| 1 | # Module 'ntpath' -- common operations on WinNT/Win95 pathnames |
| 2 | """Common pathname manipulations, WindowsNT/95 version. |
| 3 | |
| 4 | Instead of importing this module directly, import os and refer to this |
| 5 | module as os.path. |
| 6 | """ |
| 7 | |
| 8 | import os |
| 9 | import stat |
| 10 | import sys |
| 11 | |
| 12 | __all__ = ["normcase","isabs","join","splitdrive","split","splitext", |
| 13 | "basename","dirname","commonprefix","getsize","getmtime", |
| 14 | "getatime","getctime", "islink","exists","lexists","isdir","isfile", |
| 15 | "ismount","walk","expanduser","expandvars","normpath","abspath", |
| 16 | "splitunc","curdir","pardir","sep","pathsep","defpath","altsep", |
| 17 | "extsep","devnull","realpath","supports_unicode_filenames"] |
| 18 | |
| 19 | # strings representing various path-related bits and pieces |
| 20 | curdir = '.' |
| 21 | pardir = '..' |
| 22 | extsep = '.' |
| 23 | sep = '\\' |
| 24 | pathsep = ';' |
| 25 | altsep = '/' |
| 26 | defpath = '.;C:\\bin' |
| 27 | if 'ce' in sys.builtin_module_names: |
| 28 | defpath = '\\Windows' |
| 29 | elif 'os2' in sys.builtin_module_names: |
| 30 | # OS/2 w/ VACPP |
| 31 | altsep = '/' |
| 32 | devnull = 'nul' |
| 33 | |
| 34 | # Normalize the case of a pathname and map slashes to backslashes. |
| 35 | # Other normalizations (such as optimizing '../' away) are not done |
| 36 | # (this is done by normpath). |
| 37 | |
| 38 | def normcase(s): |
| 39 | """Normalize case of pathname. |
| 40 | |
| 41 | Makes all characters lowercase and all slashes into backslashes.""" |
| 42 | return s.replace("/", "\\").lower() |
| 43 | |
| 44 | |
| 45 | # Return whether a path is absolute. |
| 46 | # Trivial in Posix, harder on the Mac or MS-DOS. |
| 47 | # For DOS it is absolute if it starts with a slash or backslash (current |
| 48 | # volume), or if a pathname after the volume letter and colon / UNC resource |
| 49 | # starts with a slash or backslash. |
| 50 | |
| 51 | def isabs(s): |
| 52 | """Test whether a path is absolute""" |
| 53 | s = splitdrive(s)[1] |
| 54 | return s != '' and s[:1] in '/\\' |
| 55 | |
| 56 | |
| 57 | # Join two (or more) paths. |
| 58 | |
| 59 | def join(a, *p): |
| 60 | """Join two or more pathname components, inserting "\\" as needed""" |
| 61 | path = a |
| 62 | for b in p: |
| 63 | b_wins = 0 # set to 1 iff b makes path irrelevant |
| 64 | if path == "": |
| 65 | b_wins = 1 |
| 66 | |
| 67 | elif isabs(b): |
| 68 | # This probably wipes out path so far. However, it's more |
| 69 | # complicated if path begins with a drive letter: |
| 70 | # 1. join('c:', '/a') == 'c:/a' |
| 71 | # 2. join('c:/', '/a') == 'c:/a' |
| 72 | # But |
| 73 | # 3. join('c:/a', '/b') == '/b' |
| 74 | # 4. join('c:', 'd:/') = 'd:/' |
| 75 | # 5. join('c:/', 'd:/') = 'd:/' |
| 76 | if path[1:2] != ":" or b[1:2] == ":": |
| 77 | # Path doesn't start with a drive letter, or cases 4 and 5. |
| 78 | b_wins = 1 |
| 79 | |
| 80 | # Else path has a drive letter, and b doesn't but is absolute. |
| 81 | elif len(path) > 3 or (len(path) == 3 and |
| 82 | path[-1] not in "/\\"): |
| 83 | # case 3 |
| 84 | b_wins = 1 |
| 85 | |
| 86 | if b_wins: |
| 87 | path = b |
| 88 | else: |
| 89 | # Join, and ensure there's a separator. |
| 90 | assert len(path) > 0 |
| 91 | if path[-1] in "/\\": |
| 92 | if b and b[0] in "/\\": |
| 93 | path += b[1:] |
| 94 | else: |
| 95 | path += b |
| 96 | elif path[-1] == ":": |
| 97 | path += b |
| 98 | elif b: |
| 99 | if b[0] in "/\\": |
| 100 | path += b |
| 101 | else: |
| 102 | path += "\\" + b |
| 103 | else: |
| 104 | # path is not empty and does not end with a backslash, |
| 105 | # but b is empty; since, e.g., split('a/') produces |
| 106 | # ('a', ''), it's best if join() adds a backslash in |
