3 ###############################################################################
5 # Name : filename_helper #
6 # Author : Chris Koeritz #
7 # Rights : Copyright (C) 1996-$now by Author #
11 # Support for manipulating filenames. #
13 ###############################################################################
14 # This program is free software; you can redistribute it and/or modify it #
15 # under the terms of the GNU General Public License as published by the Free #
16 # Software Foundation; either version 2 of the License or (at your option) #
17 # any later version. See: "http://www.gruntose.com/Info/GNU/GPL.html" for a #
18 # version of the License. Please send any updates to "fred@gruntose.com". #
19 ###############################################################################
21 use Env qw(OS IS_MSYS);
23 ############################################################################
25 #hmmm: make this lower-level, a script that is inherited by all perl scripts.
27 sub yeti_interrupt_handler {
28 die "caught an interrupt; exiting.\n";
31 # hook in a ctrl-c catcher, since that seems to be universally needed.
32 sub install_interrupt_catcher {
33 $SIG{INT} = 'yeti_interrupt_handler';
34 $SIG{QUIT} = 'yeti_interrupt_handler';
35 #print "mapped int and quit signals\n";
39 ############################################################################
41 # takes an array of filenames (each possibly containing spaces and/or
42 # wildcards) and resolves it to a useful list of actual files.
45 local(@temp_list) = @_; # the names we're given.
46 local(@to_return) = (); # the final form of the name list.
47 #print "temp list is @temp_list\n";
49 # scan through the list we're given.
50 foreach $entry (@temp_list) {
51 #print "entry is $entry\n";
52 local(@chopped_filename) = &split_filename($entry);
53 #print "chopped 0=$chopped_filename[0]\n";
54 #print "chopped 1=$chopped_filename[1]\n";
55 if ( (@chopped_filename[0] eq ".") || (@chopped_filename[0] eq "..") ) {
56 # add the simple directory name into the list.
57 push @to_return, $chopped_filename[0];
60 if (@chopped_filename[1] eq ".") {
61 # add a directory that didn't have more pattern attached.
62 push @to_return, $chopped_filename[0];
65 opendir WHERE, $chopped_filename[0]; # open the directory up.
66 local(@files_found) = readdir(WHERE);
68 foreach $possible_name (@files_found) {
69 # we need to process the pattern a bit; directory patterns are different
70 # from perl regular expression patterns, so we end up massaging any "ls"
71 # wildcards into an equivalent perl-style one below.
72 local($match) = $chopped_filename[1];
73 #hmmm: would be nice to combine the replacements into a long batch instead of separate commands, but i do not seem to know how to do that yet in perl.
74 $match =~ s/\./\\./g; # replace periods with escaped ones.
75 $match =~ s/\*/.*/g; # replace asterisks with dot star.
76 $match =~ s/\+/\\+/g; # escape plusses.
77 $match =~ s/\?/\\?/g; # escape question marks.
78 $match =~ s/\|/\\?/g; # escape pipe char.
79 $match =~ s/\$/\\\$/g; # escape dollar sign.
80 $match =~ s/\[/\\[/g; # escape open bracket.
81 $match =~ s/\]/\\]/g; # escape close bracket.
82 $match =~ s/\(/\\(/g; # escape open quote.
83 $match =~ s/\)/\\)/g; # escape close quote.
84 $match =~ s/\{/\\{/g; # escape open curly bracket.
85 $match =~ s/\}/\\}/g; # escape close curly bracket.
87 $match = "^" . $match . "\$"; # make sure the whole thing matches.
88 #print "possibname is '$possible_name':\n";
89 if ($possible_name =~ /$match/) {
90 # this one matches so add it.
91 push @to_return, $chopped_filename[0] . $possible_name;
98 ############################################################################
100 # reports if two file names are the same file.
103 local($file1, $file2) = @_;
105 ($dev1, $ino1, $junk1) = stat $file1;
106 ($dev2, $ino2, $junk2) = stat $file2;
108 return ($dev1 == $dev2) && ($ino1 == $ino2);
111 ############################################################################
113 # splits a filename into a directory and file specification.
116 local($chewed_name) = &remove_trailing_slashes(@_);
117 $chewed_name = &canonicalize($chewed_name);
118 $chewed_name = &patch_name_for_pc($chewed_name);
119 if ($chewed_name =~ /\//) {
120 # there's a slash in there.
121 local($directory_part) = $chewed_name;
122 $directory_part =~ s/^(.*\/)[^\/]*$/\1/;
123 local($file_part) = $chewed_name;
124 $file_part =~ s/^.*\/([^\/]*)$/\1/;
125 if ($file_part eq "") {
126 # if there was no file specification, just add a non-matching spec.
129 return ($directory_part, $file_part);
130 } elsif ($chewed_name eq ".") {
132 } elsif ($chewed_name eq "..") {
135 # no slash in this name, so we fix that and also make sure we match
137 return ("./", $chewed_name);
141 ############################################################################
143 # returns the base part of the filename; this omits any directories.
