| 1 | =head1 NAME |
| 2 | |
| 3 | File::Basename - Parse file paths into directory, filename and suffix. |
| 4 | |
| 5 | =head1 SYNOPSIS |
| 6 | |
| 7 | use File::Basename; |
| 8 | |
| 9 | ($name,$path,$suffix) = fileparse($fullname,@suffixlist); |
| 10 | $name = fileparse($fullname,@suffixlist); |
| 11 | |
| 12 | $basename = basename($fullname,@suffixlist); |
| 13 | $dirname = dirname($fullname); |
| 14 | |
| 15 | |
| 16 | =head1 DESCRIPTION |
| 17 | |
| 18 | These routines allow you to parse file paths into their directory, filename |
| 19 | and suffix. |
| 20 | |
| 21 | B<NOTE>: C<dirname()> and C<basename()> emulate the behaviours, and |
| 22 | quirks, of the shell and C functions of the same name. See each |
| 23 | function's documentation for details. If your concern is just parsing |
| 24 | paths it is safer to use L<File::Spec>'s C<splitpath()> and |
| 25 | C<splitdir()> methods. |
| 26 | |
| 27 | It is guaranteed that |
| 28 | |
| 29 | # Where $path_separator is / for Unix, \ for Windows, etc... |
| 30 | dirname($path) . $path_separator . basename($path); |
| 31 | |
| 32 | is equivalent to the original path for all systems but VMS. |
| 33 | |
| 34 | |
| 35 | =cut |
| 36 | |
| 37 | |
| 38 | package File::Basename; |
| 39 | |
| 40 | # A bit of juggling to insure that C<use re 'taint';> always works, since |
| 41 | # File::Basename is used during the Perl build, when the re extension may |
| 42 | # not be available. |
| 43 | BEGIN { |
| 44 | unless (eval { require re; }) |
| 45 | { eval ' sub re::import { $^H |= 0x00100000; } ' } # HINT_RE_TAINT |
| 46 | import re 'taint'; |
| 47 | } |
| 48 | |
| 49 | |
| 50 | use strict; |
| 51 | use 5.006; |
| 52 | use warnings; |
| 53 | our(@ISA, @EXPORT, $VERSION, $Fileparse_fstype, $Fileparse_igncase); |
| 54 | require Exporter; |
| 55 | @ISA = qw(Exporter); |
| 56 | @EXPORT = qw(fileparse fileparse_set_fstype basename dirname); |
| 57 | $VERSION = "2.74"; |
| 58 | |
| 59 | fileparse_set_fstype($^O); |
| 60 | |
| 61 | |
| 62 | =over 4 |
| 63 | |
| 64 | =item C<fileparse> |
| 65 | |
| 66 | my($filename, $directories, $suffix) = fileparse($path); |
| 67 | my($filename, $directories, $suffix) = fileparse($path, @suffixes); |
| 68 | my $filename = fileparse($path, @suffixes); |
| 69 | |
| 70 | The C<fileparse()> routine divides a file path into its $directories, $filename |
| 71 | and (optionally) the filename $suffix. |
| 72 | |
| 73 | $directories contains everything up to and including the last |
| 74 | directory separator in the $path including the volume (if applicable). |
| 75 | The remainder of the $path is the $filename. |
| 76 | |
| 77 | # On Unix returns ("baz", "/foo/bar/", "") |
| 78 | fileparse("/foo/bar/baz"); |
| 79 | |
| 80 | # On Windows returns ("baz", "C:\foo\bar\", "") |
| 81 | fileparse("C:\foo\bar\baz"); |
| 82 | |
| 83 | # On Unix returns ("", "/foo/bar/baz/", "") |
| 84 | fileparse("/foo/bar/baz/"); |
| 85 | |
| 86 | If @suffixes are given each element is a pattern (either a string or a |
| 87 | C<qr//>) matched against the end of the $filename. The matching |
| 88 | portion is removed and becomes the $suffix. |
| 89 | |
| 90 | # On Unix returns ("baz", "/foo/bar", ".txt") |
| 91 | fileparse("/foo/bar/baz", qr/\.[^.]*/); |
| 92 | |
| 93 | If type is non-Unix (see C<fileparse_set_fstype()>) then the pattern |
| 94 | matching for suffix removal is performed case-insensitively, since |
| 95 | those systems are not case-sensitive when opening existing files. |
| 96 | |
| 97 | You are guaranteed that C<$directories . $filename . $suffix> will |
