source: main/trunk/greenstone2/perllib/util.pm@ 23314

Last change on this file since 23314 was 23314, checked in by sjm84, 13 years ago

Katherine discovered that the newly added sub setup_greenstone_env needs to clear the GSDLHOME env variable before calling setup.bash (otherwise calling setup.bash/bat terminates assuming that the GS environment was already set up.

  • Property svn:keywords set to Author Date Id Revision
File size: 38.4 KB
Line 
1###########################################################################
2#
3# util.pm -- various useful utilities
4# A component of the Greenstone digital library software
5# from the New Zealand Digital Library Project at the
6# University of Waikato, New Zealand.
7#
8# Copyright (C) 1999 New Zealand Digital Library Project
9#
10# This program is free software; you can redistribute it and/or modify
11# it under the terms of the GNU General Public License as published by
12# the Free Software Foundation; either version 2 of the License, or
13# (at your option) any later version.
14#
15# This program is distributed in the hope that it will be useful,
16# but WITHOUT ANY WARRANTY; without even the implied warranty of
17# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18# GNU General Public License for more details.
19#
20# You should have received a copy of the GNU General Public License
21# along with this program; if not, write to the Free Software
22# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
23#
24###########################################################################
25
26package util;
27
28use File::Copy;
29use File::Basename;
30
31use strict;
32
33
34# removes files (but not directories)
35sub rm {
36 my (@files) = @_;
37
38 my @filefiles = ();
39
40 # make sure the files we want to delete exist
41 # and are regular files
42 foreach my $file (@files) {
43 if (!-e $file) {
44 print STDERR "util::rm $file does not exist\n";
45 } elsif ((!-f $file) && (!-l $file)) {
46 print STDERR "util::rm $file is not a regular (or symbolic) file\n";
47 } else {
48 push (@filefiles, $file);
49 }
50 }
51
52 # remove the files
53 my $numremoved = unlink @filefiles;
54
55 # check to make sure all of them were removed
56 if ($numremoved != scalar(@filefiles)) {
57 print STDERR "util::rm Not all files were removed\n";
58 }
59}
60
61# removes files (but not directories) - can rename this to the default
62# "rm" subroutine when debugging the deletion of individual files.
63sub rm_debug {
64 my (@files) = @_;
65 my @filefiles = ();
66
67 # make sure the files we want to delete exist
68 # and are regular files
69 foreach my $file (@files) {
70 if (!-e $file) {
71 print STDERR "util::rm $file does not exist\n";
72 } elsif ((!-f $file) && (!-l $file)) {
73 print STDERR "util::rm $file is not a regular (or symbolic) file\n";
74 } else { # debug message
75 unlink($file) or warn "Could not delete file $file: $!\n";
76 }
77 }
78}
79
80
81# recursive removal
82sub filtered_rm_r {
83 my ($files,$file_accept_re,$file_reject_re) = @_;
84
85 my @files_array = (ref $files eq "ARRAY") ? @$files : ($files);
86
87 # recursively remove the files
88 foreach my $file (@files_array) {
89 $file =~ s/[\/\\]+$//; # remove trailing slashes
90
91 if (!-e $file) {
92 print STDERR "util::filtered_rm_r $file does not exist\n";
93
94 } elsif ((-d $file) && (!-l $file)) { # don't recurse down symbolic link
95 # get the contents of this directory
96 if (!opendir (INDIR, $file)) {
97 print STDERR "util::filtered_rm_r could not open directory $file\n";
98 } else {
99 my @filedir = grep (!/^\.\.?$/, readdir (INDIR));
100 closedir (INDIR);
101
102 # remove all the files in this directory
103 map {$_="$file/$_";} @filedir;
104 &filtered_rm_r (\@filedir,$file_accept_re,$file_reject_re);
105
106 if (!defined $file_accept_re && !defined $file_reject_re) {
107 # remove this directory
108 if (!rmdir $file) {
109 print STDERR "util::filtered_rm_r couldn't remove directory $file\n";
110 }
111 }
112 }
113 } else {
114 next if (defined $file_reject_re && ($file =~ m/$file_reject_re/));
115
116 if ((!defined $file_accept_re) || ($file =~ m/$file_accept_re/)) {
117 # remove this file
118 &rm ($file);
119 }
120 }
121 }
122}
123
124
125# recursive removal
126sub rm_r {
127 my (@files) = @_;
128
129 # use the more general (but reterospectively written function
130 # filtered_rm_r function()
131
132 filtered_rm_r(\@files,undef,undef); # no accept or reject expressions
133}
134
135
136
137
138# moves a file or a group of files
139sub mv {
140 my $dest = pop (@_);
141 my (@srcfiles) = @_;
142
143 # remove trailing slashes from source and destination files
144 $dest =~ s/[\\\/]+$//;
145 map {$_ =~ s/[\\\/]+$//;} @srcfiles;
146
147 # a few sanity checks
148 if (scalar (@srcfiles) == 0) {
149 print STDERR "util::mv no destination directory given\n";
150 return;
151 } elsif ((scalar (@srcfiles) > 1) && (!-d $dest)) {
152 print STDERR "util::mv if multiple source files are given the ".
153 "destination must be a directory\n";
154 return;
155 }
156
157 # move the files
158 foreach my $file (@srcfiles) {
159 my $tempdest = $dest;
160 if (-d $tempdest) {
161 my ($filename) = $file =~ /([^\\\/]+)$/;
162 $tempdest .= "/$filename";
163 }
164 if (!-e $file) {
165 print STDERR "util::mv $file does not exist\n";
166 } else {
167 rename ($file, $tempdest);
168 }
169 }
170}
171
172
173# copies a file or a group of files
174sub cp {
175 my $dest = pop (@_);
176 my (@srcfiles) = @_;
177
178 # remove trailing slashes from source and destination files
179 $dest =~ s/[\\\/]+$//;
180 map {$_ =~ s/[\\\/]+$//;} @srcfiles;
181
182 # a few sanity checks
183 if (scalar (@srcfiles) == 0) {
184 print STDERR "util::cp no destination directory given\n";
185 return;
186 } elsif ((scalar (@srcfiles) > 1) && (!-d $dest)) {
187 print STDERR "util::cp if multiple source files are given the ".