146 local(@parts) = &split_filename(@_);
150 # returns the directory part of the filename.
153 local(@parts) = &split_filename(@_);
157 # returns the extension found on the filename, if any.
159 local($base) = &basename(@_);
160 #printf "base is $base\n";
162 for (local($i) = length($base) - 1; $i >= 0; $i--) {
163 #printf "char is " . substr($base, $i, 1) . "\n";
164 if (substr($base, $i, 1) eq '.') {
166 #printf "got period found is $found\n";
171 return substr($base, $found, length($base) - $found);
173 return ""; # no extension seen.
176 # returns the portion of the filename without the extension.
178 local($full) = &remove_trailing_slashes(@_);
179 $full = &canonicalize($full);
180 $full = &patch_name_for_pc($full);
181 local($ext) = &extension($full);
182 local($to_remove) = length($ext);
183 return substr($full, 0, length($full) - $to_remove);
186 ############################################################################
188 # removes all directory slashes (either '/' or '\') from the end of a string.
190 sub remove_trailing_slashes {
191 local($directory_name) = @_;
192 # start looking at the end of the string.
193 local($inspection_point) = length($directory_name) - 1;
194 while ($inspection_point > 0) {
195 # examine the last character in the string to see if it's a slash.
196 local($final_char) = substr($directory_name, $inspection_point, 1);
197 # leave the loop if it's not a slash.
198 if ( ($final_char ne "/") && ($final_char ne "\\") ) { last; }
199 chop($directory_name); # remove the slash.
200 $inspection_point--; # check the new last character.
203 return $directory_name;
206 ############################################################################
208 # returns the proper directory separator for this platform. this requires
209 # an environment variable called "OS" for non-Unix operating systems. the
210 # valid values for that are shown below.
212 sub directory_separator {
213 if ( ($OS eq "Windows_NT") || ($OS eq "Windows_95")
214 || ($OS eq "DOS") || ($OS eq "OS2") ) { return "\\"; }
218 ############################################################################
220 # these mutate the directory slashes in a directory name.
222 # the one we use most frequently; it uses the unix slash.
224 return &canonicalizer(@_, "/");
227 # one that turns names into the style native on the current platform.
228 sub native_canonicalize {
229 return &canonicalizer(@_, &directory_separator());
232 # one that explicitly uses pc style back-slashes.
233 sub pc_canonicalize {
234 return &canonicalizer(@_, "\\");
237 # one that explicitly does unix style forward slashes.
238 sub unix_canonicalize {
239 return &canonicalizer(@_, "/");
242 # this more general routine gets a directory separator passed in. it then
243 # replaces all the separators with that one.
245 local($directory_name) = $_[0];
246 local($dirsep) = $_[1];
248 #print "old dir name is \"$directory_name\"\n";
251 #somewhat abbreviated check; only catches windoze systems, not dos or os2.
252 # IS_MSYS is calculated by feisty meow scripts startup; it will be
253 # non-empty if this is the msys tool kit.
254 if (length($IS_MSYS) > 0) {
255 # msys utilities version (http://www.mingw.org)
256 $directory_name =~ s/^(.):[\\\/](.*)$/\/\1\/\2/;
258 # cygwin utilities version (http://www.cygwin.com)
259 $directory_name =~ s/^(.):[\\\/](.*)$/\/cygdrive\/\1\/\2/;
261 #print "new dir name is \"$directory_name\"\n";
264 # turn all the non-default separators into the default.
265 for (local($j) = 0; $j < length($directory_name); $j++) {
266 if ( (substr($directory_name, $j, 1) eq "\\")
267 || (substr($directory_name, $j, 1) eq "/") ) {
268 substr($directory_name, $j, 1) = $dirsep;
271 # remove all occurrences of double separators except for the first
272 # double set, which could be a UNC filename.
274 for (local($i) = 1; $i < length($directory_name); $i++) {
275 # iterate through the string looking for redundant separators.
276 if (substr($directory_name, $i, 1) eq $dirsep) {
277 # we found a separator character.
279 # we had just seen a separator, so this is two in a row.
280 local($head, $tail) = (substr($directory_name, 0, $i - 1),
281 substr($directory_name, $i, length($directory_name) - 1));
282 $directory_name = $head . $tail;
283 # put the name back together without this redundant character.
284 $i--; # skip back one and try again.
286 # we have now seen a separator.
290 # this character was not a separator.
294 if ($directory_name =~ /^.:$/) {
295 # fix a dos style directory that's just X:, since we don't want the
296 # current directory to be used on that device. that's too random.
297 # instead, we assume they meant the root of the drive.
298 $directory_name = $directory_name . "/";
300 return $directory_name;
303 ############################################################################
305 # fixes a PC directory name if it is only a drive letter plus colon.
