'.
[python/dscho.git] / Lib / posixpath.py
blobc587b68b89ea83fef1d1b069ed74dbcc5431a1cb
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","islink","exists","isdir","isfile","ismount",
19 "walk","expanduser","expandvars","normpath","abspath",
20 "samefile","sameopenfile","samestat"]
22 # Normalize the case of a pathname. Trivial in Posix, string.lower on Mac.
23 # On MS-DOS this may also turn slashes into backslashes; however, other
24 # normalizations (such as optimizing '../' away) are not allowed
25 # (another function should be defined to do that).
27 def normcase(s):
28 """Normalize case of pathname. Has no effect under Posix"""
29 return s
32 # Return whether a path is absolute.
33 # Trivial in Posix, harder on the Mac or MS-DOS.
35 def isabs(s):
36 """Test whether a path is absolute"""
37 return s[:1] == '/'
40 # Join pathnames.
41 # Ignore the previous parts if a part is absolute.
42 # Insert a '/' unless the first part is empty or already ends in '/'.
44 def join(a, *p):
45 """Join two or more pathname components, inserting '/' as needed"""
46 path = a
47 for b in p:
48 if b[:1] == '/':
49 path = b
50 elif path == '' or path[-1:] == '/':
51 path = path + b
52 else:
53 path = path + '/' + b
54 return path
57 # Split a path in head (everything up to the last '/') and tail (the
58 # rest). If the path ends in '/', tail will be empty. If there is no
59 # '/' in the path, head will be empty.
60 # Trailing '/'es are stripped from head unless it is the root.
62 def split(p):
63 """Split a pathname. Returns tuple "(head, tail)" where "tail" is
64 everything after the final slash. Either part may be empty."""
65 i = p.rfind('/') + 1
66 head, tail = p[:i], p[i:]
67 if head and head != '/'*len(head):
68 while head[-1] == '/':
69 head = head[:-1]
70 return head, tail
73 # Split a path in root and extension.
74 # The extension is everything starting at the last dot in the last
75 # pathname component; the root is everything before that.
76 # It is always true that root + ext == p.
78 def splitext(p):
79 """Split the extension from a pathname. Extension is everything from the
80 last dot to the end. Returns "(root, ext)", either part may be empty."""
81 root, ext = '', ''
82 for c in p:
83 if c == '/':
84 root, ext = root + ext + c, ''
85 elif c == '.':
86 if ext:
87 root, ext = root + ext, c
88 else:
89 ext = c
90 elif ext:
91 ext = ext + c
92 else:
93 root = root + c
94 return root, ext
97 # Split a pathname into a drive specification and the rest of the
98 # path. Useful on DOS/Windows/NT; on Unix, the drive is always empty.
100 def splitdrive(p):
101 """Split a pathname into drive and path. On Posix, drive is always
102 empty."""
103 return '', p
106 # Return the tail (basename) part of a path.
108 def basename(p):
109 """Returns the final component of a pathname"""
110 return split(p)[1]
113 # Return the head (dirname) part of a path.
115 def dirname(p):
116 """Returns the directory component of a pathname"""
117 return split(p)[0]
120 # Return the longest prefix of all list elements.
122 def commonprefix(m):
123 "Given a list of pathnames, returns the longest common leading component"
124 if not m: return ''
125 prefix = m[0]
126 for item in m:
127 for i in range(len(prefix)):
128 if prefix[:i+1] != item[:i+1]:
129 prefix = prefix[:i]
130 if i == 0: return ''
131 break
132 return prefix
135 # Get size, mtime, atime of files.
137 def getsize(filename):
138 """Return the size of a file, reported by os.stat()."""
139 st = os.stat(filename)
140 return st[stat.ST_SIZE]
142 def getmtime(filename):
143 """Return the last modification time of a file, reported by os.stat()."""
144 st = os.stat(filename)
145 return st[stat.ST_MTIME]
147 def getatime(filename):
148 """Return the last access time of a file, reported by os.stat()."""
149 st = os.stat(filename)
150 return st[stat.ST_ATIME]
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 0
162 return stat.S_ISLNK(st[stat.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 0
174 return 1
177 # Is a path a directory?
178 # This follows symbolic links, so both islink() and isdir() can be true
179 # for the same path.
181 def isdir(path):
182 """Test whether a path is a directory"""
183 try:
184 st = os.stat(path)
185 except os.error:
186 return 0
187 return stat.S_ISDIR(st[stat.ST_MODE])
190 # Is a path a regular file?
191 # This follows symbolic links, so both islink() and isfile() can be true
192 # for the same path.
194 def isfile(path):
195 """Test whether a path is a regular file"""
196 try:
197 st = os.stat(path)
198 except os.error:
199 return 0
200 return stat.S_ISREG(st[stat.ST_MODE])
203 # Are two filenames really pointing to the same file?
205 def samefile(f1, f2):
206 """Test whether two pathnames reference the same actual file"""
207 s1 = os.stat(f1)
208 s2 = os.stat(f2)
209 return samestat(s1, s2)
212 # Are two open files really referencing the same file?
213 # (Not necessarily the same file descriptor!)
215 def sameopenfile(fp1, fp2):
216 """Test whether two open file objects reference the same file"""
217 s1 = os.fstat(fp1)
218 s2 = os.fstat(fp2)
219 return samestat(s1, s2)
222 # Are two stat buffers (obtained from stat, fstat or lstat)
223 # describing the same file?
