source: main/trunk/greenstone2/bin/script/build@ 31953

Last change on this file since 31953 was 31953, checked in by ak19, 7 years ago

Bugfix to allow Depositor to still successfully build even when GS2 is installed in a path containing spaces
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Revision
File size: 26.8 KB
Line 
1#!/usr/bin/perl -w
2
3###########################################################################
4#
5# build --
6# A component of the Greenstone digital library software
7# from the New Zealand Digital Library Project at the
8# University of Waikato, New Zealand.
9#
10# Copyright (C) 2000 New Zealand Digital Library Project
11#
12# This program is free software; you can redistribute it and/or modify
13# it under the terms of the GNU General Public License as published by
14# the Free Software Foundation; either version 2 of the License, or
15# (at your option) any later version.
16#
17# This program is distributed in the hope that it will be useful,
18# but WITHOUT ANY WARRANTY; without even the implied warranty of
19# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20# GNU General Public License for more details.
21#
22# You should have received a copy of the GNU General Public License
23# along with this program; if not, write to the Free Software
24# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
25#
26###########################################################################
27
28# This perl script may be called directly or by running build.bat on
29# windows (build.bat is in bin\windows)
30
31# Note that this script has grown over time and now has many options for
32# use when called from within the collector. If it appears to
33# over-complicate things a little, that's why.
34
35# I think that currently (Nov 2010) its only officially used from the collector
36# and depositor
37
38package build;
39
40use strict;
41no strict 'refs';
42
43use FileHandle;
44use File::Copy;
45
46BEGIN {
47
48 die "GSDLHOME not set - did you remember to source setup.bash (unix) or " .
49 "run setup.bat (windows)?\n" unless defined $ENV{'GSDLHOME'};
50 die "GSDLOS not set - did you remember to source setup.bash (unix) or " .
51 "run setup.bat (windows)?\n" unless defined $ENV{'GSDLOS'};
52 unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
53
54 STDOUT->autoflush(1);
55 STDERR->autoflush(1);
56}
57
58use lib qq($ENV{'GSDLHOME'}/perllib/cpan);
59use Mail::Sendmail;
60use parsargv;
61use util;
62use FileUtils;
63use cfgread;
64use colcfg;
65use dbutil;
66
67# set up greenstone environment
68&util::setup_greenstone_env($ENV{'GSDLHOME'}, $ENV{'GSDLOS'});
69# is the following needed with the previous line??
70# set up path - this allows for paths not to be supplied to system calls
71# and overcomes problems when GSDLHOME contains spaces (double quoting
72# the call doesn't work on win2k and probably other variants of winnt)
73my $path_separator = ":";
74$path_separator = ";" if $ENV{'GSDLOS'} =~ /^windows$/;
75$ENV{'PATH'} = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}) .
76 $path_separator . &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", "script") .
77 $path_separator . $ENV{'PATH'};
78
79# all the input option variables
80my ($optionfile, $indextype, $append, $manifest, $remove_archives, $remove_import, $buildtype, $infodbtype, $maxdocs, @download, $collectdir, $site, $dontinstall, $save_archives, $out, $make_writable, $log_events, $event_log_file, $email_events, $mail_server, $statsfile, $event_header);
81
82&parse_args (\@ARGV);
83
84my ($collection) = @ARGV;
85
86if (!defined $collection || $collection !~ /\w/) {
87 print STDERR "You must specify a collection to build\n";
88 &print_usage();
89 die "\n";
90}
91
92if ($optionfile =~ /\w/) {
93 open (OPTIONS, $optionfile) || die "Couldn't open $optionfile\n";
94 my $line = [];
95 my $options = [];
96 while (defined ($line = &cfgread::read_cfg_line ('build::OPTIONS'))) {
97 push (@$options, @$line);
98 }
99 close OPTIONS;
100 &parse_args ($options);
101}
102
103if ($maxdocs == -1) {
104 $maxdocs = "";
105} else {
106 $maxdocs = "-maxdocs $maxdocs";
107}
108
109my $cdir = $collectdir;
110my $gs_mode;
111if (defined $site && $site =~ /\w/)
112{
113 die "GSDL3HOME not set." unless $ENV{'GSDL3HOME'};