188 "destination must be a directory\n";
189 return;
190 }
191
192 # copy the files
193 foreach my $file (@srcfiles) {
194 my $tempdest = $dest;
195 if (-d $tempdest) {
196 my ($filename) = $file =~ /([^\\\/]+)$/;
197 $tempdest .= "/$filename";
198 }
199 if (!-e $file) {
200 print STDERR "util::cp $file does not exist\n";
201 } elsif (!-f $file) {
202 print STDERR "util::cp $file is not a plain file\n";
203 } else {
204 &File::Copy::copy ($file, $tempdest);
205 }
206 }
207}
208
209
210
211# recursively copies a file or group of files
212# syntax: cp_r (sourcefiles, destination directory)
213# destination must be a directory - to copy one file to
214# another use cp instead
215sub cp_r {
216 my $dest = pop (@_);
217 my (@srcfiles) = @_;
218
219 # a few sanity checks
220 if (scalar (@srcfiles) == 0) {
221 print STDERR "util::cp_r no destination directory given\n";
222 return;
223 } elsif (-f $dest) {
224 print STDERR "util::cp_r destination must be a directory\n";
225 return;
226 }
227
228 # create destination directory if it doesn't exist already
229 if (! -d $dest) {
230 my $store_umask = umask(0002);
231 mkdir ($dest, 0777);
232 umask($store_umask);
233 }
234
235 # copy the files
236 foreach my $file (@srcfiles) {
237
238 if (!-e $file) {
239 print STDERR "util::cp_r $file does not exist\n";
240
241 } elsif (-d $file) {
242 # make the new directory
243 my ($filename) = $file =~ /([^\\\/]*)$/;
244 $dest = &util::filename_cat ($dest, $filename);
245 my $store_umask = umask(0002);
246 mkdir ($dest, 0777);
247 umask($store_umask);
248
249 # get the contents of this directory
250 if (!opendir (INDIR, $file)) {
251 print STDERR "util::cp_r could not open directory $file\n";
252 } else {
253 my @filedir = readdir (INDIR);
254 closedir (INDIR);
255 foreach my $f (@filedir) {
256 next if $f =~ /^\.\.?$/;
257 # copy all the files in this directory
258 my $ff = &util::filename_cat ($file, $f);
259 &cp_r ($ff, $dest);
260 }
261 }
262
263 } else {
264 &cp($file, $dest);
265 }
266 }
267}
268# recursively copies a file or group of files
269# syntax: cp_r (sourcefiles, destination directory)
270# destination must be a directory - to copy one file to
271# another use cp instead
272sub cp_r_nosvn {
273 my $dest = pop (@_);
274 my (@srcfiles) = @_;
275
276 # a few sanity checks
277 if (scalar (@srcfiles) == 0) {
278 print STDERR "util::cp_r no destination directory given\n";
279 return;
280 } elsif (-f $dest) {
281 print STDERR "util::cp_r destination must be a directory\n";
282 return;
283 }
284
285 # create destination directory if it doesn't exist already
286 if (! -d $dest) {
287 my $store_umask = umask(0002);
288 mkdir ($dest, 0777);
289 umask($store_umask);
290 }
291
292 # copy the files
293 foreach my $file (@srcfiles) {
294
295 if (!-e $file) {
296 print STDERR "util::cp_r $file does not exist\n";
297
298 } elsif (-d $file) {
299 # make the new directory
300 my ($filename) = $file =~ /([^\\\/]*)$/;
301 $dest = &util::filename_cat ($dest, $filename);
302 my $store_umask = umask(0002);
303 mkdir ($dest, 0777);
304 umask($store_umask);
305
306 # get the contents of this directory
307 if (!opendir (INDIR, $file)) {
308 print STDERR "util::cp_r could not open directory $file\n";
309 } else {
310 my @filedir = readdir (INDIR);
311 closedir (INDIR);
312 foreach my $f (@filedir) {
313 next if $f =~ /^\.\.?$/;
314 next if $f =~ /^\.svn$/;
315 # copy all the files in this directory
316 my $ff = &util::filename_cat ($file, $f);
317 &cp_r ($ff, $dest);
318 }
319 }
320
321 } else {
322 &cp($file, $dest);
323 }
324 }
325}
326
327# copies a directory and its contents, excluding subdirectories, into a new directory
328sub cp_r_toplevel {
329 my $dest = pop (@_);
330 my (@srcfiles) = @_;
331
332 # a few sanity checks
333 if (scalar (@srcfiles) == 0) {
334 print STDERR "util::cp_r no destination directory given\n";
335 return;
336 } elsif (-f $dest) {
337 print STDERR "util::cp_r destination must be a directory\n";
338 return;
339 }
340
341 # create destination directory if it doesn't exist already
342 if (! -d $dest) {
343 my $store_umask = umask(0002);
344 mkdir ($dest, 0777);
345 umask($store_umask);
346 }
347
348 # copy the files
349 foreach my $file (@srcfiles) {
350
351 if (!-e $file) {
352 print STDERR "util::cp_r $file does not exist\n";
353
354 } elsif (-d $file) {
355 # make the new directory
356 my ($filename) = $file =~ /([^\\\/]*)$/;
357 $dest = &util::filename_cat ($dest, $filename);
358 my $store_umask = umask(0002);
359 mkdir ($dest, 0777);
360 umask($store_umask);
361
362 # get the contents of this directory
363 if (!opendir (INDIR, $file)) {
364 print STDERR "util::cp_r could not open directory $file\n";
365 } else {
366 my @filedir = readdir (INDIR);
367 closedir (INDIR);
368 foreach my $f (@filedir) {
369 next if $f =~ /^\.\.?$/;
370
371 # copy all the files in this directory, but not directories
372 my $ff = &util::filename_cat ($file, $f);
373 if (-f $ff) {
374 &cp($ff, $dest);
375 #&cp_r ($ff, $dest);
376 }
377 }
378 }
379
380 } else {
381 &cp($file, $dest);
382 }
383 }
384}
385
386sub mk_dir {
387 my ($dir) = @_;
388
389 my $store_umask = umask(0002);
390 my $mkdir_ok = mkdir ($dir, 0777);
391 umask($store_umask);
392
393 if (!$mkdir_ok)
394 {
395 print STDERR "util::mk_dir could not create directory $dir\n";
396 return;
397 }
398}
399
400# in case anyone cares - I did some testing (using perls Benchmark module)
401# on this subroutine against File::Path::mkpath (). mk_all_dir() is apparently
402# slightly faster (surprisingly) - Stefan.