| 98 | denote the same location as the original $path. |
| 99 | |
| 100 | =cut |
| 101 | |
| 102 | |
| 103 | sub fileparse { |
| 104 | my($fullname,@suffices) = @_; |
| 105 | |
| 106 | unless (defined $fullname) { |
| 107 | require Carp; |
| 108 | Carp::croak("fileparse(): need a valid pathname"); |
| 109 | } |
| 110 | |
| 111 | my $orig_type = ''; |
| 112 | my($type,$igncase) = ($Fileparse_fstype, $Fileparse_igncase); |
| 113 | |
| 114 | my($taint) = substr($fullname,0,0); # Is $fullname tainted? |
| 115 | |
| 116 | if ($type eq "VMS" and $fullname =~ m{/} ) { |
| 117 | # We're doing Unix emulation |
| 118 | $orig_type = $type; |
| 119 | $type = 'Unix'; |
| 120 | } |
| 121 | |
| 122 | my($dirpath, $basename); |
| 123 | |
| 124 | if (grep { $type eq $_ } qw(MSDOS DOS MSWin32 Epoc)) { |
| 125 | ($dirpath,$basename) = ($fullname =~ /^((?:.*[:\\\/])?)(.*)/s); |
| 126 | $dirpath .= '.\\' unless $dirpath =~ /[\\\/]\z/; |
| 127 | } |
| 128 | elsif ($type eq "OS2") { |
| 129 | ($dirpath,$basename) = ($fullname =~ m#^((?:.*[:\\/])?)(.*)#s); |
| 130 | $dirpath = './' unless $dirpath; # Can't be 0 |
| 131 | $dirpath .= '/' unless $dirpath =~ m#[\\/]\z#; |
| 132 | } |
| 133 | elsif ($type eq "MacOS") { |
| 134 | ($dirpath,$basename) = ($fullname =~ /^(.*:)?(.*)/s); |
| 135 | $dirpath = ':' unless $dirpath; |
| 136 | } |
| 137 | elsif ($type eq "AmigaOS") { |
| 138 | ($dirpath,$basename) = ($fullname =~ /(.*[:\/])?(.*)/s); |
| 139 | $dirpath = './' unless $dirpath; |
| 140 | } |
| 141 | elsif ($type eq 'VMS' ) { |
| 142 | ($dirpath,$basename) = ($fullname =~ /^(.*[:>\]])?(.*)/s); |
| 143 | $dirpath ||= ''; # should always be defined |
| 144 | } |
| 145 | else { # Default to Unix semantics. |
| 146 | ($dirpath,$basename) = ($fullname =~ m#^(.*/)?(.*)#s); |
| 147 | if ($orig_type eq 'VMS' and $fullname =~ m:^(/[^/]+/000000(/|$))(.*):) { |
| 148 | # dev:[000000] is top of VMS tree, similar to Unix '/' |
| 149 | # so strip it off and treat the rest as "normal" |
| 150 | my $devspec = $1; |
| 151 | my $remainder = $3; |
| 152 | ($dirpath,$basename) = ($remainder =~ m#^(.*/)?(.*)#s); |
| 153 | $dirpath ||= ''; # should always be defined |
| 154 | $dirpath = $devspec.$dirpath; |
| 155 | } |
| 156 | $dirpath = './' unless $dirpath; |
| 157 | } |
| 158 | |
| 159 | |
| 160 | my $tail = ''; |
| 161 | my $suffix = ''; |
| 162 | if (@suffices) { |
| 163 | foreach $suffix (@suffices) { |
| 164 | my $pat = ($igncase ? '(?i)' : '') . "($suffix)\$"; |
| 165 | if ($basename =~ s/$pat//s) { |
| 166 | $taint .= substr($suffix,0,0); |
| 167 | $tail = $1 . $tail; |
| 168 | } |
| 169 | } |
| 170 | } |
| 171 | |
| 172 | # Ensure taint is propgated from the path to its pieces. |
| 173 | $tail .= $taint; |
| 174 | wantarray ? ($basename .= $taint, $dirpath .= $taint, $tail) |
| 175 | : ($basename .= $taint); |
| 176 | } |
| 177 | |
| 178 | |
| 179 | |
| 180 | =item C<basename> |
| 181 | |
| 182 | my $filename = basename($path); |
| 183 | my $filename = basename($path, @suffixes); |
| 184 | |
| 185 | This function is provided for compatibility with the Unix shell command |
| 186 | C<basename(1)>. It does B<NOT> always return the file name portion of a |
| 187 | path as you might expect. To be safe, if you want the file name portion of |
| 188 | a path use C<fileparse()>. |
| 189 | |
| 190 | C<basename()> returns the last level of a filepath even if the last |
| 191 | level is clearly directory. In effect, it is acting like C<pop()> for |
| 192 | paths. This differs from C<fileparse()>'s behaviour. |
