Guido van Rossum | 4b8c6ea | 2000-02-04 15:39:30 +0000 | [diff] [blame] | 1 | """Common operations on DOS pathnames.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 2 | |
| 3 | import os |
| 4 | import stat |
| 5 | import string |
Skip Montanaro | 802bc5d | 2000-07-17 03:06:26 +0000 | [diff] [blame^] | 6 | import re |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 7 | |
| 8 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 9 | def normcase(s): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 10 | """Normalize the case of a pathname. |
| 11 | On MS-DOS it maps the pathname to lowercase, turns slashes into |
| 12 | backslashes. |
| 13 | Other normalizations (such as optimizing '../' away) are not allowed |
| 14 | (this is done by normpath). |
| 15 | Previously, this version mapped invalid consecutive characters to a |
| 16 | single '_', but this has been removed. This functionality should |
| 17 | possibly be added as a new function.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 18 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 19 | return string.lower(string.replace(s, "/", "\\")) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 20 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 21 | |
| 22 | def isabs(s): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 23 | """Return whether a path is absolute. |
| 24 | Trivial in Posix, harder on the Mac or MS-DOS. |
| 25 | For DOS it is absolute if it starts with a slash or backslash (current |
| 26 | volume), or if a pathname after the volume letter and colon starts with |
| 27 | a slash or backslash.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 28 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 29 | s = splitdrive(s)[1] |
| 30 | return s != '' and s[:1] in '/\\' |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 31 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 32 | |
Guido van Rossum | ae590db | 1997-10-07 14:48:23 +0000 | [diff] [blame] | 33 | def join(a, *p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 34 | """Join two (or more) paths.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 35 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 36 | path = a |
| 37 | for b in p: |
| 38 | if isabs(b): |
| 39 | path = b |
| 40 | elif path == '' or path[-1:] in '/\\': |
| 41 | path = path + b |
| 42 | else: |
| 43 | path = path + os.sep + b |
| 44 | return path |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 45 | |
Guido van Rossum | fda5c1a | 1995-08-10 19:27:42 +0000 | [diff] [blame] | 46 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 47 | def splitdrive(p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 48 | """Split a path into a drive specification (a drive letter followed |
| 49 | by a colon) and path specification. |
| 50 | It is always true that drivespec + pathspec == p.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 51 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 52 | if p[1:2] == ':': |
| 53 | return p[0:2], p[2:] |
| 54 | return '', p |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 55 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 56 | |
| 57 | def split(p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 58 | """Split a path into head (everything up to the last '/') and tail |
| 59 | (the rest). After the trailing '/' is stripped, the invariant |
| 60 | join(head, tail) == p holds. |
| 61 | The resulting head won't end in '/' unless it is the root.""" |
| 62 | |
Guido van Rossum | 8f0fa9e | 1999-03-19 21:05:12 +0000 | [diff] [blame] | 63 | d, p = splitdrive(p) |
| 64 | # set i to index beyond p's last slash |
| 65 | i = len(p) |
| 66 | while i and p[i-1] not in '/\\': |
| 67 | i = i - 1 |
| 68 | head, tail = p[:i], p[i:] # now tail has no slashes |
| 69 | # remove trailing slashes from head, unless it's all slashes |
| 70 | head2 = head |
| 71 | while head2 and head2[-1] in '/\\': |
| 72 | head2 = head2[:-1] |
| 73 | head = head2 or head |
| 74 | return d + head, tail |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 75 | |
| 76 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 77 | def splitext(p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 78 | """Split a path into root and extension. |
| 79 | The extension is everything starting at the first dot in the last |
| 80 | pathname component; the root is everything before that. |
| 81 | It is always true that root + ext == p.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 82 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 83 | root, ext = '', '' |
| 84 | for c in p: |
| 85 | if c in '/\\': |
| 86 | root, ext = root + ext + c, '' |
| 87 | elif c == '.' or ext: |
| 88 | ext = ext + c |
| 89 | else: |
| 90 | root = root + c |
| 91 | return root, ext |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 92 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 93 | |
| 94 | def basename(p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 95 | """Return the tail (basename) part of a path.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 96 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 97 | return split(p)[1] |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 98 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 99 | |
| 100 | def dirname(p): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 101 | """Return the head (dirname) part of a path.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 102 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 103 | return split(p)[0] |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 104 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 105 | |