114 $cdir = &FileUtils::filenameConcatenate ($ENV{'GSDL3HOME'}, "sites", $site, "collect") unless $collectdir =~ /\w/;
115 $gs_mode = "gs3";
116}
117else
118{
119 $cdir = &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "collect") unless $collectdir =~ /\w/;
120 $gs_mode = "gs2";
121}
122
123my $importdir = &FileUtils::filenameConcatenate ($cdir, $collection, "import");
124my $archivedir = &FileUtils::filenameConcatenate ($cdir, $collection, "archives");
125my $buildingdir = &FileUtils::filenameConcatenate ($cdir, $collection, "building");
126my $indexdir = &FileUtils::filenameConcatenate ($cdir, $collection, "index");
127my $etcdir = &FileUtils::filenameConcatenate ($cdir, $collection, "etc");
128my $bindir = &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "bin");
129
130my $use_out = 0;
131my $outfile = $out;
132if ($out !~ /^(STDERR|STDOUT)$/i) {
133 open (OUT, ">$out") || die "Couldn't open output file $out\n";
134 $out = "OUT";
135
136 # delete any existing .final file
137 &FileUtils::removeFiles ("$outfile.final") if -e "$outfile.final";
138
139 $use_out = 1;
140}
141$out->autoflush(1);
142
143# delete any .kill file left laying around from a previously aborted build
144if (-e &FileUtils::filenameConcatenate ($cdir, $collection, ".kill")) {
145 &FileUtils::removeFiles (&FileUtils::filenameConcatenate ($cdir, $collection, ".kill"));
146}
147
148# get maintainer email address from main.cfg
149my $maintainer = "NULL";
150my $main_cfg = &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "etc", "main.cfg");
151my $cfgdata = &cfgread::read_cfg_file ($main_cfg, "maintainer");
152if (defined $cfgdata->{'maintainer'} && $cfgdata->{'maintainer'} =~ /\w/) {
153 $maintainer = $cfgdata->{'maintainer'};
154}
155# if maintainer is "NULL" email_events should be disabled
156if ($maintainer =~ /^NULL$/i) {
157 $email_events = "";
158}
159
160&main();
161
162if ($use_out) {
163 close OUT;
164
165 # if we've created a build log we'll copy it to the collection's etc
166 # directory
167 my ($final_etcdir);
168 if ($dontinstall) {
169 $final_etcdir = &FileUtils::filenameConcatenate($collectdir, $collection, "etc", "build.log");
170 } else {
171 $final_etcdir = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "collect", $collection, "etc", "build.log");
172 }
173
174 &FileUtils::copyFiles($outfile, $final_etcdir);
175}
176
177sub print_usage {
178 print STDOUT "\n";
179 print STDOUT "build: Builds a Greenstone collection (i.e. runs import.pl and buildcol.pl\n";
180 print STDOUT " then copies the resulting indexes to the correct place).\n\n";
181 print STDOUT " usage: $0 [options] collection-name\n\n";
182 print STDOUT " options:\n";
183 print STDOUT " -optionfile file Get options from file, useful on systems where\n";
184 print STDOUT " long command lines may cause problems\n";
185 print STDOUT " -indextype mg|mgpp|lucene \n";
186 print STDERR " Specify the type of indexer used in this collection\n";
187 print STDERR " If -append is used then -indextype is needed to \n";
188 print STDERR " determine how to run buildcol.pl as well as update\n";
189 print STDERR " 'building' and 'index' according.\n";
190 print STDOUT " -append Add new files to existing collection\n";
191 print STDOUT " -manifest Use manifest.xml file to determine which files to process.\n";
192 print STDOUT " -remove_archives Remove archives directory after successfully\n";
193 print STDOUT " building the collection.\n";
194 print STDOUT " -remove_import Remove import directory after successfully\n";
195 print STDOUT " importing the collection.\n";
196 print STDOUT " -buildtype build|import If 'build' attempt to build directly\n";
197 print STDOUT " from archives directory (bypassing import\n";
198 print STDOUT " stage). Defaults to 'import'\n";
199 print STDOUT " -maxdocs number Maximum number of documents to build\n";
200 print STDOUT " -download directory Directory (or file) to get import documents from.\n";
201 print STDOUT " There may be multiple download directories and they\n";
202 print STDOUT " may be of type http://, ftp://, or file://\n";
203 print STDOUT " Note that any existing import directory will be\n";
204 print STDOUT " deleted to make way for the downloaded data if\n";
205 print STDOUT " a -download option is supplied\n";
206 print STDOUT " -collectdir directory Collection directory (defaults to " .
