PageRenderTime 478ms CodeModel.GetById 171ms app.highlight 169ms RepoModel.GetById 133ms app.codeStats 0ms

/Lib/posixpath.py

http://unladen-swallow.googlecode.com/
Python | 405 lines | 402 code | 0 blank | 3 comment | 1 complexity | c4da84bcb3918038717ad3b241ddab63 MD5 | raw file
  1"""Common operations on Posix pathnames.
  2
  3Instead of importing this module directly, import os and refer to
  4this module as os.path.  The "os.path" name is an alias for this
  5module on Posix systems; on other systems (e.g. Mac, Windows),
  6os.path provides the same operations in a manner specific to that
  7platform, and is an alias to another module (e.g. macpath, ntpath).
  8
  9Some of this can actually be useful on non-Posix systems too, e.g.
 10for manipulation of the pathname component of URLs.
 11"""
 12
 13import os
 14import stat
 15import genericpath
 16from genericpath import *
 17
 18__all__ = ["normcase","isabs","join","splitdrive","split","splitext",
 19           "basename","dirname","commonprefix","getsize","getmtime",
 20           "getatime","getctime","islink","exists","lexists","isdir","isfile",
 21           "ismount","walk","expanduser","expandvars","normpath","abspath",
 22           "samefile","sameopenfile","samestat",
 23           "curdir","pardir","sep","pathsep","defpath","altsep","extsep",
 24           "devnull","realpath","supports_unicode_filenames","relpath"]
 25
 26# strings representing various path-related bits and pieces
 27curdir = '.'
 28pardir = '..'
 29extsep = '.'
 30sep = '/'
 31pathsep = ':'
 32defpath = ':/bin:/usr/bin'
 33altsep = None
 34devnull = '/dev/null'
 35
 36# Normalize the case of a pathname.  Trivial in Posix, string.lower on Mac.
 37# On MS-DOS this may also turn slashes into backslashes; however, other
 38# normalizations (such as optimizing '../' away) are not allowed
 39# (another function should be defined to do that).
 40
 41def normcase(s):
 42    """Normalize case of pathname.  Has no effect under Posix"""
 43    return s
 44
 45
 46# Return whether a path is absolute.
 47# Trivial in Posix, harder on the Mac or MS-DOS.
 48
 49def isabs(s):
 50    """Test whether a path is absolute"""
 51    return s.startswith('/')
 52
 53
 54# Join pathnames.
 55# Ignore the previous parts if a part is absolute.
 56# Insert a '/' unless the first part is empty or already ends in '/'.
 57
 58def join(a, *p):
 59    """Join two or more pathname components, inserting '/' as needed.
 60    If any component is an absolute path, all previous path components
 61    will be discarded."""
 62    path = a
 63    for b in p:
 64        if b.startswith('/'):
 65            path = b
 66        elif path == '' or path.endswith('/'):
 67            path +=  b
 68        else:
 69            path += '/' + b
 70    return path
 71
 72
 73# Split a path in head (everything up to the last '/') and tail (the
 74# rest).  If the path ends in '/', tail will be empty.  If there is no
 75# '/' in the path, head  will be empty.
 76# Trailing '/'es are stripped from head unless it is the root.
 77
 78def split(p):
 79    """Split a pathname.  Returns tuple "(head, tail)" where "tail" is
 80    everything after the final slash.  Either part may be empty."""
 81    i = p.rfind('/') + 1
 82    head, tail = p[:i], p[i:]
 83    if head and head != '/'*len(head):
 84        head = head.rstrip('/')
 85    return head, tail
 86
 87
 88# Split a path in root and extension.
 89# The extension is everything starting at the last dot in the last
 90# pathname component; the root is everything before that.
 91# It is always true that root + ext == p.
 92
 93def splitext(p):
 94    return genericpath._splitext(p, sep, altsep, extsep)
 95splitext.__doc__ = genericpath._splitext.__doc__
 96
 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.
 99
100def splitdrive(p):
101    """Split a pathname into drive and path. On Posix, drive is always
102    empty."""
103    return '', p
104
105
106# Return the tail (basename) part of a path, same as split(path)[1].
107
108def basename(p):
109    """Returns the final component of a pathname"""
110    i = p.rfind('/') + 1
111    return p[i:]
112
113
114# Return the head (dirname) part of a path, same as split(path)[0].
115
116def dirname(p):
117    """Returns the directory component of a pathname"""
118    i = p.rfind('/') + 1
119    head = p[:i]
120    if head and head != '/'*len(head):
121        head = head.rstrip('/')
122    return head
123
124
125# Is a path a symbolic link?
