Lib/posixpath.py

   1 """Common operations on Posix pathnames.
   2
   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).
   8
   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 """
  12
  13 import os
  14 import stat
  15
  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","supports_unicode_filenames"]
  21
  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).
  26
  27 def normcase(s):
  28     """Normalize case of pathname.  Has no effect under Posix"""
  29     return s
  30
  31
  32 # Return whether a path is absolute.
  33 # Trivial in Posix, harder on the Mac or MS-DOS.
  34
  35 def isabs(s):
  36     """Test whether a path is absolute"""
  37     return s[:1] == '/'
  38
  39
  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 '/'.
  43
  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
  55
  56
  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.
  61
  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
  71
  72
  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.
  77
  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
  95
  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
 100 def 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.
 107
 108 def basename(p):
 109     """Returns the final component of a pathname"""
 110     return split(p)[1]
 111
 112
 113 # Return the head (dirname) part of a path.
 114
 115 def dirname(p):
 116     """Returns the directory component of a pathname"""
 117     return split(p)[0]
 118
 119
 120 # Return the longest prefix of all list elements.
 121
 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
 133
 134
 135 # Get size, mtime, atime of files.
 136
 137 def getsize(filename):
 138     """Return the size of a file, reported by os.stat()."""
 139     return os.stat(filename).st_size
 140
 141 def getmtime(filename):
 142     """Return the last modification time of a file, reported by os.stat()."""
 143     return os.stat(filename).st_mtime
 144
 145 def getatime(filename):
 146     """Return the last access time of a file, reported by os.stat()."""
 147     return os.stat(filename).st_atime
 148
 149
 150 # Is a path a symbolic link?
 151 # This will always return false on systems where os.lstat doesn't exist.
 152
 153 def islink(path):
 154     """Test whether a path is a symbolic link"""
 155     try:
 156         st = os.lstat(path)
 157     except (os.error, AttributeError):
 158         return False
 159     return stat.S_ISLNK(st.st_mode)
 160
 161
 162 # Does a path exist?
 163 # This is false for dangling symbolic links.
 164
 165 def exists(path):
 166     """Test whether a path exists.  Returns False for broken symbolic links"""
 167     try:
 168         st = os.stat(path)
 169     except os.error:
 170         return False
 171     return True
 172
 173
 174 # Is a path a directory?
 175 # This follows symbolic links, so both islink() and isdir() can be true
 176 # for the same path.
 177
 178 def isdir(path):
 179     """Test whether a path is a directory"""
 180     try:
 181         st = os.stat(path)
 182     except os.error:
 183         return False
 184     return stat.S_ISDIR(st.st_mode)
 185
 186
 187 # Is a path a regular file?
 188 # This follows symbolic links, so both islink() and isfile() can be true
 189 # for the same path.
 190
 191 def isfile(path):
 192     """Test whether a path is a regular file"""
 193     try:
 194         st = os.stat(path)
 195     except os.error:
 196         return False
 197     return stat.S_ISREG(st.st_mode)
 198
 199
 200 # Are two filenames really pointing to the same file?
 201
 202 def samefile(f1, f2):
 203     """Test whether two pathnames reference the same actual file"""
 204     s1 = os.stat(f1)
 205     s2 = os.stat(f2)
 206     return samestat(s1, s2)
 207
 208
 209 # Are two open files really referencing the same file?
 210 # (Not necessarily the same file descriptor!)
 211
 212 def sameopenfile(fp1, fp2):
 213     """Test whether two open file objects reference the same file"""
 214     s1 = os.fstat(fp1)
 215     s2 = os.fstat(fp2)
 216     return samestat(s1, s2)
 217
 218
 219 # Are two stat buffers (obtained from stat, fstat or lstat)
 220 # describing the same file?
 221
 222 def samestat(s1, s2):
 223     """Test whether two stat buffers reference the same file"""
 224     return s1.st_ino == s2.st_ino and \
 225            s1.st_dev == s2.st_dev
 226
 227
 228 # Is a path a mount point?
 229 # (Does this work for all UNIXes?  Is it even guaranteed to work by Posix?)
 230
 231 def ismount(path):
 232     """Test whether a path is a mount point"""
 233     try:
 234         s1 = os.stat(path)
 235         s2 = os.stat(join(path, '..'))
 236     except os.error:
 237         return False # It doesn't exist -- so not a mount point :-)
 238     dev1 = s1.st_dev
 239     dev2 = s2.st_dev
 240     if dev1 != dev2:
 241         return True     # path/.. on a different device as path
 242     ino1 = s1.st_ino
 243     ino2 = s2.st_ino
 244     if ino1 == ino2:
 245         return True     # path/.. is the same i-node as path
 246     return False
 247
 248
 249 # Directory tree walk.
 250 # For each directory under top (including top itself, but excluding
 251 # '.' and '..'), func(arg, dirname, filenames) is called, where
 252 # dirname is the name of the directory and filenames is the list
 253 # of files (and subdirectories etc.) in the directory.
 254 # The func may modify the filenames list, to implement a filter,
 255 # or to impose a different order of visiting.
 256
 257 def walk(top, func, arg):
 258     """Directory tree walk with callback function.
 259
 260     For each directory in the directory tree rooted at top (including top
 261     itself, but excluding '.' and '..'), call func(arg, dirname, fnames).