403sub mk_all_dir {
404 my ($dir) = @_;
405
406 # use / for the directory separator, remove duplicate and
407 # trailing slashes
408 $dir=~s/[\\\/]+/\//g;
409 $dir=~s/[\\\/]+$//;
410
411 # make sure the cache directory exists
412 my $dirsofar = "";
413 my $first = 1;
414 foreach my $dirname (split ("/", $dir)) {
415 $dirsofar .= "/" unless $first;
416 $first = 0;
417
418 $dirsofar .= $dirname;
419
420 next if $dirname =~ /^(|[a-z]:)$/i;
421 if (!-e $dirsofar)
422 {
423 my $store_umask = umask(0002);
424 my $mkdir_ok = mkdir ($dirsofar, 0777);
425 umask($store_umask);
426 if (!$mkdir_ok)
427 {
428 print STDERR "util::mk_all_dir could not create directory $dirsofar\n";
429 return;
430 }
431 }
432 }
433}
434
435# make hard link to file if supported by OS, otherwise copy the file
436sub hard_link {
437 my ($src, $dest, $verbosity) = @_;
438
439 # remove trailing slashes from source and destination files
440 $src =~ s/[\\\/]+$//;
441 $dest =~ s/[\\\/]+$//;
442
443## print STDERR "**** src = ", unicode::debug_unicode_string($src),"\n";
444 # a few sanity checks
445 if (-e $dest) {
446 # destination file already exists
447 return;
448 }
449 elsif (!-e $src) {
450 print STDERR "util::hard_link source file \"$src\" does not exist\n";
451 return 1;
452 }
453 elsif (-d $src) {
454 print STDERR "util::hard_link source \"$src\" is a directory\n";
455 return 1;
456 }
457
458 my $dest_dir = &File::Basename::dirname($dest);
459 mk_all_dir($dest_dir) if (!-e $dest_dir);
460
461
462 if (!link($src, $dest)) {
463 if ((!defined $verbosity) || ($verbosity>2)) {
464 print STDERR "util::hard_link: unable to create hard link. ";
465 print STDERR " Copying file: $src -> $dest\n";
466 }
467 &File::Copy::copy ($src, $dest);
468 }
469 return 0;
470}
471
472# make soft link to file if supported by OS, otherwise copy file
473sub soft_link {
474 my ($src, $dest, $ensure_paths_absolute) = @_;
475
476 # remove trailing slashes from source and destination files
477 $src =~ s/[\\\/]+$//;
478 $dest =~ s/[\\\/]+$//;
479
480 # Ensure file paths are absolute IF requested to do so
481 # Soft_linking didn't work for relative paths
482 if(defined $ensure_paths_absolute && $ensure_paths_absolute) {
483 # We need to ensure that the src file is the absolute path
484 # See http://perldoc.perl.org/File/Spec.html
485 if(!File::Spec->file_name_is_absolute( $src )) { # it's relative
486 $src = File::Spec->rel2abs($src); # make absolute
487 }
488 # Might as well ensure that the destination file's absolute path is used
489 if(!File::Spec->file_name_is_absolute( $dest )) {
490 $dest = File::Spec->rel2abs($dest); # make absolute
491 }
492 }
493
494 # a few sanity checks
495 if (!-e $src) {
496 print STDERR "util::soft_link source file $src does not exist\n";
497 return 0;
498 }
499
500 my $dest_dir = &File::Basename::dirname($dest);
501 mk_all_dir($dest_dir) if (!-e $dest_dir);
502
503 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
504 # symlink not supported on windows
505 &File::Copy::copy ($src, $dest);
506
507 } elsif (!eval {symlink($src, $dest)}) {
508 print STDERR "util::soft_link: unable to create soft link.\n";
509 return 0;
510 }
511
512 return 1;
513}
514
515
516
517
518# updates a copy of a directory in some other part of the filesystem
519# verbosity settings are: 0=low, 1=normal, 2=high
520# both $fromdir and $todir should be absolute paths
521sub cachedir {
522 my ($fromdir, $todir, $verbosity) = @_;
523 $verbosity = 1 unless defined $verbosity;
524
525 # use / for the directory separator, remove duplicate and
526 # trailing slashes
527 $fromdir=~s/[\\\/]+/\//g;
528 $fromdir=~s/[\\\/]+$//;
529 $todir=~s/[\\\/]+/\//g;
530 $todir=~s/[\\\/]+$//;
531
532 &mk_all_dir ($todir);
533
534 # get the directories in ascending order
535 if (!opendir (FROMDIR, $fromdir)) {
536 print STDERR "util::cachedir could not read directory $fromdir\n";
537 return;
538 }
539 my @fromdir = grep (!/^\.\.?$/, sort(readdir (FROMDIR)));
540 closedir (FROMDIR);
541
542 if (!opendir (TODIR, $todir)) {
543 print STDERR "util::cacedir could not read directory $todir\n";
544 return;
545 }
546 my @todir = grep (!/^\.\.?$/, sort(readdir (TODIR)));
547 closedir (TODIR);
548
549 my $fromi = 0;
550 my $toi = 0;
551
552 while ($fromi < scalar(@fromdir) || $toi < scalar(@todir)) {
553# print "fromi: $fromi toi: $toi\n";
554
555 # see if we should delete a file/directory
556 # this should happen if the file/directory
557 # is not in the from list or if its a different
558 # size, or has an older timestamp
559 if ($toi < scalar(@todir)) {
560 if (($fromi >= scalar(@fromdir)) ||
561 ($todir[$toi] lt $fromdir[$fromi] ||
562 ($todir[$toi] eq $fromdir[$fromi] &&
563 &differentfiles("$fromdir/$fromdir[$fromi]","$todir/$todir[$toi]",
564 $verbosity)))) {
565