| 193 | |
| 194 | # Both return "bar" |
| 195 | basename("/foo/bar"); |
| 196 | basename("/foo/bar/"); |
| 197 | |
| 198 | @suffixes work as in C<fileparse()> except all regex metacharacters are |
| 199 | quoted. |
| 200 | |
| 201 | # These two function calls are equivalent. |
| 202 | my $filename = basename("/foo/bar/baz.txt", ".txt"); |
| 203 | my $filename = fileparse("/foo/bar/baz.txt", qr/\Q.txt\E/); |
| 204 | |
| 205 | Also note that in order to be compatible with the shell command, |
| 206 | C<basename()> does not strip off a suffix if it is identical to the |
| 207 | remaining characters in the filename. |
| 208 | |
| 209 | =cut |
| 210 | |
| 211 | |
| 212 | sub basename { |
| 213 | my($path) = shift; |
| 214 | |
| 215 | # From BSD basename(1) |
| 216 | # The basename utility deletes any prefix ending with the last slash `/' |
| 217 | # character present in string (after first stripping trailing slashes) |
| 218 | _strip_trailing_sep($path); |
| 219 | |
| 220 | my($basename, $dirname, $suffix) = fileparse( $path, map("\Q$_\E",@_) ); |
| 221 | |
| 222 | # From BSD basename(1) |
| 223 | # The suffix is not stripped if it is identical to the remaining |
| 224 | # characters in string. |
| 225 | if( length $suffix and !length $basename ) { |
| 226 | $basename = $suffix; |
| 227 | } |
| 228 | |
| 229 | # Ensure that basename '/' == '/' |
| 230 | if( !length $basename ) { |
| 231 | $basename = $dirname; |
| 232 | } |
| 233 | |
| 234 | return $basename; |
| 235 | } |
| 236 | |
| 237 | |
| 238 | |
| 239 | =item C<dirname> |
| 240 | |
| 241 | This function is provided for compatibility with the Unix shell |
| 242 | command C<dirname(1)> and has inherited some of its quirks. In spite of |
| 243 | its name it does B<NOT> always return the directory name as you might |
| 244 | expect. To be safe, if you want the directory name of a path use |
| 245 | C<fileparse()>. |
| 246 | |
| 247 | Only on VMS (where there is no ambiguity between the file and directory |
| 248 | portions of a path) and AmigaOS (possibly due to an implementation quirk in |
| 249 | this module) does C<dirname()> work like C<fileparse($path)>, returning just the |
| 250 | $directories. |
| 251 | |
| 252 | # On VMS and AmigaOS |
| 253 | my $directories = dirname($path); |
| 254 | |
| 255 | When using Unix or MSDOS syntax this emulates the C<dirname(1)> shell function |
| 256 | which is subtly different from how C<fileparse()> works. It returns all but |
| 257 | the last level of a file path even if the last level is clearly a directory. |
| 258 | In effect, it is not returning the directory portion but simply the path one |
| 259 | level up acting like C<chop()> for file paths. |
| 260 | |
| 261 | Also unlike C<fileparse()>, C<dirname()> does not include a trailing slash on |
| 262 | its returned path. |
| 263 | |
| 264 | # returns /foo/bar. fileparse() would return /foo/bar/ |
| 265 | dirname("/foo/bar/baz"); |
| 266 | |
| 267 | # also returns /foo/bar despite the fact that baz is clearly a |
| 268 | # directory. fileparse() would return /foo/bar/baz/ |
| 269 | dirname("/foo/bar/baz/"); |
| 270 | |
| 271 | # returns '.'. fileparse() would return 'foo/' |
| 272 | dirname("foo/"); |
| 273 | |
| 274 | Under VMS, if there is no directory information in the $path, then the |
| 275 | current default device and directory is used. |
| 276 | |
| 277 | =cut |
| 278 | |
| 279 | |
| 280 | sub dirname { |
| 281 | my $path = shift; |
| 282 | |
| 283 | my($type) = $Fileparse_fstype; |
| 284 | |
| 285 | if( $type eq 'VMS' and $path =~ m{/} ) { |