207 &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}). "collect for Greenstone2;\n";
208 print STDOUT" for Greenstone3 use -site option and then collectdir default will be\n";
209 print STDOUT " set to the collect folder within that site.)\n";
210 print STDOUT " -site Specify the site within a Greenstone3 installation to use.\n";
211 print STDOUT " -dontinstall Only applicable if -collectdir is set to something\n";
212 print STDOUT " other than the default. -dontinstall will suppress\n";
213 print STDOUT " the default behaviour which is to install the\n";
214 print STDOUT " collection to the gsdl/collect directory once it has\n";
215 print STDOUT " been built.\n";
216 print STDOUT " -save_archives Create a copy of the existing archives directory\n";
217 print STDOUT " called archives.org\n";
218 print STDOUT " -out Filename or handle to print output status to.\n";
219 print STDOUT " The default is STDERR\n";
220 print STDOUT " -statsfile name Filename or handle to print import statistics to.\n";
221 print STDOUT " The default is STDERR\n";
222 print STDOUT " -make_writable If set build will make the collection and any\n";
223 print STDOUT " temporary files it created globally writable after\n";
224 print STDOUT " it finishes\n";
225 print STDOUT " -log_events Log important events (collection built successfully\n";
226 print STDOUT " etc.) to event_log_file\n";
227 print STDOUT " -event_log_file file File to append important events to (defaults to\n";
228 print STDOUT " " . &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "etc", "events.txt") . "\n";
229 print STDOUT " -email_events addr Comma separated list of email addresses to mail\n";
230 print STDOUT " details of important collection building events\n";
231 print STDOUT " -mail_server server The outgoing (SMTP) mail server to be used by\n";
232 print STDOUT " email_events. email_events will be disabled if\n";
233 print STDOUT " mail_server isn't set\n";
234 print STDOUT " -event_header file File containing a header to go on any event\n";
235 print STDOUT " messages. If not specified build will create a\n";
236 print STDOUT " generic header\n\n";
237 print STDOUT " [Type \"build | more\" if this help text scrolled off your screen]";
238 print STDOUT "\n" unless $ENV{'GSDLOS'} =~ /^windows$/i;
239}
240
241sub main {
242
243 if ($save_archives && -d $archivedir) {
244 print $out "caching original archives to ${archivedir}.org\n";
245 &FileUtils::copyFilesRecursive ($archivedir, "${archivedir}.org");
246 }
247
248 # do the download thing if we have any -download options
249 if (scalar (@download)) {
250 # remove any existing import data
251 if (&has_content ($importdir)) {
252 #print $out "build: WARNING: removing contents of $importdir\n";
253 #&FileUtils::removeFilesRecursive ($importdir);
254 }
255
256 foreach my $download_dir (@download) {
257
258 # remove any leading or trailing whitespace from filenames (just in case)
259 $download_dir =~ s/^\s+//;
260 $download_dir =~ s/\s+$//;
261
262 if ($download_dir =~ /^(http|ftp):\/\//) {
263 # use wget to mirror http or ftp urls
264 # options used are:
265 # -P = the directory to download documents to
266 # -np = don't ascend to parent directories. this means that only documents
267 # that live in the same directory or below on the same server as
268 # the given url will be downloaded
269 # -nv = not too verbose
270 # -r = recursively mirror
271 # -N = use time-stamping to see if an up-to-date local copy of each
272 # file already exists. this may be useful if wget fails and
273 # is restarted
274 # -l inf = infinite recursion depth
275 # -R "*\?*" = don't download cgi based urls
276 # -o = the output file to write download status to (only used if the -out
277 # option was given to build)
278
279 my $download_cmd = "\"".&util::get_perl_exec()."\" -S gsWget.pl -P \"$importdir\" -np -nv";
280 $download_cmd .= " -r -N -l inf -R \"*\\?*\"";
281 $download_cmd .= " -o \"$outfile.download\"" if $use_out;
282 $download_cmd .= " \"$download_dir\"";
283 system ($download_cmd);
284
285 # note that wget obeys the robot rules. this means that it will have
286 # downloaded a robots.txt file if one was present. since it's unlikely
287 # anyone really wants to include it in a collection we'll delete it.
