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.
15 from genericpath
import *
17 __all__
= ["normcase","isabs","join","splitdrive","split","splitext",
18 "basename","dirname","commonprefix","getsize","getmtime",
19 "getatime","getctime","islink","exists","lexists","isdir","isfile",
20 "ismount","walk","expanduser","expandvars","normpath","abspath",
21 "samefile","sameopenfile","samestat",
22 "curdir","pardir","sep","pathsep","defpath","altsep","extsep",
23 "devnull","realpath","supports_unicode_filenames"]
25 # strings representing various path-related bits and pieces
31 defpath
= ':/bin:/usr/bin'
35 # Normalize the case of a pathname. Trivial in Posix, string.lower on Mac.
36 # On MS-DOS this may also turn slashes into backslashes; however, other
37 # normalizations (such as optimizing '../' away) are not allowed
38 # (another function should be defined to do that).
41 """Normalize case of pathname. Has no effect under Posix"""
45 # Return whether a path is absolute.
46 # Trivial in Posix, harder on the Mac or MS-DOS.
49 """Test whether a path is absolute"""
50 return s
.startswith('/')
54 # Ignore the previous parts if a part is absolute.
55 # Insert a '/' unless the first part is empty or already ends in '/'.
58 """Join two or more pathname components, inserting '/' as needed"""
63 elif path
== '' or path
.endswith('/'):
70 # Split a path in head (everything up to the last '/') and tail (the
71 # rest). If the path ends in '/', tail will be empty. If there is no
72 # '/' in the path, head will be empty.
73 # Trailing '/'es are stripped from head unless it is the root.
76 """Split a pathname. Returns tuple "(head, tail)" where "tail" is
77 everything after the final slash. Either part may be empty."""
79 head
, tail
= p
[:i
], p
[i
:]
80 if head
and head
!= '/'*len(head
):
81 head
= head
.rstrip('/')
85 # Split a path in root and extension.
86 # The extension is everything starting at the last dot in the last
87 # pathname component; the root is everything before that.
88 # It is always true that root + ext == p.
91 """Split the extension from a pathname. Extension is everything from the
92 last dot to the end. Returns "(root, ext)", either part may be empty."""
100 # Split a pathname into a drive specification and the rest of the
101 # path. Useful on DOS/Windows/NT; on Unix, the drive is always empty.
104 """Split a pathname into drive and path. On Posix, drive is always
109 # Return the tail (basename) part of a path, same as split(path)[1].
112 """Returns the final component of a pathname"""
117 # Return the head (dirname) part of a path, same as split(path)[0].
120 """Returns the directory component of a pathname"""
123 if head
and head
!= '/'*len(head
):
124 head
= head
.rstrip('/')
128 # Is a path a symbolic link?
129 # This will always return false on systems where os.lstat doesn't exist.
132 """Test whether a path is a symbolic link"""
135 except (os
.error
, AttributeError):
137 return stat
.S_ISLNK(st
.st_mode
)
139 # Being true for dangling symbolic links is also useful.
142 """Test whether a path exists. Returns True for broken symbolic links"""
150 # Are two filenames really pointing to the same file?
152 def samefile(f1
, f2
):
153 """Test whether two pathnames reference the same actual file"""
156 return samestat(s1
, s2
)
159 # Are two open files really referencing the same file?
160 # (Not necessarily the same file descriptor!)
162 def sameopenfile(fp1
, fp2
):
163 """Test whether two open file objects reference the same file"""
166 return samestat(s1
, s2
)
169 # Are two stat buffers (obtained from stat, fstat or lstat)
170 # describing the same file?
172 def samestat(s1
, s2
):
173 """Test whether two stat buffers reference the same file"""
174 return s1
.st_ino
== s2
.st_ino
and \
175 s1
.st_dev
== s2
.st_dev
178 # Is a path a mount point?
179 # (Does this work for all UNIXes? Is it even guaranteed to work by Posix?)
182 """Test whether a path is a mount point"""
185 s2
= os
.stat(join(path
, '..'))
187 return False # It doesn't exist -- so not a mount point :-)
191 return True # path/.. on a different device as path
195 return True # path/.. is the same i-node as path
199 # Directory tree walk.
200 # For each directory under top (including top itself, but excluding
201 # '.' and '..'), func(arg, dirname, filenames) is called, where
202 # dirname is the name of the directory and filenames is the list
203 # of files (and subdirectories etc.) in the directory.
204 # The func may modify the filenames list, to implement a filter,
205 # or to impose a different order of visiting.
207 def walk(top
, func
, arg
):
208 """Directory tree walk with callback function.