566 # the files are different
567 &rm_r("$todir/$todir[$toi]");
568 splice(@todir, $toi, 1); # $toi stays the same
569
570 } elsif ($todir[$toi] eq $fromdir[$fromi]) {
571 # the files are the same
572 # if it is a directory, check its contents
573 if (-d "$todir/$todir[$toi]") {
574 &cachedir ("$fromdir/$fromdir[$fromi]",
575 "$todir/$todir[$toi]", $verbosity);
576 }
577
578 $toi++;
579 $fromi++;
580 next;
581 }
582 }
583
584 # see if we should insert a file/directory
585 # we should insert a file/directory if there
586 # is no tofiles left or if the tofile does not exist
587 if ($fromi < scalar(@fromdir) && ($toi >= scalar(@todir) ||
588 $todir[$toi] gt $fromdir[$fromi])) {
589 &cp_r ("$fromdir/$fromdir[$fromi]", "$todir/$fromdir[$fromi]");
590 splice (@todir, $toi, 0, $fromdir[$fromi]);
591
592 $toi++;
593 $fromi++;
594 }
595 }
596}
597
598# this function returns -1 if either file is not found
599# assumes that $file1 and $file2 are absolute file names or
600# in the current directory
601# $file2 is allowed to be newer than $file1
602sub differentfiles {
603 my ($file1, $file2, $verbosity) = @_;
604 $verbosity = 1 unless defined $verbosity;
605
606 $file1 =~ s/\/+$//;
607 $file2 =~ s/\/+$//;
608
609 my ($file1name) = $file1 =~ /\/([^\/]*)$/;
610 my ($file2name) = $file2 =~ /\/([^\/]*)$/;
611
612 return -1 unless (-e $file1 && -e $file2);
613 if ($file1name ne $file2name) {
614 print STDERR "filenames are not the same\n" if ($verbosity >= 2);
615 return 1;
616 }
617
618 my @file1stat = stat ($file1);
619 my @file2stat = stat ($file2);
620
621 if (-d $file1) {
622 if (! -d $file2) {
623 print STDERR "one file is a directory\n" if ($verbosity >= 2);
624 return 1;
625 }
626 return 0;
627 }
628
629 # both must be regular files
630 unless (-f $file1 && -f $file2) {
631 print STDERR "one file is not a regular file\n" if ($verbosity >= 2);
632 return 1;
633 }
634
635 # the size of the files must be the same
636 if ($file1stat[7] != $file2stat[7]) {
637 print STDERR "different sized files\n" if ($verbosity >= 2);
638 return 1;
639 }
640
641 # the second file cannot be older than the first
642 if ($file1stat[9] > $file2stat[9]) {
643 print STDERR "file is older\n" if ($verbosity >= 2);
644 return 1;
645 }
646
647 return 0;
648}
649
650
651sub get_tmp_filename
652{
653 my $file_ext = shift(@_) || undef;
654
655 my $opt_dot_file_ext = "";
656 if (defined $file_ext) {
657 if ($file_ext !~ m/\./) {
658 # no dot, so needs one added in at start
659 $opt_dot_file_ext = ".$file_ext"
660 }
661 else {
662 # allow for "extensions" such as _metadata.txt to be handled
663 # gracefully
664 $opt_dot_file_ext = $file_ext;
665 }
666 }
667
668 my $tmpdir = filename_cat($ENV{'GSDLHOME'}, "tmp");
669 &mk_all_dir ($tmpdir) unless -e $tmpdir;
670
671 my $count = 1000;
672 my $rand = int(rand $count);
673 my $full_tmp_filename = &filename_cat($tmpdir, "F$rand$opt_dot_file_ext");
674
675 while (-e $full_tmp_filename) {
676 $rand = int(rand $count);
677 $full_tmp_filename = &filename_cat($tmpdir, "F$rand$opt_dot_file_ext");
678 $count++;
679 }
680
681 return $full_tmp_filename;
682}
683
684sub get_timestamped_tmp_folder
685{
686
687 my $tmp_dirname;
688 if(defined $ENV{'GSDLCOLLECTDIR'}) {
689 $tmp_dirname = $ENV{'GSDLCOLLECTDIR'};
690 } elsif(defined $ENV{'GSDLHOME'}) {
691 $tmp_dirname = $ENV{'GSDLHOME'};
692 } else {
693 return undef;
694 }
695
696 $tmp_dirname = &util::filename_cat($tmp_dirname, "tmp");
697 &util::mk_dir($tmp_dirname) if (!-e $tmp_dirname);
698
699 # add the timestamp into the path otherwise we can run into problems
700 # if documents have the same name
701 my $timestamp = time;
702 my $time_tmp_dirname = &util::filename_cat($tmp_dirname, $timestamp);
703 $tmp_dirname = $time_tmp_dirname;
704 my $i = 1;
705 while (-e $tmp_dirname) {
706 $tmp_dirname = "$time_tmp_dirname$i";
707 $i++;
708 }
709 &util::mk_dir($tmp_dirname);
710
711 return $tmp_dirname;
712}
713
714sub get_timestamped_tmp_filename_in_collection
715{
716
717 my ($input_filename, $output_ext) = @_;
718 # derive tmp filename from input filename
719 my ($tailname, $dirname, $suffix)
720 = &File::Basename::fileparse($input_filename, "\\.[^\\.]+\$");
721
722 # softlink to collection tmp dir
723 my $tmp_dirname = &util::get_timestamped_tmp_folder();
724 $tmp_dirname = $dirname unless defined $tmp_dirname;
725
726 # following two steps copied from ConvertBinaryFile
727 # do we need them?? can't use them as is, as they use plugin methods.
728
729 #$tailname = $self->SUPER::filepath_to_utf8($tailname) unless &unicode::check_is_utf8($tailname);
730
731 # URLEncode this since htmls with images where the html filename is utf8 don't seem
732 # to work on Windows (IE or Firefox), as browsers are looking for filesystem-encoded
733 # files on the filesystem.