288 # robots.txt shouldn't be more than two directories deep (I think it will
289 # always be exactly two deep but will look for it in the top directory too)
290 # so that's as far as we'll go looking for it.
291 if (opendir (DIR, $importdir)) {
292 my @files = readdir DIR;
293 closedir DIR;
294 foreach my $file (@files) {
295 next if $file =~ /^\.\.?$/;
296 if ($file =~ /^robots.txt$/i) {
297 &FileUtils::removeFiles (&FileUtils::filenameConcatenate ($importdir, $file));
298 last;
299 } else {
300 $file = &FileUtils::filenameConcatenate ($importdir, $file);
301 if (-d $file) {
302 if (opendir (DIR, $file)) {
303 my @files2 = readdir DIR;
304 closedir DIR;
305 foreach my $file2 (@files2) {
306 if ($file2 =~ /^robots.txt$/i) {
307 &FileUtils::removeFiles (&FileUtils::filenameConcatenate ($file, $file2));
308 last;
309 }
310 }
311 }
312 }
313 }
314 }
315 }
316
317 # if using output directory append the file download output to it
318 &append_file ($out, "$outfile.download");
319
320 } else {
321 # we assume anything not beginning with http:// or ftp://
322 # is a file or directory on the local file system.
323 $download_dir =~ s/^file:(\/\/)?//;
324 $download_dir =~ s/^\s+//; # may be whitespace between "file://" and the rest
325
326 if (-e $download_dir) {
327 # copy download_dir and all it contains to the import directory
328 #Config{perlpath}, like $^X, is a special variable containing the full path to the current perl executable we are in
329 my $download_cmd = "\"".&util::get_perl_exec()."\" -S filecopy.pl";
330 $download_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
331 $download_cmd .= " -site \"$site\"" if $site =~ /\w/;
332 $download_cmd .= " -out \"$outfile.download\"" if $use_out;
333 $download_cmd .= " \"" . $download_dir . "\" " . $collection;
334
335 my $download_status = system ($download_cmd);
336 if ($download_status > 0)
337 {
338 die "Failed to execute: $download_cmd\n";
339 }
340
341
342 # if using output directory append the file download output to it
343 &append_file ($out, "$outfile.download");
344 } else {
345 print $out "WARNING: '$download_dir' does not exist\n";
346 }
347 }
348 }
349 }
350
351 `echo $importdir ; ls $importdir `;
352
353 my $col_cfg_file;
354 if ($gs_mode eq "gs3") {
355 $col_cfg_file = &FileUtils::filenameConcatenate($etcdir, "collectionConfig.xml");
356 } else {
357 $col_cfg_file = &FileUtils::filenameConcatenate($etcdir, "collect.cfg");
358 }
359
360 my $collect_cfg = &colcfg::read_collection_cfg ($col_cfg_file, $gs_mode);
361 # get the database type for this collection from its configuration file (may be undefined)
362 $infodbtype = $collect_cfg->{'infodbtype'} || &dbutil::get_default_infodb_type();
363
364 my $archiveinf_doc_file_path = &dbutil::get_infodb_file_path($infodbtype, "archiveinf-doc", $archivedir);
365 if (-e $archiveinf_doc_file_path) {
366 if (&has_content ($importdir)) {
367 if ($buildtype eq "build") {
368 &gsdl_build();
369 } else {
370 &gsdl_import();
371 &gsdl_build();
372 }
373 } else {
374 # there are archives but no import, build directly from archives
375 print $out "build: no import material was found, building directly\n";
376 print $out " from archives\n";
377 &gsdl_build();
378 }
379 } else {
380 if (&has_content ($importdir)) {
381 if ($buildtype eq "build") {
382 print $out "build: can't build directly from archives as no\n";
383 print $out " imported archives exist (did you forget to\n";
384 print $out " move the contents of $collection/import to\n";
385 print $out " collection/archives?)\n";