| 107 | # this case. |
| 108 | path += '\\' |
| 109 | |
| 110 | return path |
| 111 | |
| 112 | |
| 113 | # Split a path in a drive specification (a drive letter followed by a |
| 114 | # colon) and the path specification. |
| 115 | # It is always true that drivespec + pathspec == p |
| 116 | def splitdrive(p): |
| 117 | """Split a pathname into drive and path specifiers. Returns a 2-tuple |
| 118 | "(drive,path)"; either part may be empty""" |
| 119 | if p[1:2] == ':': |
| 120 | return p[0:2], p[2:] |
| 121 | return '', p |
| 122 | |
| 123 | |
| 124 | # Parse UNC paths |
| 125 | def splitunc(p): |
| 126 | """Split a pathname into UNC mount point and relative path specifiers. |
| 127 | |
| 128 | Return a 2-tuple (unc, rest); either part may be empty. |
| 129 | If unc is not empty, it has the form '//host/mount' (or similar |
| 130 | using backslashes). unc+rest is always the input path. |
| 131 | Paths containing drive letters never have an UNC part. |
| 132 | """ |
| 133 | if p[1:2] == ':': |
| 134 | return '', p # Drive letter present |
| 135 | firstTwo = p[0:2] |
| 136 | if firstTwo == '//' or firstTwo == '\\\\': |
| 137 | # is a UNC path: |
| 138 | # vvvvvvvvvvvvvvvvvvvv equivalent to drive letter |
| 139 | # \\machine\mountpoint\directories... |
| 140 | # directory ^^^^^^^^^^^^^^^ |
| 141 | normp = normcase(p) |
| 142 | index = normp.find('\\', 2) |
| 143 | if index == -1: |
| 144 | ##raise RuntimeError, 'illegal UNC path: "' + p + '"' |
| 145 | return ("", p) |
| 146 | index = normp.find('\\', index + 1) |
| 147 | if index == -1: |
| 148 | index = len(p) |
| 149 | return p[:index], p[index:] |
| 150 | return '', p |
| 151 | |
| 152 | |
| 153 | # Split a path in head (everything up to the last '/') and tail (the |
| 154 | # rest). After the trailing '/' is stripped, the invariant |
| 155 | # join(head, tail) == p holds. |
| 156 | # The resulting head won't end in '/' unless it is the root. |
| 157 | |
| 158 | def split(p): |
| 159 | """Split a pathname. |
| 160 | |
| 161 | Return tuple (head, tail) where tail is everything after the final slash. |
| 162 | Either part may be empty.""" |
| 163 | |
| 164 | d, p = splitdrive(p) |
| 165 | # set i to index beyond p's last slash |
| 166 | i = len(p) |
| 167 | while i and p[i-1] not in '/\\': |
| 168 | i = i - 1 |
| 169 | head, tail = p[:i], p[i:] # now tail has no slashes |
| 170 | # remove trailing slashes from head, unless it's all slashes |
| 171 | head2 = head |
| 172 | while head2 and head2[-1] in '/\\': |
| 173 | head2 = head2[:-1] |
| 174 | head = head2 or head |
| 175 | return d + head, tail |
| 176 | |
| 177 | |
| 178 | # Split a path in root and extension. |
| 179 | # The extension is everything starting at the last dot in the last |
| 180 | # pathname component; the root is everything before that. |
| 181 | # It is always true that root + ext == p. |
| 182 | |
| 183 | def splitext(p): |
| 184 | """Split the extension from a pathname. |
| 185 | |
| 186 | Extension is everything from the last dot to the end. |
| 187 | Return (root, ext), either part may be empty.""" |
| 188 | |
| 189 | i = p.rfind('.') |
| 190 | if i<=max(p.rfind('/'), p.rfind('\\')): |
| 191 | return p, '' |
| 192 | else: |
| 193 | return p[:i], p[i:] |
| 194 | |
| 195 | |
| 196 | # Return the tail (basename) part of a path. |
| 197 | |
| 198 | def basename(p): |
| 199 | """Returns the final component of a pathname""" |
| 200 | return split(p)[1] |
| 201 | |
| 202 | |
| 203 | # Return the head (dirname) part of a path. |
| 204 | |
| 205 | def dirname(p): |
| 206 | """Returns the directory component of a pathname""" |
| 207 | return split(p)[0] |
| 208 | |
| 209 | |
| 210 | # Return the longest prefix of all list elements. |
| 211 | |