734 #$tailname = &util::rename_file($tailname, $self->{'file_rename_method'}, "without_suffix");
735 if (defined $output_ext) {
736 $output_ext = ".$output_ext"; # add the dot
737 } else {
738 $output_ext = $suffix;
739 }
740 $output_ext= lc($output_ext);
741 my $tmp_filename = &util::filename_cat($tmp_dirname, "$tailname$output_ext");
742
743 return $tmp_filename;
744}
745
746sub get_toplevel_tmp_dir
747{
748 return filename_cat($ENV{'GSDLHOME'}, "tmp");
749}
750
751
752sub filename_to_regex {
753 my $filename = shift (@_);
754
755 # need to put single backslash back to double so that regex works
756 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
757 $filename =~ s/\\/\\\\/g;
758 }
759 return $filename;
760}
761
762sub filename_cat {
763 my $first_file = shift(@_);
764 my (@filenames) = @_;
765
766# Useful for debugging
767# -- might make sense to call caller(0) rather than (1)??
768# my ($cpackage,$cfilename,$cline,$csubr,$chas_args,$cwantarray) = caller(1);
769# print STDERR "Calling method: $cfilename:$cline $cpackage->$csubr\n";
770
771 # If first_file is not null or empty, then add it back into the list
772 if (defined $first_file && $first_file =~ /\S/) {
773 unshift(@filenames, $first_file);
774 }
775
776 my $filename = join("/", @filenames);
777
778 # remove duplicate slashes and remove the last slash
779 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
780 $filename =~ s/[\\\/]+/\\/g;
781 } else {
782 $filename =~ s/[\/]+/\//g;
783 # DB: want a filename abc\de.html to remain like this
784 }
785 $filename =~ s/[\\\/]$//;
786
787 return $filename;
788}
789
790
791sub pathname_cat {
792 my $first_path = shift(@_);
793 my (@pathnames) = @_;
794
795 # If first_path is not null or empty, then add it back into the list
796 if (defined $first_path && $first_path =~ /\S/) {
797 unshift(@pathnames, $first_path);
798 }
799
800 my $join_char;
801 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
802 $join_char = ";";
803 } else {
804 $join_char = ":";
805 }
806
807 my $pathname = join($join_char, @pathnames);
808
809 # remove duplicate slashes
810 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
811 $pathname =~ s/[\\\/]+/\\/g;
812 } else {
813 $pathname =~ s/[\/]+/\//g;
814 # DB: want a pathname abc\de.html to remain like this
815 }
816
817 return $pathname;
818}
819
820
821sub tidy_up_oid {
822 my ($OID) = @_;
823 if ($OID =~ /\./) {
824 print STDERR "Warning, identifier $OID contains periods (.), removing them\n";
825 $OID =~ s/\.//g; #remove any periods
826 }
827 if ($OID =~ /^\s.*\s$/) {
828 print STDERR "Warning, identifier $OID starts or ends with whitespace. Removing it\n";
829 # remove starting and trailing whitespace
830 $OID =~ s/^\s+//;
831 $OID =~ s/\s+$//;
832 }
833 if ($OID =~ /^[\d]*$/) {
834 print STDERR "Warning, identifier $OID contains only digits. Prepending 'D'.\n";
835 $OID = "D" . $OID;
836 }
837
838 return $OID;
839}
840sub envvar_prepend {
841 my ($var,$val) = @_;
842
843 # do not prepend any value/path that's already in the environment variable
844 if ($ENV{'GSDLOS'} =~ /^windows$/i)
845 {
846 my $escaped_val = $val;
847 $escaped_val =~ s/\\/\\\\/g; # escape any Windows backslashes for upcoming regex
848 if (!defined($ENV{$var})) {
849 $ENV{$var} = "$val";
850 }
851 elsif($ENV{$var} !~ m/$escaped_val/) {
852 $ENV{$var} = "$val;".$ENV{$var};
853 }
854 }
855 else {
856 if (!defined($ENV{$var})) {
857 $ENV{$var} = "$val";
858 }
859 elsif($ENV{$var} !~ m/$val/) {
860 $ENV{$var} = "$val:".$ENV{$var};
861 }
862 }
863}
864
865sub envvar_append {
866 my ($var,$val) = @_;
867
868 # do not append any value/path that's already in the environment variable
869 if ($ENV{'GSDLOS'} =~ /^windows$/i)
870 {
871 my $escaped_val = $val;
872 $escaped_val =~ s/\\/\\\\/g; # escape any Windows backslashes for upcoming regex
873 if (!defined($ENV{$var})) {
874 $ENV{$var} = "$val";
875 }
876 elsif($ENV{$var} !~ m/$escaped_val/) {
877 $ENV{$var} .= ";$val";
878 }
879 }
880 else {
881 if (!defined($ENV{$var})) {
882 $ENV{$var} = "$val";
883 }
884 elsif($ENV{$var} !~ m/$val/) {
885 $ENV{$var} .= ":$val";
886 }
887 }
888}
889
890
891# splits a filename into a prefix and a tail extension using the tail_re, or
892# if that fails, splits on the file_extension . (dot)
893sub get_prefix_and_tail_by_regex {
894
895 my ($filename,$tail_re) = @_;
896
897 my ($file_prefix,$file_ext) = ($filename =~ m/^(.*?)($tail_re)$/);
898 if ((!defined $file_prefix) || (!defined $file_ext)) {
899 ($file_prefix,$file_ext) = ($filename =~ m/^(.*)(\..*?)$/);
900 }
901
902 return ($file_prefix,$file_ext);
903}
904
905# get full path and file only path from a base_dir (which may be empty) and
906# file (which may contain directories)
907sub get_full_filenames {
908 my ($base_dir, $file) = @_;
909
910 my $filename_full_path = $file;
911 # add on directory if present
912 $filename_full_path = &util::filename_cat ($base_dir, $file) if $base_dir =~ /\S/;
913
914 my $filename_no_path = $file;
915
916 # remove directory if present
917 $filename_no_path =~ s/^.*[\/\\]//;
918 return ($filename_full_path, $filename_no_path);
919}
920
921# returns the path of a file without the filename -- ie. the directory the file is in
922sub filename_head {
923 my $filename = shift(@_);
924
925 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
926 $filename =~ s/[^\\\\]*$//;
927 }
928 else {
929 $filename =~ s/[^\\\/]*$//;
930 }
931
932 return $filename;
933}
934
935
936# returns 1 if filename1 and filename2 point to the same
937# file or directory
938sub filenames_equal {
939 my ($filename1, $filename2) = @_;
940
941 # use filename_cat to clean up trailing slashes and
942 # multiple slashes
943 $filename1 = filename_cat ($filename1);
944 $filename2 = filename_cat ($filename2);
945
946 # filenames not case sensitive on windows
947 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
948 $filename1 =~ tr/[A-Z]/[a-z]/;
949 $filename2 =~ tr/[A-Z]/[a-z]/;
950 }
951 return 1 if $filename1 eq $filename2;
952 return 0;
953}
954
955sub filename_within_collection
956{
957 my ($filename) = @_;
958
959 my $collect_dir = $ENV{'GSDLCOLLECTDIR'};
960
961 if (defined $collect_dir) {
962 my $dirsep = &util::get_dirsep();
963 if ($collect_dir !~ m/$dirsep$/) {
964 $collect_dir .= $dirsep;
965 }
966
967 $collect_dir =~ s/\\/\\\\/g; # escape DOS style file separator
968
969 # if from within GSDLCOLLECTDIR, then remove directory prefix
970 # so source_filename is realative to it. This is done to aid
971 # portability, i.e. the collection can be moved to somewhere
972 # else on the file system and the archives directory will still
973 # work. This is needed, for example in the applet version of
974 # GLI where GSDLHOME/collect on the server will be different to
975 # the collect directory of the remove user. Of course,
976 # GSDLCOLLECTDIR subsequently needs to be put back on to turn
977 # it back into a full pathname.