386 }
387 &gsdl_import();
388 if (&has_content ($archivedir, "^archiveinf-doc\..*\$")) {
389 &gsdl_build();
390 } else {
391 my $msg = "build: ERROR: The collection could not be built as no\n";
392 $msg .= " valid data was imported. Are at least some of\n";
393 $msg .= " the files you imported in a format that can be\n";
394 $msg .= " processed by the specified Greenstone plugins?\n";
395 print $out $msg;
396 &log_event ($msg);
397 &final_out (6) if $use_out;
398 die "\n";
399 }
400 } else {
401 # no import or archives
402 my $msg = "build: ERROR: The collection could not be built as it contains no data.\n";
403 print $out $msg;
404 &log_event ($msg);
405 &final_out (1) if $use_out;
406 die "\n";
407 }
408 }
409
410 if ($collectdir ne "" && !$dontinstall) {
411
412 my $install_collectdir;
413 if (defined $ENV{'GSDL3HOME'})
414 {
415
416 if ((defined $site) && ($site ne ""))
417 {
418 $install_collectdir = &FileUtils::filenameConcatenate ($ENV{'GSDL3HOME'}, "sites", $site, "collect");
419 }
420 else
421 {
422 my $msg = "build: ERROR: Need to specify the site within the Greenstone3 installation.";
423 print $out $msg;
424 &log_event ($msg);
425 &final_out (6) if $use_out;
426 die "\n";
427 }
428 }
429 else
430 {
431 $install_collectdir = &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "collect");
432 }
433
434 if (!&util::filenames_equal ($collectdir, $install_collectdir)) {
435
436 # install collection to gsdl/collect
437 print $out "installing the $collection collection\n";
438 my $newdir = &FileUtils::filenameConcatenate ($install_collectdir, $collection);
439 my $olddir = &FileUtils::filenameConcatenate ($collectdir, $collection);
440 if (-d $newdir) {
441 my $msg = "build: Could not install collection as $newdir\n" .
442 " already exists. Collection will remain at\n$olddir\n";
443
444 print $out $msg;
445 &log_event ($msg);
446 &final_out (4) if $use_out;
447 die "\n";
448 }
449 if (!&File::Copy::move ($olddir, $newdir)) {
450 my $msg = "build: Failed to install collection to $newdir\n" .
451 " Collection will remain at $olddir\n";
452 print $out $msg;
453 &log_event ($msg);
454 &final_out (5) if $use_out;
455 die "\n";
456 }
457 }
458 }
459
460 &log_event ("The $collection collection was built successfully\n");
461 &final_out (0) if $use_out;
462}
463
464sub gsdl_import {
465
466 print $out "importing the $collection collection\n\n";
467
468 #Config{perlpath}, like $^X, is a special variable containing the full path to the current perl executable we are in
469 my $import_cmd = "\"".&util::get_perl_exec()."\" -S import.pl";
470 $import_cmd .= " -out \"$outfile.import\"" if $use_out;
471 if ($append) {
472 $import_cmd .= " -keepold";
473 if (not $manifest) {
474 # if we are appending, with no manifest, assume incremental
475 $import_cmd .= " -incremental";
476 }
477 } else {
478 $import_cmd .= " -removeold";
479 }
480
481 $import_cmd .= " -manifest manifest.xml" if ($manifest);
482 $import_cmd .= " -site \"$site\"" if $site =~ /\w/;
483 $import_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
484 $import_cmd .= " -statsfile \"$statsfile\"" if $statsfile =~ /\w/;
485 $import_cmd .= " $maxdocs $collection";
486
487 system ($import_cmd);
488 # if using output directory append the import output to it
489 &append_file ($out, "$outfile.import");
490
491 my $archiveinf_doc_file_path = &dbutil::get_infodb_file_path($infodbtype, "archiveinf-doc", $archivedir);
492
493 if (-e $archiveinf_doc_file_path) {
494 print $out "$collection collection imported successfully\n\n";