| 286 | # Parse as Unix |
| 287 | local($File::Basename::Fileparse_fstype) = ''; |
| 288 | return dirname($path); |
| 289 | } |
| 290 | |
| 291 | my($basename, $dirname) = fileparse($path); |
| 292 | |
| 293 | if ($type eq 'VMS') { |
| 294 | $dirname ||= $ENV{DEFAULT}; |
| 295 | } |
| 296 | elsif ($type eq 'MacOS') { |
| 297 | if( !length($basename) && $dirname !~ /^[^:]+:\z/) { |
| 298 | _strip_trailing_sep($dirname); |
| 299 | ($basename,$dirname) = fileparse $dirname; |
| 300 | } |
| 301 | $dirname .= ":" unless $dirname =~ /:\z/; |
| 302 | } |
| 303 | elsif (grep { $type eq $_ } qw(MSDOS DOS MSWin32 OS2)) { |
| 304 | _strip_trailing_sep($dirname); |
| 305 | unless( length($basename) ) { |
| 306 | ($basename,$dirname) = fileparse $dirname; |
| 307 | _strip_trailing_sep($dirname); |
| 308 | } |
| 309 | } |
| 310 | elsif ($type eq 'AmigaOS') { |
| 311 | if ( $dirname =~ /:\z/) { return $dirname } |
| 312 | chop $dirname; |
| 313 | $dirname =~ s#[^:/]+\z## unless length($basename); |
| 314 | } |
| 315 | else { |
| 316 | _strip_trailing_sep($dirname); |
| 317 | unless( length($basename) ) { |
| 318 | ($basename,$dirname) = fileparse $dirname; |
| 319 | _strip_trailing_sep($dirname); |
| 320 | } |
| 321 | } |
| 322 | |
| 323 | $dirname; |
| 324 | } |
| 325 | |
| 326 | |
| 327 | # Strip the trailing path separator. |
| 328 | sub _strip_trailing_sep { |
| 329 | my $type = $Fileparse_fstype; |
| 330 | |
| 331 | if ($type eq 'MacOS') { |
| 332 | $_[0] =~ s/([^:]):\z/$1/s; |
| 333 | } |
| 334 | elsif (grep { $type eq $_ } qw(MSDOS DOS MSWin32 OS2)) { |
| 335 | $_[0] =~ s/([^:])[\\\/]*\z/$1/; |
| 336 | } |
| 337 | else { |
| 338 | $_[0] =~ s{(.)/*\z}{$1}s; |
| 339 | } |
| 340 | } |
| 341 | |
| 342 | |
| 343 | =item C<fileparse_set_fstype> |
| 344 | |
| 345 | my $type = fileparse_set_fstype(); |
| 346 | my $previous_type = fileparse_set_fstype($type); |
| 347 | |
| 348 | Normally File::Basename will assume a file path type native to your current |
| 349 | operating system (ie. /foo/bar style on Unix, \foo\bar on Windows, etc...). |
| 350 | With this function you can override that assumption. |
| 351 | |
| 352 | Valid $types are "MacOS", "VMS", "AmigaOS", "OS2", "RISCOS", |
| 353 | "MSWin32", "DOS" (also "MSDOS" for backwards bug compatibility), |
| 354 | "Epoc" and "Unix" (all case-insensitive). If an unrecognized $type is |
| 355 | given "Unix" will be assumed. |
| 356 | |
| 357 | If you've selected VMS syntax, and the file specification you pass to |
| 358 | one of these routines contains a "/", they assume you are using Unix |
| 359 | emulation and apply the Unix syntax rules instead, for that function |
| 360 | call only. |
| 361 | |
| 362 | =back |
| 363 | |
| 364 | =cut |
| 365 | |
| 366 | |
| 367 | BEGIN { |
| 368 | |
| 369 | my @Ignore_Case = qw(MacOS VMS AmigaOS OS2 RISCOS MSWin32 MSDOS DOS Epoc); |
| 370 | my @Types = (@Ignore_Case, qw(Unix)); |
| 371 | |
| 372 | sub fileparse_set_fstype { |
| 373 | my $old = $Fileparse_fstype; |
| 374 | |
| 375 | if (@_) { |
| 376 | my $new_type = shift; |
| 377 | |
| 378 | $Fileparse_fstype = 'Unix'; # default |
| 379 | foreach my $type (@Types) { |
| 380 | $Fileparse_fstype = $type if $new_type =~ /^$type/i; |
| 381 | } |
| 382 | |
| 383 | $Fileparse_igncase = |
| 384 | (grep $Fileparse_fstype eq $_, @Ignore_Case) ? 1 : 0; |
| 385 | } |
| 386 | |
| 387 | return $old; |
| 388 | } |
| 389 | |
| 390 | } |
| 391 | |
| 392 | |
| 393 | 1; |
| 394 | |
| 395 | |
| 396 | =head1 SEE ALSO |
| 397 | |
| 398 | L<dirname(1)>, L<basename(1)>, L<File::Spec> |