978
979 if ($filename =~ /^$collect_dir(.*)$/) {
980 $filename = $1;
981 }
982 }
983
984 return $filename;
985}
986
987sub filename_is_absolute
988{
989 my ($filename) = @_;
990
991 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
992 return ($filename =~ m/^(\w:)?\\/);
993 }
994 else {
995 return ($filename =~ m/^\//);
996 }
997}
998
999
1000## @method make_absolute()
1001#
1002# Ensure the given file path is absolute in respect to the given base path.
1003#
1004# @param $base_dir A string denoting the base path the given dir must be
1005# absolute to.
1006# @param $dir The directory to be made absolute as a string. Note that the
1007# dir may already be absolute, in which case it will remain
1008# unchanged.
1009# @return The now absolute form of the directory as a string.
1010#
1011# @author John Thompson, DL Consulting Ltd.
1012# @copy 2006 DL Consulting Ltd.
1013#
1014#used in buildcol.pl, doesn't work for all cases --kjdon
1015sub make_absolute {
1016
1017 my ($base_dir, $dir) = @_;
1018### print STDERR "dir = $dir\n";
1019 $dir =~ s/[\\\/]+/\//g;
1020 $dir = $base_dir . "/$dir" unless ($dir =~ m|^(\w:)?/|);
1021 $dir =~ s|^/tmp_mnt||;
1022 1 while($dir =~ s|/[^/]*/\.\./|/|g);
1023 $dir =~ s|/[.][.]?/|/|g;
1024 $dir =~ tr|/|/|s;
1025### print STDERR "dir = $dir\n";
1026
1027 return $dir;
1028}
1029## make_absolute() ##
1030
1031sub get_dirsep {
1032
1033 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
1034 return "\\";
1035 } else {
1036 return "\/";
1037 }
1038}
1039
1040sub get_os_dirsep {
1041
1042 if ($ENV{'GSDLOS'} =~ /^windows$/i) {
1043 return "\\\\";
1044 } else {
1045 return "\\\/";
1046 }
1047}
1048
1049sub get_re_dirsep {
1050
1051 return "\\\\|\\\/";
1052}
1053
1054
1055sub get_dirsep_tail {
1056 my ($filename) = @_;
1057
1058 # returns last part of directory or filename
1059 # On unix e.g. a/b.d => b.d
1060 # a/b/c => c
1061
1062 my $dirsep = get_re_dirsep();
1063 my @dirs = split (/$dirsep/, $filename);
1064 my $tail = pop @dirs;
1065
1066 # - caused problems under windows
1067 #my ($tail) = ($filename =~ m/^(?:.*?$dirsep)?(.*?)$/);
1068
1069 return $tail;
1070}
1071
1072
1073# if this is running on windows we want binaries to end in
1074# .exe, otherwise they don't have to end in any extension
1075sub get_os_exe {
1076 return ".exe" if $ENV{'GSDLOS'} =~ /^windows$/i;
1077 return "";
1078}
1079
1080
1081# test to see whether this is a big or little endian machine
1082sub is_little_endian
1083{
1084 # To determine the name of the operating system, the variable $^O is a cheap alternative to pulling it out of the Config module;
1085 # If it is a Macintosh machine (i.e. the Darwin operating system), regardless if it's running on the IBM power-pc cpu or the x86 Intel-based chip with a power-pc emulator running on top of it, it's big-endian
1086 # Otherwise, it's little endian
1087
1088 #return 0 if $^O =~ /^darwin$/i;
1089 #return 0 if $ENV{'GSDLOS'} =~ /^darwin$/i;
1090
1091 # Going back to stating exactly whether the machine is little endian
1092 # or big endian, without any special case for Macs. Since for rata it comes
1093 # back with little endian and for shuttle with bigendian.
1094 return (ord(substr(pack("s",1), 0, 1)) == 1);
1095}
1096
1097
1098# will return the collection name if successful, "" otherwise
1099sub use_collection {
1100 my ($collection, $collectdir) = @_;
1101
1102 if (!defined $collectdir || $collectdir eq "") {
1103 $collectdir = &filename_cat ($ENV{'GSDLHOME'}, "collect");
1104 }
1105
1106 # get and check the collection
1107 if (!defined($collection) || $collection eq "") {
1108 if (defined $ENV{'GSDLCOLLECTION'}) {
1109 $collection = $ENV{'GSDLCOLLECTION'};
1110 } else {
1111 print STDOUT "No collection specified\n";
1112 return "";
1113 }
1114 }
1115
1116 if ($collection eq "modelcol") {
1117 print STDOUT "You can't use modelcol.\n";
1118 return "";
1119 }
1120
1121 # make sure the environment variables GSDLCOLLECTION and GSDLCOLLECTDIR
1122 # are defined
1123 $ENV{'GSDLCOLLECTION'} = $collection;
1124 $ENV{'GSDLCOLLECTDIR'} = &filename_cat ($collectdir, $collection);
1125
1126 # make sure this collection exists
1127 if (!-e $ENV{'GSDLCOLLECTDIR'}) {
1128 print STDOUT "Invalid collection ($collection).\n";
1129 return "";
1130 }
1131
1132 # everything is ready to go
1133 return $collection;
1134}
1135
1136sub get_current_collection_name {
1137 return $ENV{'GSDLCOLLECTION'};
1138}
1139
1140
1141# will return the collection name if successful, "" otherwise.