495 if ($remove_import) {
496 print $out "removing import directory ($importdir)\n";
497 &FileUtils::removeFilesRecursive ($importdir);
498 }
499 } else {
500 print $out "$archiveinf_doc_file_path not found. archives contents:\n";
501 print $out `ls $archivedir`;
502
503 my $msg = "build: ERROR: import.pl failed\n";
504 print $out "\n$msg";
505 &log_event ($msg);
506 &final_out (2) if $use_out;
507 die "\n";
508 }
509}
510
511sub gsdl_build {
512
513 print $out "building the $collection collection\n\n";
514
515 #Config{perlpath}, like $^X, is a special variable containing the full path to the current perl executable we are in
516 my $build_cmd = "\"".&util::get_perl_exec()."\" -S buildcol.pl";
517
518 my $removeold = 1;
519 if ($append) {
520 if ($indextype eq "lucene") {
521 $build_cmd .= " -keepold -incremental -builddir \"$indexdir\"";
522 $removeold = 0;
523 }
524 else {
525 $build_cmd .= " -removeold";
526 }
527 }
528 else {
529 $build_cmd .= " -removeold";
530 }
531
532 $build_cmd .= " -out \"$outfile.build\"" if $use_out;
533 $build_cmd .= " -site \"$site\"" if $site =~ /\w/;
534 $build_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
535 $build_cmd .= " $maxdocs $collection";
536
537 ##print $out "Running command: |$build_cmd|\n\n";
538
539 system ($build_cmd);
540 # if using output directory append the buildcol output to it
541 &append_file ($out, "$outfile.build");
542
543 my @db_exts = ( ".ldb", ".bdb", ".gdb", ".db" );
544 my $build_ok = 0;
545 foreach my $db_ext (@db_exts) {
546 if ($removeold && (-e &FileUtils::filenameConcatenate ($buildingdir, "text", "$collection$db_ext"))) {
547 $build_ok = 1;
548 last;
549 }
550 if (($removeold==0) && (-e &FileUtils::filenameConcatenate ($indexdir, "text", "$collection$db_ext"))) {
551 $build_ok = 1;
552 last;
553 }
554 }
555
556 if ($build_ok) {
557 print $out "$collection collection built successfully\n\n";
558 if ($remove_archives) {
559 print $out "removing archives directory ($archivedir)\n";
560 &FileUtils::removeFilesRecursive ($archivedir);
561 }
562 } else {
563 my $msg = "build: ERROR: buildcol.pl failed\n";
564 print $out "\n$msg";
565 &log_event ($msg);
566 &final_out (3) if $use_out;
567 die "\n";
568 }
569
570 if ($removeold) {
571 # replace old indexes with new ones
572 if (&has_content ($indexdir)) {
573 print $out "removing old indexes\n";
574 &FileUtils::removeFilesRecursive ($indexdir);
575 }
576 rmdir ($indexdir) if -d $indexdir;
577 &File::Copy::move ($buildingdir, $indexdir);
578 }
579 else {
580 # Do nothing. We have built into index dir rather than building dir
581 }
582
583 # remove the cached archives
584 if ($save_archives && -d "${archivedir}.org") {
585 &FileUtils::removeFilesRecursive ("${archivedir}.org");
586 }
587}
588
589# return 1 if $dir directory contains any files or sub-directories (other
590# than those specified in the $ignore regular expression)
591sub has_content {
592 my ($dir, $ignore) = @_;
593
594 if (!-d $dir) {return 0;}
595
596 opendir (DIR, $dir) || return 0;
597 my @files = readdir DIR;
598 closedir DIR;
599
600 foreach my $file (@files) {
601 if ($file !~ /^\.{1,2}$/) {
602 return 1 unless (defined $ignore && $file =~ /$ignore/);
603 }
604 }
605
606 return 0;
607}
608
609sub append_file {
610 my ($handle, $file) = @_;
611
612 open (FILE, $file) || return;
613 undef $/;
614 print $handle <FILE>;
615 $/ = "\n";
616 close FILE;
617 &FileUtils::removeFiles ($file);
618}
619
620# creates a file called $outfile.final and writes an output code to it.