126# This will always return false on systems where os.lstat doesn't exist.
127
128def islink(path):
129    """Test whether a path is a symbolic link"""
130    try:
131        st = os.lstat(path)
132    except (os.error, AttributeError):
133        return False
134    return stat.S_ISLNK(st.st_mode)
135
136# Being true for dangling symbolic links is also useful.
137
138def lexists(path):
139    """Test whether a path exists.  Returns True for broken symbolic links"""
140    try:
141        st = os.lstat(path)
142    except os.error:
143        return False
144    return True
145
146
147# Are two filenames really pointing to the same file?
148
149def samefile(f1, f2):
150    """Test whether two pathnames reference the same actual file"""
151    s1 = os.stat(f1)
152    s2 = os.stat(f2)
153    return samestat(s1, s2)
154
155
156# Are two open files really referencing the same file?
157# (Not necessarily the same file descriptor!)
158
159def sameopenfile(fp1, fp2):
160    """Test whether two open file objects reference the same file"""
161    s1 = os.fstat(fp1)
162    s2 = os.fstat(fp2)
163    return samestat(s1, s2)
164
165
166# Are two stat buffers (obtained from stat, fstat or lstat)
167# describing the same file?
168
169def samestat(s1, s2):
170    """Test whether two stat buffers reference the same file"""
171    return s1.st_ino == s2.st_ino and \
172           s1.st_dev == s2.st_dev
173
174
175# Is a path a mount point?
176# (Does this work for all UNIXes?  Is it even guaranteed to work by Posix?)
177
178def ismount(path):
179    """Test whether a path is a mount point"""
180    try:
181        s1 = os.lstat(path)
182        s2 = os.lstat(join(path, '..'))
183    except os.error:
184        return False # It doesn't exist -- so not a mount point :-)
185    dev1 = s1.st_dev
186    dev2 = s2.st_dev
187    if dev1 != dev2:
188        return True     # path/.. on a different device as path
189    ino1 = s1.st_ino
190    ino2 = s2.st_ino
191    if ino1 == ino2:
192        return True     # path/.. is the same i-node as path
193    return False
194
195
196# Directory tree walk.
197# For each directory under top (including top itself, but excluding
198# '.' and '..'), func(arg, dirname, filenames) is called, where
199# dirname is the name of the directory and filenames is the list
200# of files (and subdirectories etc.) in the directory.
201# The func may modify the filenames list, to implement a filter,
202# or to impose a different order of visiting.
203
204def walk(top, func, arg):
205    """Directory tree walk with callback function.
206
207    For each directory in the directory tree rooted at top (including top
208    itself, but excluding '.' and '..'), call func(arg, dirname, fnames).
209    dirname is the name of the directory, and fnames a list of the names of
210    the files and subdirectories in dirname (excluding '.' and '..').  func
211    may modify the fnames list in-place (e.g. via del or slice assignment),
212    and walk will only recurse into the subdirectories whose names remain in
213    fnames; this can be used to implement a filter, or to impose a specific
214    order of visiting.  No semantics are defined for, or required of, arg,
215    beyond that arg is always passed to func.  It can be used, e.g., to pass
216    a filename pattern, or a mutable object designed to accumulate
217    statistics.  Passing None for arg is common."""
218    import warnings
219    warnings.warnpy3k("In 3.x, os.path.walk is removed in favor of os.walk.",
220                      stacklevel=2)
221    try:
222        names = os.listdir(top)
223    except os.error:
224        return
225    func(arg, top, names)
226    for name in names:
227        name = join(top, name)
228        try:
229            st = os.lstat(name)
230        except os.error:
231            continue
232        if stat.S_ISDIR(st.st_mode):
233            walk(name, func, arg)
234
235
236# Expand paths beginning with '~' or '~user'.
237# '~' means $HOME; '~user' means that user's home directory.
238# If the path doesn't begin with '~', or if the user or $HOME is unknown,
239# the path is returned unchanged (leaving error reporting to whatever
240# function is called with the expanded path as argument).
241# See also module 'glob' for expansion of *, ? and [...] in pathnames.
242# (A function should also be defined to do full *sh-style environment
243# variable expansion.)
244
245def expanduser(path):
246    """Expand ~ and ~user constructions.  If user or $HOME is unknown,
247    do nothing."""
