Guido van Rossum | e7b146f | 2000-02-04 15:28:42 +0000 | [diff] [blame] | 1 | """Utility functions for copying files and directory trees. |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 2 | |
Guido van Rossum | 959fa01 | 1999-08-18 20:03:17 +0000 | [diff] [blame] | 3 | XXX The functions here don't copy the resource fork or other metadata on Mac. |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 4 | |
| 5 | """ |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 6 | |
Guido van Rossum | c96207a | 1992-03-31 18:55:40 +0000 | [diff] [blame] | 7 | import os |
Guido van Rossum | 83c03e2 | 1999-02-23 23:07:51 +0000 | [diff] [blame] | 8 | import sys |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 9 | import stat |
Brett Cannon | 1c3fa18 | 2004-06-19 21:11:35 +0000 | [diff] [blame] | 10 | from os.path import abspath |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 11 | import fnmatch |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 12 | |
Skip Montanaro | 0de6580 | 2001-02-15 22:15:14 +0000 | [diff] [blame] | 13 | __all__ = ["copyfileobj","copyfile","copymode","copystat","copy","copy2", |
Raymond Hettinger | 2b9bfb3 | 2002-10-30 05:44:50 +0000 | [diff] [blame] | 14 | "copytree","move","rmtree","Error"] |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 15 | |
Neal Norwitz | 4ce69a5 | 2005-09-01 00:45:28 +0000 | [diff] [blame] | 16 | class Error(EnvironmentError): |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 17 | pass |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 18 | |
Georg Brandl | 6aa2d1f | 2008-08-12 08:35:52 +0000 | [diff] [blame] | 19 | try: |
| 20 | WindowsError |
| 21 | except NameError: |
| 22 | WindowsError = None |
| 23 | |
Greg Stein | 42bb8b3 | 2000-07-12 09:55:30 +0000 | [diff] [blame] | 24 | def copyfileobj(fsrc, fdst, length=16*1024): |
| 25 | """copy data from file-like object fsrc to file-like object fdst""" |
| 26 | while 1: |
| 27 | buf = fsrc.read(length) |
| 28 | if not buf: |
| 29 | break |
| 30 | fdst.write(buf) |
| 31 | |
Johannes Gijsbers | 46f1459 | 2004-08-14 13:30:02 +0000 | [diff] [blame] | 32 | def _samefile(src, dst): |
| 33 | # Macintosh, Unix. |
| 34 | if hasattr(os.path,'samefile'): |
Johannes Gijsbers | f9a098e | 2004-08-14 14:51:01 +0000 | [diff] [blame] | 35 | try: |
| 36 | return os.path.samefile(src, dst) |
| 37 | except OSError: |
| 38 | return False |
Johannes Gijsbers | 46f1459 | 2004-08-14 13:30:02 +0000 | [diff] [blame] | 39 | |
| 40 | # All other platforms: check for same pathname. |
| 41 | return (os.path.normcase(os.path.abspath(src)) == |
| 42 | os.path.normcase(os.path.abspath(dst))) |
Tim Peters | 495ad3c | 2001-01-15 01:36:40 +0000 | [diff] [blame] | 43 | |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 44 | def copyfile(src, dst): |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 45 | """Copy data from src to dst""" |
Johannes Gijsbers | 46f1459 | 2004-08-14 13:30:02 +0000 | [diff] [blame] | 46 | if _samefile(src, dst): |
Collin Winter | ce36ad8 | 2007-08-30 01:19:48 +0000 | [diff] [blame] | 47 | raise Error("`%s` and `%s` are the same file" % (src, dst)) |
Johannes Gijsbers | 46f1459 | 2004-08-14 13:30:02 +0000 | [diff] [blame] | 48 | |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 49 | fsrc = None |
| 50 | fdst = None |
| 51 | try: |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 52 | fsrc = open(src, 'rb') |
| 53 | fdst = open(dst, 'wb') |
Greg Stein | 42bb8b3 | 2000-07-12 09:55:30 +0000 | [diff] [blame] | 54 | copyfileobj(fsrc, fdst) |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 55 | finally: |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 56 | if fdst: |
| 57 | fdst.close() |
| 58 | if fsrc: |
| 59 | fsrc.close() |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 60 | |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 61 | def copymode(src, dst): |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 62 | """Copy mode bits from src to dst""" |
Tim Peters | 0c94724 | 2001-01-21 20:00:00 +0000 | [diff] [blame] | 63 | if hasattr(os, 'chmod'): |
| 64 | st = os.stat(src) |
Walter Dörwald | 294bbf3 | 2002-06-06 09:48:13 +0000 | [diff] [blame] | 65 | mode = stat.S_IMODE(st.st_mode) |
Tim Peters | 0c94724 | 2001-01-21 20:00:00 +0000 | [diff] [blame] | 66 | os.chmod(dst, mode) |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 67 | |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 68 | def copystat(src, dst): |