621# An output code of 0 specifies that there was no error
622sub final_out {
623 my ($exit_code) = @_;
624
625 if ($use_out && (!-e "$outfile.final")) {
626
627 if (open (FINAL, ">$outfile.final")) {
628 print FINAL $exit_code;
629 close FINAL;
630 }
631 }
632}
633
634sub log_event {
635 my ($msg) = @_;
636
637 return unless ($log_events || $email_events);
638
639 # get the event header
640 my $eheader = "[Build Event]\n";
641 $eheader .= "Date: " . scalar localtime() . "\n";
642 if ($event_header ne "" && open (HEADER, $event_header)) {
643 undef $/;
644 $eheader .= <HEADER>;
645 $/ = "\n";
646 close HEADER;
647 } else {
648 $eheader .= "Collection: $collection\n";
649 $eheader .= "GSDLHOME: $ENV{'GSDLHOME'}\n";
650 $eheader .= "Build Location: $collectdir\n";
651 }
652
653 if ($log_events) {
654 my $fail = 0;
655 # append the event to the event log file
656 if ($event_log_file eq "" || !open (LOG, ">>$event_log_file")) {
657 # log file defaults to $GSDLHOME/etc/events.txt
658 $event_log_file = &FileUtils::filenameConcatenate ($ENV{'GSDLHOME'}, "etc", "events.txt");
659 if (!open (LOG, ">>$event_log_file")) {
660 print $out "build: ERROR: Couldn't open event log file $event_log_file\n";
661 $fail = 1;
662 }
663 }
664 if (!$fail) {
665 print LOG $eheader;
666 print LOG $msg;
667 print LOG "\n";
668 close LOG;
669 }
670 }
671
672 if ($email_events) {
673 # if mail_server isn't set email_events does nothing
674 if ($mail_server eq "") {
675 print $out "build: WARNING: mail_server was not set - email_events option was ignored\n";
676 return;
677 }
678
679 my %mail = ('SMTP' => $mail_server,
680 'To' => $email_events,
681 'From' => $maintainer,
682 'Subject' => 'Greenstone Build Event'
683 );
684 $mail{'Message'} = $eheader . $msg;
685
686 if (!sendmail %mail) {
687 print $out "build: ERROR sending mail to $email_events\n";
688 print $out "'$Mail::Sendmail::error'\n";
689 }
690 }
691}
692
693
694sub parse_args {
695 my ($argref) = @_;
696
697 if (!parsargv::parse($argref,
698 'optionfile/.*/', \$optionfile,
699 'indextype/^(mg|mgpp|lucene)$/mg', \$indextype,
700 'append', \$append,
701 'manifest', \$manifest,
702 'remove_archives', \$remove_archives,
703 'remove_import', \$remove_import,
704 'buildtype/^(build|import)$/import', \$buildtype,
705 'maxdocs/^\-?\d+/-1', \$maxdocs,
706 'download/.+', \@download,
707 'collectdir/.*/', \$collectdir,
708 'site/.*/', \$site,
709 'dontinstall', \$dontinstall,
710 'save_archives', \$save_archives,
711 'out/.*/STDERR', \$out,
712 'make_writable', \$make_writable,
713 'log_events', \$log_events,
714 'event_log_file/.*/', \$event_log_file,
715 'email_events/.*/', \$email_events,
716 'mail_server/.*/', \$mail_server,
717 'statsfile/.*/STDERR', \$statsfile,
718 'event_header/.*/', \$event_header)) {
719
720 &print_usage();
721 die "\n";
722 }
723}
724
725
726END {
727
728 if ($make_writable) {
729 # chmod a+rw new collection
730 my $installed_collection = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "collect", $collection);
731 &recursive_chmod($installed_collection);
732
733 # chmod a+rw anything we've left laying about in the tmp directory
734 if (($collectdir ne "") &&
735 (!&util::filenames_equal ($collectdir, &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "collect")))) {
736 &recursive_chmod($collectdir);
737 }
738 }
739
740 # this will produce a .final file if one doesn't exist yet - that
741 # should only happen if there's been an error somewhere in the perl
742 # code
743 &final_out(7);
744
745 sub recursive_chmod {
746 my ($dir) = @_;
747 return unless -d $dir;
748
749 chmod (0777, $dir);
750
751 opendir (DIR, $dir) || die;
752 my @files = readdir DIR;
753 closedir DIR;
754
755 foreach my $file (@files) {
756 next if $file =~ /^\.\.?$/;
757 $file = &FileUtils::filenameConcatenate($dir, $file);
758 if (-d $file) {
759 &recursive_chmod ($file);
760 } else {
761 chmod (0777, $file);
762 }
763 }
764 }
765}
Note: See TracBrowser for help on using the repository browser.