210 For each directory in the directory tree rooted at top (including top
211 itself, but excluding '.' and '..'), call func(arg, dirname, fnames).
212 dirname is the name of the directory, and fnames a list of the names of
213 the files and subdirectories in dirname (excluding '.' and '..'). func
214 may modify the fnames list in-place (e.g. via del or slice assignment),
215 and walk will only recurse into the subdirectories whose names remain in
216 fnames; this can be used to implement a filter, or to impose a specific
217 order of visiting. No semantics are defined for, or required of, arg,
218 beyond that arg is always passed to func. It can be used, e.g., to pass
219 a filename pattern, or a mutable object designed to accumulate
220 statistics. Passing None for arg is common."""
223 names
= os
.listdir(top
)
226 func(arg
, top
, names
)
228 name
= join(top
, name
)
233 if stat
.S_ISDIR(st
.st_mode
):
234 walk(name
, func
, arg
)
237 # Expand paths beginning with '~' or '~user'.
238 # '~' means $HOME; '~user' means that user's home directory.
239 # If the path doesn't begin with '~', or if the user or $HOME is unknown,
240 # the path is returned unchanged (leaving error reporting to whatever
241 # function is called with the expanded path as argument).
242 # See also module 'glob' for expansion of *, ? and [...] in pathnames.
243 # (A function should also be defined to do full *sh-style environment
244 # variable expansion.)
246 def expanduser(path
):
247 """Expand ~ and ~user constructions. If user or $HOME is unknown,
249 if not path
.startswith('~'):
251 i
= path
.find('/', 1)
255 if 'HOME' not in os
.environ
:
257 userhome
= pwd
.getpwuid(os
.getuid()).pw_dir
259 userhome
= os
.environ
['HOME']
263 pwent
= pwd
.getpwnam(path
[1:i
])
266 userhome
= pwent
.pw_dir
267 userhome
= userhome
.rstrip('/')
268 return userhome
+ path
[i
:]
271 # Expand paths containing shell variable substitutions.
272 # This expands the forms $variable and ${variable} only.
273 # Non-existent variables are left unchanged.
277 def expandvars(path
):
278 """Expand shell variables of form $var and ${var}. Unknown variables
279 are left unchanged."""
285 _varprog
= re
.compile(r
'\$(\w+|\{[^}]*\})')
288 m
= _varprog
.search(path
, i
)
293 if name
.startswith('{') and name
.endswith('}'):
295 if name
in os
.environ
:
297 path
= path
[:i
] + os
.environ
[name
]
305 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
306 # It should be understood that this may change the meaning of the path
307 # if it contains symbolic links!
310 """Normalize path, eliminating double slashes, etc."""
313 initial_slashes
= path
.startswith('/')
314 # POSIX allows one or two initial slashes, but treats three or more
316 if (initial_slashes
and
317 path
.startswith('//') and not path
.startswith('///')):
319 comps
= path
.split('/')
322 if comp
in ('', '.'):
324 if (comp
!= '..' or (not initial_slashes
and not new_comps
) or
325 (new_comps
and new_comps
[-1] == '..')):
326 new_comps
.append(comp
)
330 path
= '/'.join(comps
)
332 path
= '/'*initial_slashes
+ path
337 """Return an absolute path."""
339 path
= join(os
.getcwd(), path
)
340 return normpath(path
)
343 # Return a canonical path (i.e. the absolute location of a file on the
346 def realpath(filename
):
347 """Return the canonical path of the specified filename, eliminating any
348 symbolic links encountered in the path."""
350 bits
= ['/'] + filename
.split('/')[1:]
352 bits
= [''] + filename
.split('/')
354 for i
in range(2, len(bits
)+1):
355 component
= join(*bits
[0:i
])
356 # Resolve symbolic links.
357 if islink(component
):
358 resolved
= _resolve_link(component
)
360 # Infinite loop -- return original component + rest of the path
361 return abspath(join(*([component
] + bits
[i
:])))
363 newpath
= join(*([resolved
] + bits
[i
:]))
364 return realpath(newpath
)
366 return abspath(filename
)
369 def _resolve_link(path
):
370 """Internal helper function. Takes a path and follows symlinks
371 until we either arrive at something that isn't a symlink, or
372 encounter a path we've seen before (meaning that there's a loop).
376 if path
in paths_seen
:
377 # Already seen this path, so we must have a symlink loop
379 paths_seen
.append(path
)
380 # Resolve where the link points to
381 resolved
= os
.readlink(path
)
382 if not isabs(resolved
):
384 path
= normpath(join(dir, resolved
))
386 path
= normpath(resolved
)
389 supports_unicode_filenames
= False