1142# Like use_collection (above) but for greenstone 3 (taking account of site level)
1143
1144sub use_site_collection {
1145 my ($site, $collection, $collectdir) = @_;
1146
1147 if (!defined $collectdir || $collectdir eq "") {
1148 die "GSDL3HOME not set.\n" unless defined $ENV{'GSDL3HOME'};
1149 $collectdir = &filename_cat ($ENV{'GSDL3HOME'}, "sites", $site, "collect");
1150 }
1151
1152 # collectdir explicitly set by this point (using $site variable if required).
1153 # Can call "old" gsdl2 use_collection now.
1154
1155 return use_collection($collection,$collectdir);
1156}
1157
1158
1159
1160sub locate_config_file
1161{
1162 my ($file) = @_;
1163
1164 my $locations = locate_config_files($file);
1165
1166 return shift @$locations; # returns undef if 'locations' is empty
1167}
1168
1169
1170sub locate_config_files
1171{
1172 my ($file) = @_;
1173
1174 my @locations = ();
1175
1176 if (-e $file) {
1177 # Clearly specified (most likely full filename)
1178 # No need to hunt in 'etc' directories, return value unchanged
1179 push(@locations,$file);
1180 }
1181 else {
1182 # Check for collection specific one before looking in global GSDL 'etc'
1183 if (defined $ENV{'GSDLCOLLECTDIR'} && $ENV{'GSDLCOLLECTDIR'} ne "") {
1184 my $test_collect_etc_filename
1185 = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"etc", $file);
1186
1187 if (-e $test_collect_etc_filename) {
1188 push(@locations,$test_collect_etc_filename);
1189 }
1190 }
1191 my $test_main_etc_filename
1192 = &util::filename_cat($ENV{'GSDLHOME'},"etc", $file);
1193 if (-e $test_main_etc_filename) {
1194 push(@locations,$test_main_etc_filename);
1195 }
1196 }
1197
1198 return \@locations;
1199}
1200
1201
1202sub hyperlink_text
1203{
1204 my ($text) = @_;
1205
1206 $text =~ s/(http:\/\/[^\s]+)/<a href=\"$1\">$1<\/a>/mg;
1207 $text =~ s/(^|\s+)(www\.(\w|\.)+)/<a href=\"http:\/\/$2\">$2<\/a>/mg;
1208
1209 return $text;
1210}
1211
1212
1213# A method to check if a directory is empty (note that an empty directory still has non-zero size!!!)
1214# Code is from http://episteme.arstechnica.com/eve/forums/a/tpc/f/6330927813/m/436007700831
1215sub is_dir_empty
1216{
1217 my ($path) = @_;
1218 opendir DIR, $path;
1219 while(my $entry = readdir DIR) {
1220 next if($entry =~ /^\.\.?$/);
1221 closedir DIR;
1222 return 0;
1223 }
1224 closedir DIR;
1225 return 1;
1226}
1227
1228# Returns the given filename converted using either URL encoding or base64
1229# encoding, as specified by $rename_method. If the given filename has no suffix
1230# (if it is just the tailname), then $no_suffix should be some defined value.
1231# rename_method can be url, none, base64
1232sub rename_file {
1233 my ($filename, $rename_method, $no_suffix) = @_;
1234
1235 if(!$filename) { # undefined or empty string
1236 return $filename;
1237 }
1238
1239 if (!$rename_method) {
1240 print STDERR "WARNING: no file renaming method specified. Defaulting to using URL encoding...\n";
1241 # Debugging information
1242 # my ($cpackage,$cfilename,$cline,$csubr,$chas_args,$cwantarray) = caller(1);
1243 # print STDERR "Called from method: $cfilename:$cline $cpackage->$csubr\n";
1244 $rename_method = "url";
1245 } elsif($rename_method eq "none") {
1246 return $filename; # would have already been renamed
1247 }
1248
1249 # No longer replace spaces with underscores, since underscores mess with incremental rebuild
1250 ### Replace spaces with underscore. Do this first else it can go wrong below when getting tailname
1251 ###$filename =~ s/ /_/g;
1252
1253 my ($tailname,$dirname,$suffix);
1254 if($no_suffix) { # given a tailname, no suffix
1255 ($tailname,$dirname) = File::Basename::fileparse($filename);
1256 }
1257 else {
1258 ($tailname,$dirname,$suffix) = File::Basename::fileparse($filename, "\\.(?:[^\\.]+?)\$");
1259 }
1260 $suffix = "" if !$suffix;
1261
1262 if ($rename_method eq "url") {
1263 $tailname = &unicode::url_encode($tailname);
1264 }
1265 elsif ($rename_method eq "base64") {
1266 $tailname = &unicode::base64_encode($tailname);
1267 $tailname =~ s/\s*//sg; # for some reason it adds spaces not just at end but also in middle
1268 }
1269
1270 $filename = "$tailname$suffix";
1271 $filename = "$dirname$filename" if ($dirname ne "./" && $dirname ne ".\\");
1272
1273 return $filename;
1274}
1275
1276
1277# BACKWARDS COMPATIBILITY: Just in case there are old .ldb/.bdb files
1278sub rename_ldb_or_bdb_file {
1279 my ($filename_no_ext) = @_;
1280
1281 my $new_filename = "$filename_no_ext.gdb";
1282 return if (-f $new_filename); # if the file has the right extension, don't need to do anything
1283 # try ldb
1284 my $old_filename = "$filename_no_ext.ldb";
1285
1286 if (-f $old_filename) {
1287 print STDERR "Renaming $old_filename to $new_filename\n";
1288 rename ($old_filename, $new_filename)
1289 || print STDERR "Rename failed: $!\n";
1290 return;
1291 }
1292 # try bdb
1293 $old_filename = "$filename_no_ext.bdb";
1294 if (-f $old_filename) {
1295 print STDERR "Renaming $old_filename to $new_filename\n";
1296 rename ($old_filename, $new_filename)
1297 || print STDERR "Rename failed: $!\n";
1298 return;
1299 }
1300}
1301
1302
1303# Returns the greenstone URL prefix extracted from the appropriate GS2/GS3 config file.