Skip Montanaro | 97bc98a | 2000-07-12 16:55:57 +0000 | [diff] [blame] | 106 | # Return the longest prefix of all list elements. |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 107 | |
Skip Montanaro | 97bc98a | 2000-07-12 16:55:57 +0000 | [diff] [blame] | 108 | def commonprefix(m): |
| 109 | "Given a list of pathnames, returns the longest common leading component" |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 110 | if not m: return '' |
Skip Montanaro | 6222c05 | 2000-07-17 03:04:19 +0000 | [diff] [blame] | 111 | n = map(string.lower, m) |
Skip Montanaro | 97bc98a | 2000-07-12 16:55:57 +0000 | [diff] [blame] | 112 | for i in range(len(n)): |
Skip Montanaro | 6222c05 | 2000-07-17 03:04:19 +0000 | [diff] [blame] | 113 | n[i] = re.split(r"[/\\]", n[i]) |
Skip Montanaro | 97bc98a | 2000-07-12 16:55:57 +0000 | [diff] [blame] | 114 | |
| 115 | prefix = n[0] |
| 116 | for item in n: |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 117 | for i in range(len(prefix)): |
| 118 | if prefix[:i+1] <> item[:i+1]: |
| 119 | prefix = prefix[:i] |
| 120 | if i == 0: return '' |
| 121 | break |
Skip Montanaro | 6222c05 | 2000-07-17 03:04:19 +0000 | [diff] [blame] | 122 | return "\\".join(prefix) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 123 | |
| 124 | |
Guido van Rossum | 2bc1f8f | 1998-07-24 20:49:26 +0000 | [diff] [blame] | 125 | # Get size, mtime, atime of files. |
| 126 | |
| 127 | def getsize(filename): |
| 128 | """Return the size of a file, reported by os.stat().""" |
| 129 | st = os.stat(filename) |
| 130 | return st[stat.ST_SIZE] |
| 131 | |
| 132 | def getmtime(filename): |
| 133 | """Return the last modification time of a file, reported by os.stat().""" |
| 134 | st = os.stat(filename) |
| 135 | return st[stat.ST_MTIME] |
| 136 | |
| 137 | def getatime(filename): |
| 138 | """Return the last access time of a file, reported by os.stat().""" |
| 139 | st = os.stat(filename) |
Guido van Rossum | 46d5651 | 2000-07-01 10:52:49 +0000 | [diff] [blame] | 140 | return st[stat.ST_ATIME] |
Guido van Rossum | 2bc1f8f | 1998-07-24 20:49:26 +0000 | [diff] [blame] | 141 | |
| 142 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 143 | def islink(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 144 | """Is a path a symbolic link? |
| 145 | This will always return false on systems where posix.lstat doesn't exist.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 146 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 147 | return 0 |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 148 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 149 | |
| 150 | def exists(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 151 | """Does a path exist? |
| 152 | This is false for dangling symbolic links.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 153 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 154 | try: |
| 155 | st = os.stat(path) |
| 156 | except os.error: |
| 157 | return 0 |
| 158 | return 1 |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 159 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 160 | |
| 161 | def isdir(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 162 | """Is a path a dos directory?""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 163 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 164 | try: |
| 165 | st = os.stat(path) |
| 166 | except os.error: |
| 167 | return 0 |
| 168 | return stat.S_ISDIR(st[stat.ST_MODE]) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 169 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 170 | |
| 171 | def isfile(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 172 | """Is a path a regular file?""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 173 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 174 | try: |
| 175 | st = os.stat(path) |
| 176 | except os.error: |
| 177 | return 0 |
| 178 | return stat.S_ISREG(st[stat.ST_MODE]) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 179 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 180 | |
| 181 | def ismount(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 182 | """Is a path a mount point?""" |
| 183 | # XXX This degenerates in: 'is this the root?' on DOS |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 184 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 185 | return isabs(splitdrive(path)[1]) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 186 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 187 | |
| 188 | def walk(top, func, arg): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 189 | """Directory tree walk. |
| 190 | For each directory under top (including top itself, but excluding |
| 191 | '.' and '..'), func(arg, dirname, filenames) is called, where |
| 192 | dirname is the name of the directory and filenames is the list |
| 193 | files files (and subdirectories etc.) in the directory. |
| 194 | The func may modify the filenames list, to implement a filter, |
| 195 | or to impose a different order of visiting.""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 196 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 197 | try: |
| 198 | names = os.listdir(top) |
| 199 | except os.error: |
| 200 | return |
| 201 | func(arg, top, names) |
| 202 | exceptions = ('.', '..') |
| 203 | for name in names: |
| 204 | if name not in exceptions: |