| 212 | def commonprefix(m): |
| 213 | "Given a list of pathnames, returns the longest common leading component" |
| 214 | if not m: return '' |
| 215 | prefix = m[0] |
| 216 | for item in m: |
| 217 | for i in range(len(prefix)): |
| 218 | if prefix[:i+1] != item[:i+1]: |
| 219 | prefix = prefix[:i] |
| 220 | if i == 0: return '' |
| 221 | break |
| 222 | return prefix |
| 223 | |
| 224 | |
| 225 | # Get size, mtime, atime of files. |
| 226 | |
| 227 | def getsize(filename): |
| 228 | """Return the size of a file, reported by os.stat()""" |
| 229 | return os.stat(filename).st_size |
| 230 | |
| 231 | def getmtime(filename): |
| 232 | """Return the last modification time of a file, reported by os.stat()""" |
| 233 | return os.stat(filename).st_mtime |
| 234 | |
| 235 | def getatime(filename): |
| 236 | """Return the last access time of a file, reported by os.stat()""" |
| 237 | return os.stat(filename).st_atime |
| 238 | |
| 239 | def getctime(filename): |
| 240 | """Return the creation time of a file, reported by os.stat().""" |
| 241 | return os.stat(filename).st_ctime |
| 242 | |
| 243 | # Is a path a symbolic link? |
| 244 | # This will always return false on systems where posix.lstat doesn't exist. |
| 245 | |
| 246 | def islink(path): |
| 247 | """Test for symbolic link. On WindowsNT/95 always returns false""" |
| 248 | return False |
| 249 | |
| 250 | |
| 251 | # Does a path exist? |
| 252 | |
| 253 | def exists(path): |
| 254 | """Test whether a path exists""" |
| 255 | try: |
| 256 | st = os.stat(path) |
| 257 | except os.error: |
| 258 | return False |
| 259 | return True |
| 260 | |
| 261 | lexists = exists |
| 262 | |
| 263 | |
| 264 | # Is a path a dos directory? |
| 265 | # This follows symbolic links, so both islink() and isdir() can be true |
| 266 | # for the same path. |
| 267 | |
| 268 | def isdir(path): |
| 269 | """Test whether a path is a directory""" |
| 270 | try: |
| 271 | st = os.stat(path) |
| 272 | except os.error: |
| 273 | return False |
| 274 | return stat.S_ISDIR(st.st_mode) |
| 275 | |
| 276 | |
| 277 | # Is a path a regular file? |
| 278 | # This follows symbolic links, so both islink() and isdir() can be true |
| 279 | # for the same path. |
| 280 | |
| 281 | def isfile(path): |
| 282 | """Test whether a path is a regular file""" |
| 283 | try: |
| 284 | st = os.stat(path) |
| 285 | except os.error: |
| 286 | return False |
| 287 | return stat.S_ISREG(st.st_mode) |
| 288 | |
| 289 | |
| 290 | # Is a path a mount point? Either a root (with or without drive letter) |
| 291 | # or an UNC path with at most a / or \ after the mount point. |
| 292 | |
| 293 | def ismount(path): |
| 294 | """Test whether a path is a mount point (defined as root of drive)""" |
| 295 | unc, rest = splitunc(path) |
| 296 | if unc: |
| 297 | return rest in ("", "/", "\\") |
| 298 | p = splitdrive(path)[1] |
| 299 | return len(p) == 1 and p[0] in '/\\' |
| 300 | |
| 301 | |
| 302 | # Directory tree walk. |
| 303 | # For each directory under top (including top itself, but excluding |
| 304 | # '.' and '..'), func(arg, dirname, filenames) is called, where |
| 305 | # dirname is the name of the directory and filenames is the list |
| 306 | # of files (and subdirectories etc.) in the directory. |
| 307 | # The func may modify the filenames list, to implement a filter, |
| 308 | # or to impose a different order of visiting. |
| 309 | |
| 310 | def walk(top, func, arg): |
| 311 | """Directory tree walk with callback function. |
| 312 | |
| 313 | For each directory in the directory tree rooted at top (including top |
| 314 | itself, but excluding '.' and '..'), call func(arg, dirname, fnames). |
| 315 | dirname is the name of the directory, and fnames a list of the names of |