1304# By default, /greenstone3 for GS3 or /greenstone for GS2.
1305sub get_greenstone_url_prefix() {
1306 # if already set on a previous occasion, just return that
1307 # (Don't want to keep repeating this: cost of re-opening and scanning files.)
1308 return $ENV{'GREENSTONE_URL_PREFIX'} if($ENV{'GREENSTONE_URL_PREFIX'});
1309
1310 my ($configfile, $urlprefix, $defaultUrlprefix);
1311 my @propertynames = ();
1312
1313 if($ENV{'GSDL3SRCHOME'}) {
1314 $defaultUrlprefix = "/greenstone3";
1315 $configfile = &util::filename_cat($ENV{'GSDL3SRCHOME'}, "packages", "tomcat", "conf", "Catalina", "localhost", "greenstone3.xml");
1316 push(@propertynames, qw/path\s*\=/);
1317 } else {
1318 $defaultUrlprefix = "/greenstone";
1319 $configfile = &util::filename_cat($ENV{'GSDLHOME'}, "cgi-bin", "gsdlsite.cfg");
1320 push(@propertynames, (qw/\nhttpprefix/, qw/\ngwcgi/)); # inspect one property then the other
1321 }
1322
1323 $urlprefix = &extract_propvalue_from_file($configfile, \@propertynames);
1324
1325 if(!$urlprefix) { # no values found for URL prefix, use default values
1326 $urlprefix = $defaultUrlprefix;
1327 } else {
1328 #gwcgi can contain more than the wanted prefix, we split on / to get the first "directory" level
1329 $urlprefix =~ s/^\///; # remove the starting slash
1330 my @dirs = split(/(\\|\/)/, $urlprefix);
1331 $urlprefix = shift(@dirs);
1332
1333 if($urlprefix !~ m/^\//) { # in all cases: ensure the required forward slash is at the front
1334 $urlprefix = "/$urlprefix";
1335 }
1336 }
1337
1338 # set for the future
1339 $ENV{'GREENSTONE_URL_PREFIX'} = $urlprefix;
1340# print STDERR "*** in get_greenstone_url_prefix(): $urlprefix\n\n";
1341 return $urlprefix;
1342}
1343
1344
1345# Given a config file (xml or java properties file) and a list/array of regular expressions
1346# that represent property names to match on, this function will return the value for the 1st
1347# matching property name. If the return value is undefined, no matching property was found.
1348sub extract_propvalue_from_file() {
1349 my ($configfile, $propertynames) = @_;
1350
1351 my $value;
1352 unless(open(FIN, "<$configfile")) {
1353 print STDERR "extract_propvalue_from_file(): Unable to open $configfile. $!\n";
1354 return $value; # not initialised
1355 }
1356
1357 # Read the entire file at once, as one single line, then close it
1358 my $filecontents;
1359 {
1360 local $/ = undef;
1361 $filecontents = <FIN>;
1362 }
1363 close(FIN);
1364
1365 foreach my $regex (@$propertynames) {
1366 ($value) = $filecontents=~ m/$regex\s*(\S*)/s; # read value of the property given by regex up to the 1st space
1367 if($value) {
1368 $value =~ s/^\"//; # remove any startquotes
1369 $value =~ s/\".*$//; # remove the 1st endquotes (if any) followed by any xml
1370 last; # found value for a matching property, break from loop
1371 }
1372 }
1373
1374 return $value;
1375}
1376
1377# Subroutine that sources setup.bash, given GSDLHOME and GSDLOS and
1378# given that perllib is in @INC in order to invoke this subroutine.
1379# Call as follows -- after setting up INC to include perllib and
1380# after setting up GSDLHOME and GSDLOS:
1381#
1382# require util;
1383# &util::setup_greenstone_env($ENV{'GSDLHOME'}, $ENV{'GSDLOS'});
1384#
1385sub setup_greenstone_env() {
1386 my ($GSDLHOME, $GSDLOS) = @_;
1387
1388 #my %env_map = ();
1389 # Get the localised ENV settings of running a localised source setup.bash
1390 # and put it into the ENV here. Need to clear GSDLHOME before running setup
1391 #my $perl_command = "(cd $GSDLHOME; export GSDLHOME=; . ./setup.bash > /dev/null; env)";
1392 my $perl_command = "(cd $GSDLHOME; /bin/bash -c \"export GSDLHOME=; source setup.bash > /dev/null; env\")";
1393 if($GSDLOS =~ m/windows/i) {
1394 #$perl_command = "cmd /C \"cd $GSDLHOME&& set GSDLHOME=&& setup.bat > nul&& set\"";
1395 $perl_command = "(cd $GSDLHOME&& set GSDLHOME=&& setup.bat > nul&& set)";
1396 }
1397 if (!open(PIN, "$perl_command |")) {
1398 print STDERR ("Unable to execute command: $perl_command. $!\n");
1399 }
1400
1401 while (defined (my $perl_output_line = <PIN>)) {
1402 my($key,$value) = ($perl_output_line =~ m/^([^=]*)[=](.*)$/);
1403 #$env_map{$key}=$value;
1404 $ENV{$key}=$value;
1405 }
1406
1407 # If any keys in $ENV don't occur in Greenstone's localised env
1408 # (stored in $env_map), delete those entries from $ENV
1409 #foreach $key (keys %ENV) {
1410 # if(!defined $env_map{$key}) {
1411 # print STDOUT "**** DELETING ENV KEY: $key\tVALUE: $ENV{'$key'}\n";
1412 # delete $ENV{$key}; # del $ENV(key, value) pair
1413 # }
1414 #}
1415 #undef %env_map;
1416}
1417
14181;
Note: See TracBrowser for help on using the repository browser.