248    if not path.startswith('~'):
249        return path
250    i = path.find('/', 1)
251    if i < 0:
252        i = len(path)
253    if i == 1:
254        if 'HOME' not in os.environ:
255            import pwd
256            userhome = pwd.getpwuid(os.getuid()).pw_dir
257        else:
258            userhome = os.environ['HOME']
259    else:
260        import pwd
261        try:
262            pwent = pwd.getpwnam(path[1:i])
263        except KeyError:
264            return path
265        userhome = pwent.pw_dir
266    userhome = userhome.rstrip('/') or userhome
267    return userhome + path[i:]
268
269
270# Expand paths containing shell variable substitutions.
271# This expands the forms $variable and ${variable} only.
272# Non-existent variables are left unchanged.
273
274_varprog = None
275
276def expandvars(path):
277    """Expand shell variables of form $var and ${var}.  Unknown variables
278    are left unchanged."""
279    global _varprog
280    if '$' not in path:
281        return path
282    if not _varprog:
283        import re
284        _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
285    i = 0
286    while True:
287        m = _varprog.search(path, i)
288        if not m:
289            break
290        i, j = m.span(0)
291        name = m.group(1)
292        if name.startswith('{') and name.endswith('}'):
293            name = name[1:-1]
294        if name in os.environ:
295            tail = path[j:]
296            path = path[:i] + os.environ[name]
297            i = len(path)
298            path += tail
299        else:
300            i = j
301    return path
302
303
304# Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
305# It should be understood that this may change the meaning of the path
306# if it contains symbolic links!
307
308def normpath(path):
309    """Normalize path, eliminating double slashes, etc."""
310    if path == '':
311        return '.'
312    initial_slashes = path.startswith('/')
313    # POSIX allows one or two initial slashes, but treats three or more
314    # as single slash.
315    if (initial_slashes and
316        path.startswith('//') and not path.startswith('///')):
317        initial_slashes = 2
318    comps = path.split('/')
319    new_comps = []
320    for comp in comps:
321        if comp in ('', '.'):
322            continue
323        if (comp != '..' or (not initial_slashes and not new_comps) or
324             (new_comps and new_comps[-1] == '..')):
325            new_comps.append(comp)
326        elif new_comps:
327            new_comps.pop()
328    comps = new_comps
329    path = '/'.join(comps)
330    if initial_slashes:
331        path = '/'*initial_slashes + path
332    return path or '.'
333
334
335def abspath(path):
336    """Return an absolute path."""
337    if not isabs(path):
338        path = join(os.getcwd(), path)
339    return normpath(path)
340
341
342# Return a canonical path (i.e. the absolute location of a file on the
343# filesystem).
344
345def realpath(filename):
346    """Return the canonical path of the specified filename, eliminating any
347symbolic links encountered in the path."""
348    if isabs(filename):
349        bits = ['/'] + filename.split('/')[1:]
350    else:
351        bits = [''] + filename.split('/')
352
353    for i in range(2, len(bits)+1):
354        component = join(*bits[0:i])
355        # Resolve symbolic links.
356        if islink(component):
357            resolved = _resolve_link(component)
358            if resolved is None:
359                # Infinite loop -- return original component + rest of the path
360                return abspath(join(*([component] + bits[i:])))
361            else:
362                newpath = join(*([resolved] + bits[i:]))
363                return realpath(newpath)
364
365    return abspath(filename)
366
367
368def _resolve_link(path):
369    """Internal helper function.  Takes a path and follows symlinks
370    until we either arrive at something that isn't a symlink, or
371    encounter a path we've seen before (meaning that there's a loop).
372    """
373    paths_seen = []
374    while islink(path):
375        if path in paths_seen:
376            # Already seen this path, so we must have a symlink loop
377            return None
378        paths_seen.append(path)
379        # Resolve where the link points to
380        resolved = os.readlink(path)
381        if not isabs(resolved):
382            dir = dirname(path)
383            path = normpath(join(dir, resolved))
384        else:
385            path = normpath(resolved)
386    return path
387
388supports_unicode_filenames = False
389
390def relpath(path, start=curdir):
391    """Return a relative version of a path"""
392
393    if not path:
394        raise ValueError("no path specified")
395
396    start_list = abspath(start).split(sep)
397    path_list = abspath(path).split(sep)
398
399    # Work out how much of the filepath is shared by start and path.
400    i = len(commonprefix([start_list, path_list]))
401
402    rel_list = [pardir] * (len(start_list)-i) + path_list[i:]
403    if not rel_list:
404        return curdir
405    return join(*rel_list)