| 316 | the files and subdirectories in dirname (excluding '.' and '..'). func |
| 317 | may modify the fnames list in-place (e.g. via del or slice assignment), |
| 318 | and walk will only recurse into the subdirectories whose names remain in |
| 319 | fnames; this can be used to implement a filter, or to impose a specific |
| 320 | order of visiting. No semantics are defined for, or required of, arg, |
| 321 | beyond that arg is always passed to func. It can be used, e.g., to pass |
| 322 | a filename pattern, or a mutable object designed to accumulate |
| 323 | statistics. Passing None for arg is common.""" |
| 324 | |
| 325 | try: |
| 326 | names = os.listdir(top) |
| 327 | except os.error: |
| 328 | return |
| 329 | func(arg, top, names) |
| 330 | exceptions = ('.', '..') |
| 331 | for name in names: |
| 332 | if name not in exceptions: |
| 333 | name = join(top, name) |
| 334 | if isdir(name): |
| 335 | walk(name, func, arg) |
| 336 | |
| 337 | |
| 338 | # Expand paths beginning with '~' or '~user'. |
| 339 | # '~' means $HOME; '~user' means that user's home directory. |
| 340 | # If the path doesn't begin with '~', or if the user or $HOME is unknown, |
| 341 | # the path is returned unchanged (leaving error reporting to whatever |
| 342 | # function is called with the expanded path as argument). |
| 343 | # See also module 'glob' for expansion of *, ? and [...] in pathnames. |
| 344 | # (A function should also be defined to do full *sh-style environment |
| 345 | # variable expansion.) |
| 346 | |
| 347 | def expanduser(path): |
| 348 | """Expand ~ and ~user constructs. |
| 349 | |
| 350 | If user or $HOME is unknown, do nothing.""" |
| 351 | if path[:1] != '~': |
| 352 | return path |
| 353 | i, n = 1, len(path) |
| 354 | while i < n and path[i] not in '/\\': |
| 355 | i = i + 1 |
| 356 | if i == 1: |
| 357 | if 'HOME' in os.environ: |
| 358 | userhome = os.environ['HOME'] |
| 359 | elif not 'HOMEPATH' in os.environ: |
| 360 | return path |
| 361 | else: |
| 362 | try: |
| 363 | drive = os.environ['HOMEDRIVE'] |
| 364 | except KeyError: |
| 365 | drive = '' |
| 366 | userhome = join(drive, os.environ['HOMEPATH']) |
| 367 | else: |
| 368 | return path |
| 369 | return userhome + path[i:] |
| 370 | |
| 371 | |
| 372 | # Expand paths containing shell variable substitutions. |
| 373 | # The following rules apply: |
| 374 | # - no expansion within single quotes |
| 375 | # - no escape character, except for '$$' which is translated into '$' |
| 376 | # - ${varname} is accepted. |
| 377 | # - varnames can be made out of letters, digits and the character '_' |
| 378 | # XXX With COMMAND.COM you can use any characters in a variable name, |
| 379 | # XXX except '^|<>='. |
| 380 | |
| 381 | def expandvars(path): |
| 382 | """Expand shell variables of form $var and ${var}. |
| 383 | |
| 384 | Unknown variables are left unchanged.""" |
| 385 | if '$' not in path: |
| 386 | return path |
| 387 | import string |
| 388 | varchars = string.ascii_letters + string.digits + '_-' |
| 389 | res = '' |
| 390 | index = 0 |
| 391 | pathlen = len(path) |
| 392 | while index < pathlen: |
| 393 | c = path[index] |
| 394 | if c == '\'': # no expansion within single quotes |
| 395 | path = path[index + 1:] |
| 396 | pathlen = len(path) |
| 397 | try: |
| 398 | index = path.index('\'') |
| 399 | res = res + '\'' + path[:index + 1] |
| 400 | except ValueError: |
| 401 | res = res + path |
| 402 | index = pathlen - 1 |
| 403 | elif c == '$': # variable or '$$' |
| 404 | if path[index + 1:index + 2] == '$': |
| 405 | res = res + c |
| 406 | index = index + 1 |
| 407 | elif path[index + 1:index + 2] == '{': |
| 408 | path = path[index+2:] |
| 409 | pathlen = len(path) |
| 410 | try: |