Thomas Wouters | cf297e4 | 2007-02-23 15:07:44 +0000 | [diff] [blame] | 69 | """Copy all stat info (mode bits, atime, mtime, flags) from src to dst""" |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 70 | st = os.stat(src) |
Walter Dörwald | 294bbf3 | 2002-06-06 09:48:13 +0000 | [diff] [blame] | 71 | mode = stat.S_IMODE(st.st_mode) |
Tim Peters | 0c94724 | 2001-01-21 20:00:00 +0000 | [diff] [blame] | 72 | if hasattr(os, 'utime'): |
Walter Dörwald | 294bbf3 | 2002-06-06 09:48:13 +0000 | [diff] [blame] | 73 | os.utime(dst, (st.st_atime, st.st_mtime)) |
Tim Peters | 0c94724 | 2001-01-21 20:00:00 +0000 | [diff] [blame] | 74 | if hasattr(os, 'chmod'): |
| 75 | os.chmod(dst, mode) |
Thomas Wouters | cf297e4 | 2007-02-23 15:07:44 +0000 | [diff] [blame] | 76 | if hasattr(os, 'chflags') and hasattr(st, 'st_flags'): |
| 77 | os.chflags(dst, st.st_flags) |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 78 | |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 79 | |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 80 | def copy(src, dst): |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 81 | """Copy data and mode bits ("cp src dst"). |
Tim Peters | 495ad3c | 2001-01-15 01:36:40 +0000 | [diff] [blame] | 82 | |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 83 | The destination may be a directory. |
| 84 | |
| 85 | """ |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 86 | if os.path.isdir(dst): |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 87 | dst = os.path.join(dst, os.path.basename(src)) |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 88 | copyfile(src, dst) |
| 89 | copymode(src, dst) |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 90 | |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 91 | def copy2(src, dst): |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 92 | """Copy data and all stat info ("cp -p src dst"). |
| 93 | |
| 94 | The destination may be a directory. |
| 95 | |
| 96 | """ |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 97 | if os.path.isdir(dst): |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 98 | dst = os.path.join(dst, os.path.basename(src)) |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 99 | copyfile(src, dst) |
| 100 | copystat(src, dst) |
Guido van Rossum | c636014 | 1990-10-13 19:23:40 +0000 | [diff] [blame] | 101 | |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 102 | def ignore_patterns(*patterns): |
| 103 | """Function that can be used as copytree() ignore parameter. |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 104 | |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 105 | Patterns is a sequence of glob-style patterns |
| 106 | that are used to exclude files""" |
| 107 | def _ignore_patterns(path, names): |
| 108 | ignored_names = [] |
| 109 | for pattern in patterns: |
| 110 | ignored_names.extend(fnmatch.filter(names, pattern)) |
| 111 | return set(ignored_names) |
| 112 | return _ignore_patterns |
| 113 | |
| 114 | def copytree(src, dst, symlinks=False, ignore=None): |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 115 | """Recursively copy a directory tree using copy2(). |
| 116 | |
| 117 | The destination directory must not already exist. |
Neal Norwitz | a4c93b6 | 2003-02-23 21:36:32 +0000 | [diff] [blame] | 118 | If exception(s) occur, an Error is raised with a list of reasons. |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 119 | |
| 120 | If the optional symlinks flag is true, symbolic links in the |
| 121 | source tree result in symbolic links in the destination tree; if |
| 122 | it is false, the contents of the files pointed to by symbolic |
| 123 | links are copied. |
| 124 | |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 125 | The optional ignore argument is a callable. If given, it |
| 126 | is called with the `src` parameter, which is the directory |
| 127 | being visited by copytree(), and `names` which is the list of |
| 128 | `src` contents, as returned by os.listdir(): |
| 129 | |
| 130 | callable(src, names) -> ignored_names |
| 131 | |
| 132 | Since copytree() is called recursively, the callable will be |
| 133 | called once for each directory that is copied. It returns a |
