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"]
  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             return path
 307         userhome = os.environ['HOME']
 308     else:
 309         import pwd
 310         try:
 311             pwent = pwd.getpwnam(path[1:i])
 312         except KeyError:
 313             return path
 314         userhome = pwent[5]
 315     if userhome[-1:] == '/': i = i + 1
 316     return userhome + path[i:]
 317
 318
 319 # Expand paths containing shell variable substitutions.
 320 # This expands the forms $variable and ${variable} only.
 321 # Non-existent variables are left unchanged.
 322
 323 _varprog = None
 324
 325 def expandvars(path):
 326     """Expand shell variables of form $var and ${var}.  Unknown variables
 327     are left unchanged."""
 328     global _varprog
 329     if '$' not in path:
 330         return path
 331     if not _varprog:
 332         import re
 333         _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
 334     i = 0
 335     while True:
 336         m = _varprog.search(path, i)
 337         if not m:
 338             break
 339         i, j = m.span(0)
 340         name = m.group(1)
 341         if name[:1] == '{' and name[-1:] == '}':
 342             name = name[1:-1]
 343         if name in os.environ:
 344             tail = path[j:]
 345             path = path[:i] + os.environ[name]
 346             i = len(path)
 347             path = path + tail
 348         else:
 349             i = j
 350     return path
 351
 352
 353 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
 354 # It should be understood that this may change the meaning of the path
 355 # if it contains symbolic links!
 356
 357 def normpath(path):
 358     """Normalize path, eliminating double slashes, etc."""
 359     if path == '':
 360         return '.'
 361     initial_slashes = path.startswith('/')
 362     # POSIX allows one or two initial slashes, but treats three or more
 363     # as single slash.
 364     if (initial_slashes and
 365         path.startswith('//') and not path.startswith('///')):
 366         initial_slashes = 2
 367     comps = path.split('/')
 368     new_comps = []
 369     for comp in comps:
 370         if comp in ('', '.'):
 371             continue
 372         if (comp != '..' or (not initial_slashes and not new_comps) or
 373              (new_comps and new_comps[-1] == '..')):
 374             new_comps.append(comp)
 375         elif new_comps:
 376             new_comps.pop()
 377     comps = new_comps
 378     path = '/'.join(comps)
 379     if initial_slashes:
 380         path = '/'*initial_slashes + path
 381     return path or '.'
 382
 383
 384 def abspath(path):
 385     """Return an absolute path."""
 386     if not isabs(path):
 387         path = join(os.getcwd(), path)
 388     return normpath(path)
 389
 390
 391 # Return a canonical path (i.e. the absolute location of a file on the
 392 # filesystem).
 393
 394 def realpath(filename):
 395     """Return the canonical path of the specified filename, eliminating any
 396 symbolic links encountered in the path."""
 397     filename = abspath(filename)
 398
 399     bits = ['/'] + filename.split('/')[1:]
 400     for i in range(2, len(bits)+1):
 401         component = join(*bits[0:i])
 402         if islink(component):
 403             resolved = os.readlink(component)
 404             (dir, file) = split(component)
 405             resolved = normpath(join(dir, resolved))
 406             newpath = join(*([resolved] + bits[i:]))
 407             return realpath(newpath)
 408
 409     return filename