| 411 | index = path.index('}') |
| 412 | var = path[:index] |
| 413 | if var in os.environ: |
| 414 | res = res + os.environ[var] |
| 415 | except ValueError: |
| 416 | res = res + path |
| 417 | index = pathlen - 1 |
| 418 | else: |
| 419 | var = '' |
| 420 | index = index + 1 |
| 421 | c = path[index:index + 1] |
| 422 | while c != '' and c in varchars: |
| 423 | var = var + c |
| 424 | index = index + 1 |
| 425 | c = path[index:index + 1] |
| 426 | if var in os.environ: |
| 427 | res = res + os.environ[var] |
| 428 | if c != '': |
| 429 | res = res + c |
| 430 | else: |
| 431 | res = res + c |
| 432 | index = index + 1 |
| 433 | return res |
| 434 | |
| 435 | |
| 436 | # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A\B. |
| 437 | # Previously, this function also truncated pathnames to 8+3 format, |
| 438 | # but as this module is called "ntpath", that's obviously wrong! |
| 439 | |
| 440 | def normpath(path): |
| 441 | """Normalize path, eliminating double slashes, etc.""" |
| 442 | path = path.replace("/", "\\") |
| 443 | prefix, path = splitdrive(path) |
| 444 | # We need to be careful here. If the prefix is empty, and the path starts |
| 445 | # with a backslash, it could either be an absolute path on the current |
| 446 | # drive (\dir1\dir2\file) or a UNC filename (\\server\mount\dir1\file). It |
| 447 | # is therefore imperative NOT to collapse multiple backslashes blindly in |
| 448 | # that case. |
| 449 | # The code below preserves multiple backslashes when there is no drive |
| 450 | # letter. This means that the invalid filename \\\a\b is preserved |
| 451 | # unchanged, where a\\\b is normalised to a\b. It's not clear that there |
| 452 | # is any better behaviour for such edge cases. |
| 453 | if prefix == '': |
| 454 | # No drive letter - preserve initial backslashes |
| 455 | while path[:1] == "\\": |
| 456 | prefix = prefix + "\\" |
| 457 | path = path[1:] |
| 458 | else: |
| 459 | # We have a drive letter - collapse initial backslashes |
| 460 | if path.startswith("\\"): |
| 461 | prefix = prefix + "\\" |
| 462 | path = path.lstrip("\\") |
| 463 | comps = path.split("\\") |
| 464 | i = 0 |
| 465 | while i < len(comps): |
| 466 | if comps[i] in ('.', ''): |
| 467 | del comps[i] |
| 468 | elif comps[i] == '..': |
| 469 | if i > 0 and comps[i-1] != '..': |
| 470 | del comps[i-1:i+1] |
| 471 | i -= 1 |
| 472 | elif i == 0 and prefix.endswith("\\"): |
| 473 | del comps[i] |
| 474 | else: |
| 475 | i += 1 |
| 476 | else: |
| 477 | i += 1 |
| 478 | # If the path is now empty, substitute '.' |
| 479 | if not prefix and not comps: |
| 480 | comps.append('.') |
| 481 | return prefix + "\\".join(comps) |
| 482 | |
| 483 | |
| 484 | # Return an absolute path. |
| 485 | def abspath(path): |
| 486 | """Return the absolute version of a path""" |
| 487 | try: |
| 488 | from nt import _getfullpathname |
| 489 | except ImportError: # Not running on Windows - mock up something sensible. |
| 490 | global abspath |
| 491 | def _abspath(path): |
| 492 | if not isabs(path): |
| 493 | path = join(os.getcwd(), path) |
| 494 | return normpath(path) |
| 495 | abspath = _abspath |
| 496 | return _abspath(path) |
| 497 | |
| 498 | if path: # Empty path must return current working directory. |
| 499 | try: |
| 500 | path = _getfullpathname(path) |
| 501 | except WindowsError: |
| 502 | pass # Bad path - return unchanged. |
| 503 | else: |
| 504 | path = os.getcwd() |
| 505 | return normpath(path) |
| 506 | |
| 507 | # realpath is a no-op on systems without islink support |
| 508 | realpath = abspath |
| 509 | # Win9x family and earlier have no Unicode filename support. |
| 510 | supports_unicode_filenames = (hasattr(sys, "getwindowsversion") and |
| 511 | sys.getwindowsversion()[3] >= 2) |