| 134 | list of names relative to the `src` directory that should |
| 135 | not be copied. |
| 136 | |
Guido van Rossum | 9d0a3df | 1997-04-29 14:45:19 +0000 | [diff] [blame] | 137 | XXX Consider this example code rather than the ultimate tool. |
| 138 | |
| 139 | """ |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 140 | names = os.listdir(src) |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 141 | if ignore is not None: |
| 142 | ignored_names = ignore(src, names) |
| 143 | else: |
| 144 | ignored_names = set() |
| 145 | |
Johannes Gijsbers | e4172ea | 2005-01-08 12:31:29 +0000 | [diff] [blame] | 146 | os.makedirs(dst) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 147 | errors = [] |
Guido van Rossum | a2baf46 | 1997-04-29 14:06:46 +0000 | [diff] [blame] | 148 | for name in names: |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 149 | if name in ignored_names: |
| 150 | continue |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 151 | srcname = os.path.join(src, name) |
| 152 | dstname = os.path.join(dst, name) |
| 153 | try: |
| 154 | if symlinks and os.path.islink(srcname): |
| 155 | linkto = os.readlink(srcname) |
| 156 | os.symlink(linkto, dstname) |
| 157 | elif os.path.isdir(srcname): |
Georg Brandl | 2ee470f | 2008-07-16 12:55:28 +0000 | [diff] [blame] | 158 | copytree(srcname, dstname, symlinks, ignore) |
Guido van Rossum | 45e2fbc | 1998-03-26 21:13:24 +0000 | [diff] [blame] | 159 | else: |
| 160 | copy2(srcname, dstname) |
| 161 | # XXX What about devices, sockets etc.? |
Guido van Rossum | b940e11 | 2007-01-10 16:19:56 +0000 | [diff] [blame] | 162 | except (IOError, os.error) as why: |
Thomas Wouters | 477c8d5 | 2006-05-27 19:21:47 +0000 | [diff] [blame] | 163 | errors.append((srcname, dstname, str(why))) |
Georg Brandl | a1be88e | 2005-08-31 22:48:45 +0000 | [diff] [blame] | 164 | # catch the Error from the recursive copytree so that we can |
| 165 | # continue with other files |
Guido van Rossum | b940e11 | 2007-01-10 16:19:56 +0000 | [diff] [blame] | 166 | except Error as err: |
Georg Brandl | a1be88e | 2005-08-31 22:48:45 +0000 | [diff] [blame] | 167 | errors.extend(err.args[0]) |
Thomas Wouters | 0e3f591 | 2006-08-11 14:57:12 +0000 | [diff] [blame] | 168 | try: |
| 169 | copystat(src, dst) |
Guido van Rossum | b940e11 | 2007-01-10 16:19:56 +0000 | [diff] [blame] | 170 | except OSError as why: |
Georg Brandl | 6aa2d1f | 2008-08-12 08:35:52 +0000 | [diff] [blame] | 171 | if WindowsError is not None and isinstance(why, WindowsError): |
| 172 | # Copying file access times may fail on Windows |
| 173 | pass |
| 174 | else: |
| 175 | errors.extend((src, dst, str(why))) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 176 | if errors: |
Collin Winter | ce36ad8 | 2007-08-30 01:19:48 +0000 | [diff] [blame] | 177 | raise Error(errors) |
Guido van Rossum | d767329 | 1998-02-06 21:38:09 +0000 | [diff] [blame] | 178 | |
Barry Warsaw | 234d9a9 | 2003-01-24 17:36:15 +0000 | [diff] [blame] | 179 | def rmtree(path, ignore_errors=False, onerror=None): |
Guido van Rossum | d767329 | 1998-02-06 21:38:09 +0000 | [diff] [blame] | 180 | """Recursively delete a directory tree. |
| 181 | |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 182 | If ignore_errors is set, errors are ignored; otherwise, if onerror |
| 183 | is set, it is called to handle the error with arguments (func, |
| 184 | path, exc_info) where func is os.listdir, os.remove, or os.rmdir; |
| 185 | path is the argument to that function that caused it to fail; and |
| 186 | exc_info is a tuple returned by sys.exc_info(). If ignore_errors |
| 187 | is false and onerror is None, an exception is raised. |
| 188 | |
Guido van Rossum | d767329 | 1998-02-06 21:38:09 +0000 | [diff] [blame] | 189 | """ |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 190 | if ignore_errors: |
| 191 | def onerror(*args): |
Barry Warsaw | 234d9a9 | 2003-01-24 17:36:15 +0000 | [diff] [blame] | 192 | pass |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 193 | elif onerror is None: |
| 194 | def onerror(*args): |
| 195 | raise |
Christian Heimes | 9bd667a | 2008-01-20 15:14:11 +0000 | [diff] [blame] | 196 | try: |
| 197 | if os.path.islink(path): |
| 198 | # symlinks to directories are forbidden, see bug #1669 |
| 199 | raise OSError("Cannot call rmtree on a symbolic link") |
| 200 | except OSError: |
