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 import genericpath
  16 from 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
  27 curdir = '.'
  28 pardir = '..'
  29 extsep = '.'
  30 sep = '/'
  31 pathsep = ':'
  32 defpath = ':/bin:/usr/bin'
  33 altsep = None
  34 devnull = '/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
  41 def 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
  49 def 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
  58 def 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
  78 def 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
  93 def splitext(p):
  94     return genericpath._splitext(p, sep, altsep, extsep)
  95 splitext.__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
 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, same as split(path)[1].
 107
 108 def 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
 116 def 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
 128 def 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
 138 def 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
 149 def 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
 159 def 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
 169 def 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
 178 def 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
 204 def 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
 219     try:
 220         names = os.listdir(top)
 221     except os.error:
 222         return
 223     func(arg, top, names)
 224     for name in names:
 225         name = join(top, name)
 226         try:
 227             st = os.lstat(name)
 228         except os.error:
 229             continue
 230         if stat.S_ISDIR(st.st_mode):
 231             walk(name, func, arg)
 232
 233
 234 # Expand paths beginning with '~' or '~user'.
 235 # '~' means $HOME; '~user' means that user's home directory.
 236 # If the path doesn't begin with '~', or if the user or $HOME is unknown,
 237 # the path is returned unchanged (leaving error reporting to whatever
 238 # function is called with the expanded path as argument).
 239 # See also module 'glob' for expansion of *, ? and [...] in pathnames.
 240 # (A function should also be defined to do full *sh-style environment
 241 # variable expansion.)
 242
 243 def expanduser(path):
 244     """Expand ~ and ~user constructions.  If user or $HOME is unknown,
 245     do nothing."""
 246     if not path.startswith('~'):
 247         return path
 248     i = path.find('/', 1)
 249     if i < 0:
 250         i = len(path)
 251     if i == 1:
 252         if 'HOME' not in os.environ:
 253             import pwd
 254             userhome = pwd.getpwuid(os.getuid()).pw_dir
 255         else:
 256             userhome = os.environ['HOME']
 257     else:
 258         import pwd
 259         try:
 260             pwent = pwd.getpwnam(path[1:i])
 261         except KeyError:
 262             return path
 263         userhome = pwent.pw_dir
 264     userhome = userhome.rstrip('/')
 265     return userhome + path[i:]
 266
 267
 268 # Expand paths containing shell variable substitutions.
 269 # This expands the forms $variable and ${variable} only.
 270 # Non-existent variables are left unchanged.
 271
 272 _varprog = None
 273
 274 def expandvars(path):
 275     """Expand shell variables of form $var and ${var}.  Unknown variables
 276     are left unchanged."""
 277     global _varprog
 278     if '$' not in path:
 279         return path
 280     if not _varprog:
 281         import re
 282         _varprog = re.compile(r'\$(\w+|\{[^}]*\})')
 283     i = 0
 284     while True:
 285         m = _varprog.search(path, i)
 286         if not m:
 287             break
 288         i, j = m.span(0)
 289         name = m.group(1)
 290         if name.startswith('{') and name.endswith('}'):
 291             name = name[1:-1]
 292         if name in os.environ:
 293             tail = path[j:]
 294             path = path[:i] + os.environ[name]
 295             i = len(path)
 296             path += tail
 297         else:
 298             i = j
 299     return path
 300
 301
 302 # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B.
 303 # It should be understood that this may change the meaning of the path
 304 # if it contains symbolic links!
 305
 306 def normpath(path):
 307     """Normalize path, eliminating double slashes, etc."""
 308     if path == '':
 309         return '.'
 310     initial_slashes = path.startswith('/')
 311     # POSIX allows one or two initial slashes, but treats three or more
 312     # as single slash.
 313     if (initial_slashes and
 314         path.startswith('//') and not path.startswith('///')):
 315         initial_slashes = 2
 316     comps = path.split('/')
 317     new_comps = []
 318     for comp in comps:
 319         if comp in ('', '.'):
 320             continue
 321         if (comp != '..' or (not initial_slashes and not new_comps) or
 322              (new_comps and new_comps[-1] == '..')):
 323             new_comps.append(comp)
 324         elif new_comps:
 325             new_comps.pop()
 326     comps = new_comps
 327     path = '/'.join(comps)
 328     if initial_slashes:
 329         path = '/'*initial_slashes + path
 330     return path or '.'
 331
 332
 333 def abspath(path):
 334     """Return an absolute path."""
 335     if not isabs(path):
 336         path = join(os.getcwd(), path)
 337     return normpath(path)
 338
 339
 340 # Return a canonical path (i.e. the absolute location of a file on the
 341 # filesystem).
 342
 343 def realpath(filename):
 344     """Return the canonical path of the specified filename, eliminating any
 345 symbolic links encountered in the path."""
 346     if isabs(filename):
 347         bits = ['/'] + filename.split('/')[1:]
 348     else:
 349         bits = [''] + filename.split('/')
 350
 351     for i in range(2, len(bits)+1):
 352         component = join(*bits[0:i])
 353         # Resolve symbolic links.
 354         if islink(component):
 355             resolved = _resolve_link(component)
 356             if resolved is None:
 357                 # Infinite loop -- return original component + rest of the path
 358                 return abspath(join(*([component] + bits[i:])))
 359             else:
 360                 newpath = join(*([resolved] + bits[i:]))
 361                 return realpath(newpath)
 362
 363     return abspath(filename)
 364
 365
 366 def _resolve_link(path):
 367     """Internal helper function.  Takes a path and follows symlinks
 368     until we either arrive at something that isn't a symlink, or
 369     encounter a path we've seen before (meaning that there's a loop).
 370     """
 371     paths_seen = []
 372     while islink(path):
 373         if path in paths_seen:
 374             # Already seen this path, so we must have a symlink loop
 375             return None
 376         paths_seen.append(path)
 377         # Resolve where the link points to
 378         resolved = os.readlink(path)
 379         if not isabs(resolved):
 380             dir = dirname(path)
 381             path = normpath(join(dir, resolved))
 382         else:
 383             path = normpath(resolved)
 384     return path
 385
 386 supports_unicode_filenames = False
 387
 388 def relpath(path, start=curdir):
 389     """Return a relative version of a path"""
 390
 391     if not path:
 392         raise ValueError("no path specified")
 393
 394     start_list = abspath(start).split(sep)
 395     path_list = abspath(path).split(sep)
 396
 397     # Work out how much of the filepath is shared by start and path.
 398     i = len(commonprefix([start_list, path_list]))
 399
 400     rel_list = [pardir] * (len(start_list)-i) + path_list[i:]
 401     if not rel_list:
 402         return curdir
 403     return join(*rel_list)