 262     dirname is the name of the directory, and fnames a list of the names of
 263     the files and subdirectories in dirname (excluding '.' and '..').  func
 264     may modify the fnames list in-place (e.g. via del or slice assignment),
 265     and walk will only recurse into the subdirectories whose names remain in
 266     fnames; this can be used to implement a filter, or to impose a specific
 267     order of visiting.  No semantics are defined for, or required of, arg,
 268     beyond that arg is always passed to func.  It can be used, e.g., to pass
 269     a filename pattern, or a mutable object designed to accumulate
 270     statistics.  Passing None for arg is common."""
 271
 272     try:
 273         names = os.listdir(top)
 274     except os.error:
 275         return
 276     func(arg, top, names)
 277     for name in names:
 278         name = join(top, name)
 279         try:
 280             st = os.lstat(name)
 281         except os.error:
 282             continue
 283         if stat.S_ISDIR(st.st_mode):
 284             walk(name, func, arg)
 285
 286
 287 # Expand paths beginning with '~' or '~user'.
 288 # '~' means $HOME; '~user' means that user's home directory.
 289 # If the path doesn't begin with '~', or if the user or $HOME is unknown,
 290 # the path is returned unchanged (leaving error reporting to whatever
 291 # function is called with the expanded path as argument).
 292 # See also module 'glob' for expansion of *, ? and [...] in pathnames.
 293 # (A function should also be defined to do full *sh-style environment
 294 # variable expansion.)
 295
 296 def expanduser(path):
 297     """Expand ~ and ~user constructions.  If user or $HOME is unknown,
 298     do nothing."""
 299     if path[:1] != '~':
 300         return path
 301     i, n = 1, len(path)
 302     while i < n and path[i] != '/':
 303         i = i + 1
 304     if i == 1:
 305         if not 'HOME' in os.environ:
 306             import pwd
 307             userhome = pwd.getpwuid(os.getuid())[5]
 308         else:
 309             userhome = os.environ['HOME']
 310     else:
 311         import pwd
 312         try:
 313             pwent = pwd.getpwnam(path[1:i])
 314         except KeyError:
 315             return path
 316         userhome = pwent[5]
 317     if userhome[-1:] == '/': i = i + 1
 318     return userhome + path[i:]
 319
 320
 321 # Expand paths containing shell variable substitutions.
 322 # This expands the forms $variable and ${variable} only.
 323 # Non-existent variables are left unchanged.
 324
 325 _varprog = None
 326
 327 def expandvars(path):
 328     """Expand shell variables of form $var and ${var}.  Unknown variables
 329     are left unchanged."""
 330     global _varprog
 331     if '$' not in path:
 332         return path
 333     if not _varprog:
 334         import re
 335         _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
 336     i = 0
 337     while True:
 338         m = _varprog.search(path, i)
 339         if not m:
 340             break
 341         i, j = m.span(0)
 342         name = m.group(1)
 343         if name[:1] == '{' and name[-1:] == '}':
 344             name = name[1:-1]
 345         if name in os.environ:
 346             tail = path[j:]
 347             path = path[:i] + os.environ[name]
 348             i = len(path)
 349             path = path + tail
 350         else:
 351             i = j
 352     return path
 353
 354
 355 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
 356 # It should be understood that this may change the meaning of the path
 357 # if it contains symbolic links!
 358
 359 def normpath(path):
 360     """Normalize path, eliminating double slashes, etc."""
 361     if path == '':
 362         return '.'
 363     initial_slashes = path.startswith('/')
 364     # POSIX allows one or two initial slashes, but treats three or more
 365     # as single slash.
 366     if (initial_slashes and
 367         path.startswith('//') and not path.startswith('///')):
 368         initial_slashes = 2
 369     comps = path.split('/')
 370     new_comps = []
 371     for comp in comps:
 372         if comp in ('', '.'):
 373             continue
 374         if (comp != '..' or (not initial_slashes and not new_comps) or
 375              (new_comps and new_comps[-1] == '..')):
 376             new_comps.append(comp)
 377         elif new_comps:
 378             new_comps.pop()
 379     comps = new_comps
 380     path = '/'.join(comps)
 381     if initial_slashes:
 382         path = '/'*initial_slashes + path
 383     return path or '.'
 384
 385
 386 def abspath(path):
 387     """Return an absolute path."""
 388     if not isabs(path):
 389         path = join(os.getcwd(), path)
 390     return normpath(path)
 391
 392
 393 # Return a canonical path (i.e. the absolute location of a file on the
 394 # filesystem).
 395
 396 def realpath(filename):
 397     """Return the canonical path of the specified filename, eliminating any
 398 symbolic links encountered in the path."""
 399     filename = abspath(filename)
 400
 401     bits = ['/'] + filename.split('/')[1:]
 402     for i in range(2, len(bits)+1):
 403         component = join(*bits[0:i])
 404         if islink(component):
 405             resolved = os.readlink(component)
 406             (dir, file) = split(component)
 407             resolved = normpath(join(dir, resolved))
 408             newpath = join(*([resolved] + bits[i:]))
 409             return realpath(newpath)
 410
 411     return filename
 412
 413 supports_unicode_filenames = False
 414