| 201 | onerror(os.path.islink, path, sys.exc_info()) |
| 202 | # can't continue even if onerror hook returns |
| 203 | return |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 204 | names = [] |
| 205 | try: |
| 206 | names = os.listdir(path) |
Guido van Rossum | b940e11 | 2007-01-10 16:19:56 +0000 | [diff] [blame] | 207 | except os.error as err: |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 208 | onerror(os.listdir, path, sys.exc_info()) |
| 209 | for name in names: |
| 210 | fullname = os.path.join(path, name) |
| 211 | try: |
| 212 | mode = os.lstat(fullname).st_mode |
| 213 | except os.error: |
| 214 | mode = 0 |
| 215 | if stat.S_ISDIR(mode): |
| 216 | rmtree(fullname, ignore_errors, onerror) |
Barry Warsaw | 234d9a9 | 2003-01-24 17:36:15 +0000 | [diff] [blame] | 217 | else: |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 218 | try: |
| 219 | os.remove(fullname) |
Guido van Rossum | b940e11 | 2007-01-10 16:19:56 +0000 | [diff] [blame] | 220 | except os.error as err: |
Johannes Gijsbers | ef5ffc4 | 2004-10-31 12:05:31 +0000 | [diff] [blame] | 221 | onerror(os.remove, fullname, sys.exc_info()) |
| 222 | try: |
| 223 | os.rmdir(path) |
| 224 | except os.error: |
| 225 | onerror(os.rmdir, path, sys.exc_info()) |
Guido van Rossum | d767329 | 1998-02-06 21:38:09 +0000 | [diff] [blame] | 226 | |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 227 | |
Christian Heimes | ada8c3b | 2008-03-18 18:26:33 +0000 | [diff] [blame] | 228 | def _basename(path): |
| 229 | # A basename() variant which first strips the trailing slash, if present. |
| 230 | # Thus we always get the last component of the path, even for directories. |
| 231 | return os.path.basename(path.rstrip(os.path.sep)) |
| 232 | |
| 233 | def move(src, dst): |
| 234 | """Recursively move a file or directory to another location. This is |
| 235 | similar to the Unix "mv" command. |
| 236 | |
| 237 | If the destination is a directory or a symlink to a directory, the source |
| 238 | is moved inside the directory. The destination path must not already |
| 239 | exist. |
| 240 | |
| 241 | If the destination already exists but is not a directory, it may be |
| 242 | overwritten depending on os.rename() semantics. |
| 243 | |
| 244 | If the destination is on our current filesystem, then rename() is used. |
| 245 | Otherwise, src is copied to the destination and then removed. |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 246 | A lot more could be done here... A look at a mv.c shows a lot of |
| 247 | the issues this implementation glosses over. |
| 248 | |
| 249 | """ |
Christian Heimes | ada8c3b | 2008-03-18 18:26:33 +0000 | [diff] [blame] | 250 | real_dst = dst |
| 251 | if os.path.isdir(dst): |
| 252 | real_dst = os.path.join(dst, _basename(src)) |
| 253 | if os.path.exists(real_dst): |
| 254 | raise Error("Destination path '%s' already exists" % real_dst) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 255 | try: |
Christian Heimes | ada8c3b | 2008-03-18 18:26:33 +0000 | [diff] [blame] | 256 | os.rename(src, real_dst) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 257 | except OSError: |
| 258 | if os.path.isdir(src): |
Benjamin Peterson | 247a9b8 | 2009-02-20 04:09:19 +0000 | [diff] [blame] | 259 | if _destinsrc(src, dst): |
Collin Winter | ce36ad8 | 2007-08-30 01:19:48 +0000 | [diff] [blame] | 260 | raise Error("Cannot move a directory '%s' into itself '%s'." % (src, dst)) |
Christian Heimes | ada8c3b | 2008-03-18 18:26:33 +0000 | [diff] [blame] | 261 | copytree(src, real_dst, symlinks=True) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 262 | rmtree(src) |
| 263 | else: |
Christian Heimes | ada8c3b | 2008-03-18 18:26:33 +0000 | [diff] [blame] | 264 | copy2(src, real_dst) |
Martin v. Löwis | e9ce0b0 | 2002-10-07 13:23:24 +0000 | [diff] [blame] | 265 | os.unlink(src) |
Brett Cannon | 1c3fa18 | 2004-06-19 21:11:35 +0000 | [diff] [blame] | 266 | |
Benjamin Peterson | 247a9b8 | 2009-02-20 04:09:19 +0000 | [diff] [blame] | 267 | def _destinsrc(src, dst): |
Antoine Pitrou | 0dcc3cd | 2009-01-29 20:26:59 +0000 | [diff] [blame] | 268 | src = abspath(src) |
| 269 | dst = abspath(dst) |
| 270 | if not src.endswith(os.path.sep): |
| 271 | src += os.path.sep |
| 272 | if not dst.endswith(os.path.sep): |
| 273 | dst += os.path.sep |
| 274 | return dst.startswith(src) |