| 205 | name = join(top, name) |
| 206 | if isdir(name): |
| 207 | walk(name, func, arg) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 208 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 209 | |
| 210 | def expanduser(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 211 | """Expand paths beginning with '~' or '~user'. |
| 212 | '~' means $HOME; '~user' means that user's home directory. |
| 213 | If the path doesn't begin with '~', or if the user or $HOME is unknown, |
| 214 | the path is returned unchanged (leaving error reporting to whatever |
| 215 | function is called with the expanded path as argument). |
| 216 | See also module 'glob' for expansion of *, ? and [...] in pathnames. |
| 217 | (A function should also be defined to do full *sh-style environment |
| 218 | variable expansion.)""" |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 219 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 220 | if path[:1] <> '~': |
| 221 | return path |
| 222 | i, n = 1, len(path) |
| 223 | while i < n and path[i] not in '/\\': |
| 224 | i = i+1 |
| 225 | if i == 1: |
| 226 | if not os.environ.has_key('HOME'): |
| 227 | return path |
| 228 | userhome = os.environ['HOME'] |
| 229 | else: |
| 230 | return path |
| 231 | return userhome + path[i:] |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 232 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 233 | |
| 234 | varchars = string.letters + string.digits + '_-' |
| 235 | |
| 236 | def expandvars(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 237 | """Expand paths containing shell variable substitutions. |
| 238 | The following rules apply: |
| 239 | - no expansion within single quotes |
| 240 | - no escape character, except for '$$' which is translated into '$' |
| 241 | - ${varname} is accepted. |
| 242 | - varnames can be made out of letters, digits and the character '_'""" |
| 243 | # XXX With COMMAND.COM you can use any characters in a variable name, |
| 244 | # XXX except '^|<>='. |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 245 | |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 246 | if '$' not in path: |
| 247 | return path |
| 248 | res = '' |
| 249 | index = 0 |
| 250 | pathlen = len(path) |
| 251 | while index < pathlen: |
| 252 | c = path[index] |
| 253 | if c == '\'': # no expansion within single quotes |
| 254 | path = path[index + 1:] |
| 255 | pathlen = len(path) |
| 256 | try: |
| 257 | index = string.index(path, '\'') |
| 258 | res = res + '\'' + path[:index + 1] |
| 259 | except string.index_error: |
| 260 | res = res + path |
| 261 | index = pathlen -1 |
| 262 | elif c == '$': # variable or '$$' |
| 263 | if path[index + 1:index + 2] == '$': |
| 264 | res = res + c |
| 265 | index = index + 1 |
| 266 | elif path[index + 1:index + 2] == '{': |
| 267 | path = path[index+2:] |
| 268 | pathlen = len(path) |
| 269 | try: |
| 270 | index = string.index(path, '}') |
| 271 | var = path[:index] |
| 272 | if os.environ.has_key(var): |
| 273 | res = res + os.environ[var] |
| 274 | except string.index_error: |
| 275 | res = res + path |
| 276 | index = pathlen - 1 |
| 277 | else: |
| 278 | var = '' |
| 279 | index = index + 1 |
| 280 | c = path[index:index + 1] |
| 281 | while c != '' and c in varchars: |
| 282 | var = var + c |
| 283 | index = index + 1 |
| 284 | c = path[index:index + 1] |
| 285 | if os.environ.has_key(var): |
| 286 | res = res + os.environ[var] |
| 287 | if c != '': |
| 288 | res = res + c |
| 289 | else: |
| 290 | res = res + c |
| 291 | index = index + 1 |
| 292 | return res |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 293 | |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 294 | |
| 295 | def normpath(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 296 | """Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B. |
| 297 | Also, components of the path are silently truncated to 8+3 notation.""" |
| 298 | |
| 299 | path = string.replace(path, "/", "\\") |
| 300 | prefix, path = splitdrive(path) |
| 301 | while path[:1] == os.sep: |
| 302 | prefix = prefix + os.sep |
| 303 | path = path[1:] |
| 304 | comps = string.splitfields(path, os.sep) |
| 305 | i = 0 |
| 306 | while i < len(comps): |
| 307 | if comps[i] == '.': |
| 308 | del comps[i] |
| 309 | elif comps[i] == '..' and i > 0 and \ |
| 310 | comps[i-1] not in ('', '..'): |
| 311 | del comps[i-1:i+1] |
| 312 | i = i-1 |
| 313 | elif comps[i] == '' and i > 0 and comps[i-1] <> '': |
| 314 | del comps[i] |
| 315 | elif '.' in comps[i]: |
| 316 | comp = string.splitfields(comps[i], '.') |
| 317 | comps[i] = comp[0][:8] + '.' + comp[1][:3] |
| 318 | i = i+1 |
| 319 | elif len(comps[i]) > 8: |
| 320 | comps[i] = comps[i][:8] |
| 321 | i = i+1 |
| 322 | else: |
| 323 | i = i+1 |
| 324 | # If the path is now empty, substitute '.' |
| 325 | if not prefix and not comps: |
| 326 | comps.append('.') |
| 327 | return prefix + string.joinfields(comps, os.sep) |
Guido van Rossum | 3ed23cc | 1994-02-15 15:57:15 +0000 | [diff] [blame] | 328 | |
Guido van Rossum | e294cf6 | 1999-01-29 18:05:18 +0000 | [diff] [blame] | 329 | |
| 330 | |
Guido van Rossum | e294cf6 | 1999-01-29 18:05:18 +0000 | [diff] [blame] | 331 | def abspath(path): |
Guido van Rossum | 54f22ed | 2000-02-04 15:10:34 +0000 | [diff] [blame] | 332 | """Return an absolute path.""" |
Guido van Rossum | e294cf6 | 1999-01-29 18:05:18 +0000 | [diff] [blame] | 333 | if not isabs(path): |
| 334 | path = join(os.getcwd(), path) |
| 335 | return normpath(path) |