307 sub patch_name_for_pc {
309 #print "name=$name\n";
310 if (length($name) != 2) { return $name; }
311 local($colon) = substr($name, 1, 1);
312 #print "colon=$colon\n";
313 # check whether the string needs patching.
315 # name is currently in feeble form of "X:"; fix it.
318 #print "returning=$name\n";
322 ############################################################################
324 # tells whether a filename is important or not. the unimportant category
325 # can usually be safely ignored or deleted.
327 sub important_filename {
328 local($name) = &basename($_[0]);
330 # these are endings that we consider unimportant. where a caret is used
331 # at the front, we will match only the whole string. double slashes are
332 # used before periods to ensure we match a real period character.
333 local(@junk_files) = ("~", "^\\.#.*", "^\\._.*", "\\.aps", "\\.bak",
334 "^binaries", "^bin.ant", "^bin.eclipse",
335 "\\.clw", "^cpdiff_tmp\\.txt", "^\\.ds_store", "^diffs\\.txt",
336 "^diff_tmp\\.txt", "\\.dsp", "\\.dsw", "\\.gid", "gmon\\.out", "\\.isr",
337 "^isconfig\\.ini", "\\.log", "^manifest.txt", "^obj",
338 "\\.obj", "\\.output", "\\.plg", "^RCa.*", "^Release", "\\.res",
339 "\\.sbr", ".*scc", "^Setup\\.dbg", "^Setup\\.inx",
340 "^Setup\\.map", "^Setup\\.obs", "^Selenium_.*Login.html",
341 "\\.stackdump", "^string1033\\.txt", "\\.suo", "\\.swp",
342 "^thumbs.db", "[a-zA-Z0-9]\\.tmp", "^trans\\.tbl", "\\.user", "_version\\.h",
343 "_version\\.rc", "^waste", "\\.ws4", "\\.wsm");
345 foreach $temp (@junk_files) {
347 if ($name =~ /${temp}/i) { return 0; }
348 # we hit a match on it being unimportant.
351 return 1; # anything else is considered important.
354 ############################################################################
357 return &patch_name_for_pc
358 (&remove_trailing_slashes
359 (&canonicalize(@_)));
362 ############################################################################
364 sub get_drive_letter {
366 if (substr($path, 0, 1) =~ /[a-zA-Z]/) {
367 if (substr($path, 1, 1) eq ":") { return substr($path, 0, 1); }
372 ############################################################################
374 sub remove_drive_letter {
376 if (substr($path, 0, 1) =~ /[a-zA-Z]/) {
377 if (substr($path, 1, 1) eq ":") { return substr($path, 2); }
382 ############################################################################
384 # these return their argument with the case flipped to lower or upper case.
388 $name =~ tr/A-Z/a-z/;
394 $name =~ tr/a-z/A-Z/;
398 ############################################################################
400 # recursively deletes a directory that is passed as the single parameter.
401 # from http://developer.novell.com/wiki/index.php/Recursive_Directory_Remove
402 sub recursive_delete {
406 print "this is not a dir: $dir\nshould whack it here?\n";
411 # if we can't open the dir, just skip to the next one.
412 opendir DIR, $dir or next;
413 while ($_ = readdir DIR) {
415 my $path = "$dir/$_";
416 unlink $path if -f $path;
417 recursive_delete($path) if -d $path;
420 rmdir $dir or print "error - $!";
424 ############################################################################
426 # finds any directories under the arguments, which can be a list of directories.
427 sub find_directories {
432 # if we can't open the dir, just skip to the next one.
433 opendir DIR, $dir or next;
434 while ($_ = readdir DIR) {
435 # skip if it's current or parent dir.
437 my $path = "$dir/$_";
438 # skip if this entry is not itself a directory.
440 push @dirs_found, $path;
447 ############################################################################
449 # given a list of paths, this returns an array of all the filenames found therein.
451 my @files_found = ();
455 # that's actually just a file, so add it.
456 push @files_found, $dir;
460 # if we can't open the dir, just skip to the next one.
461 opendir DIR, $dir or next;
462 while ($_ = readdir DIR) {
463 # skip if it's current or parent dir.
465 my $path = "$dir/$_";
466 # skip if this entry is not a file.
468 push @files_found, $path;
475 ############################################################################
477 # finds all directories starting at a particular directory and returns them
478 # in an array. does not include the starting directory.
479 sub recursive_find_directories {
480 # first find all the directories within the parameters.
481 my @toplevel = find_directories(@_);
484 push(@to_return, @toplevel);
486 # return the composition of the list we found here plus any directories under those.
487 # we only recurse if there's something to chew on in our directory list.
488 # otherwise, we've hit the bottom of that tree.
489 if (scalar @toplevel > 0) {
490 my @subs_found = recursive_find_directories(@toplevel);
491 push(@to_return, @subs_found);
496 ############################################################################