225 def samestat(s1, s2):
226 """Test whether two stat buffers reference the same file"""
227 return s1[stat.ST_INO] == s2[stat.ST_INO] and \
228 s1[stat.ST_DEV] == s2[stat.ST_DEV]
231 # Is a path a mount point?
232 # (Does this work for all UNIXes? Is it even guaranteed to work by Posix?)
234 def ismount(path):
235 """Test whether a path is a mount point"""
236 try:
237 s1 = os.stat(path)
238 s2 = os.stat(join(path, '..'))
239 except os.error:
240 return 0 # It doesn't exist -- so not a mount point :-)
241 dev1 = s1[stat.ST_DEV]
242 dev2 = s2[stat.ST_DEV]
243 if dev1 != dev2:
244 return 1 # path/.. on a different device as path
245 ino1 = s1[stat.ST_INO]
246 ino2 = s2[stat.ST_INO]
247 if ino1 == ino2:
248 return 1 # path/.. is the same i-node as path
249 return 0
252 # Directory tree walk.
253 # For each directory under top (including top itself, but excluding
254 # '.' and '..'), func(arg, dirname, filenames) is called, where
255 # dirname is the name of the directory and filenames is the list
256 # of files (and subdirectories etc.) in the directory.
257 # The func may modify the filenames list, to implement a filter,
258 # or to impose a different order of visiting.
260 def walk(top, func, arg):
261 """walk(top,func,arg) calls func(arg, d, files) for each directory "d"
262 in the tree rooted at "top" (including "top" itself). "files" is a list
263 of all the files and subdirs in directory "d".
265 try:
266 names = os.listdir(top)
267 except os.error:
268 return
269 func(arg, top, names)
270 for name in names:
271 name = join(top, name)
272 try:
273 st = os.lstat(name)
274 except os.error:
275 continue
276 if stat.S_ISDIR(st[stat.ST_MODE]):
277 walk(name, func, arg)
280 # Expand paths beginning with '~' or '~user'.
281 # '~' means $HOME; '~user' means that user's home directory.
282 # If the path doesn't begin with '~', or if the user or $HOME is unknown,
283 # the path is returned unchanged (leaving error reporting to whatever
284 # function is called with the expanded path as argument).
285 # See also module 'glob' for expansion of *, ? and [...] in pathnames.
286 # (A function should also be defined to do full *sh-style environment
287 # variable expansion.)
289 def expanduser(path):
290 """Expand ~ and ~user constructions. If user or $HOME is unknown,
291 do nothing."""
292 if path[:1] != '~':
293 return path
294 i, n = 1, len(path)
295 while i < n and path[i] != '/':
296 i = i + 1
297 if i == 1:
298 if not os.environ.has_key('HOME'):
299 return path
300 userhome = os.environ['HOME']
301 else:
302 import pwd
303 try:
304 pwent = pwd.getpwnam(path[1:i])
305 except KeyError:
306 return path
307 userhome = pwent[5]
308 if userhome[-1:] == '/': i = i + 1
309 return userhome + path[i:]
312 # Expand paths containing shell variable substitutions.
313 # This expands the forms $variable and ${variable} only.
314 # Non-existent variables are left unchanged.
316 _varprog = None
318 def expandvars(path):
319 """Expand shell variables of form $var and ${var}. Unknown variables
320 are left unchanged."""
321 global _varprog
322 if '$' not in path:
323 return path
324 if not _varprog:
325 import re
326 _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
327 i = 0
328 while 1:
329 m = _varprog.search(path, i)
330 if not m:
331 break
332 i, j = m.span(0)
333 name = m.group(1)
334 if name[:1] == '{' and name[-1:] == '}':
335 name = name[1:-1]
336 if os.environ.has_key(name):
337 tail = path[j:]
338 path = path[:i] + os.environ[name]
339 i = len(path)
340 path = path + tail
341 else:
342 i = j
343 return path
346 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
347 # It should be understood that this may change the meaning of the path
348 # if it contains symbolic links!
350 def normpath(path):
351 """Normalize path, eliminating double slashes, etc."""
352 if path == '':
353 return '.'
354 initial_slashes = path.startswith('/')
355 # POSIX allows one or two initial slashes, but treats three or more
356 # as single slash.
357 if (initial_slashes and
358 path.startswith('//') and not path.startswith('///')):
359 initial_slashes = 2
360 comps = path.split('/')
361 new_comps = []
362 for comp in comps:
363 if comp in ('', '.'):
364 continue
365 if (comp != '..' or (not initial_slashes and not new_comps) or
366 (new_comps and new_comps[-1] == '..')):
367 new_comps.append(comp)
368 elif new_comps:
369 new_comps.pop()
370 comps = new_comps
371 path = '/'.join(comps)
372 if initial_slashes:
373 path = '/'*initial_slashes + path
374 return path or '.'
377 def abspath(path):
378 """Return an absolute path."""
379 if not isabs(path):
380 path = join(os.getcwd(), path)
381 return normpath(path)
384 # Return a canonical path (i.e. the absolute location of a file on the
385 # filesystem).
387 def realpath(filename):
388 """Return the canonical path of the specified filename, eliminating any
389 symbolic links encountered in the path."""
390 filename = abspath(filename)
392 bits = ['/'] + filename.split('/')[1:]
393 for i in range(2, len(bits)+1):
394 component = join(*bits[0:i])
395 if islink(component):
396 resolved = os.readlink(component)
397 (dir, file) = split(component)
398 resolved = normpath(join(dir, resolved))
399 newpath = join(*([resolved] + bits[i:]))
400 return realpath(newpath)
402 return filename