| 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 | package build;
|
|---|
| 36 |
|
|---|
| 37 | use FileHandle;
|
|---|
| 38 | use File::Copy;
|
|---|
| 39 |
|
|---|
| 40 | BEGIN {
|
|---|
| 41 |
|
|---|
| 42 | die "GSDLHOME not set - did you remember to source setup.bash (unix) or " .
|
|---|
| 43 | "run setup.bat (windows)?\n" unless defined $ENV{'GSDLHOME'};
|
|---|
| 44 | die "GSDLOS not set - did you remember to source setup.bash (unix) or " .
|
|---|
| 45 | "run setup.bat (windows)?\n" unless defined $ENV{'GSDLOS'};
|
|---|
| 46 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
|---|
| 47 |
|
|---|
| 48 | STDOUT->autoflush(1);
|
|---|
| 49 | STDERR->autoflush(1);
|
|---|
| 50 | }
|
|---|
| 51 |
|
|---|
| 52 | use lib qq($ENV{'GSDLHOME'}/perllib/cpan);
|
|---|
| 53 | use Mail::Sendmail;
|
|---|
| 54 | use parsargv;
|
|---|
| 55 | use util;
|
|---|
| 56 | use cfgread;
|
|---|
| 57 |
|
|---|
| 58 | # set up path - this allows for paths not to be supplied to system calls
|
|---|
| 59 | # and overcomes problems when GSDLHOME contains spaces (double quoting
|
|---|
| 60 | # the call doesn't work on win2k and probably other variants of winnt)
|
|---|
| 61 | my $path_separator = ":";
|
|---|
| 62 | $path_separator = ";" if $ENV{'GSDLOS'} =~ /^windows$/;
|
|---|
| 63 | $ENV{'PATH'} = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}) .
|
|---|
| 64 | $path_separator . &util::filename_cat($ENV{'GSDLHOME'}, "bin", "script") .
|
|---|
| 65 | $path_separator . $ENV{'PATH'};
|
|---|
| 66 |
|
|---|
| 67 | &parse_args (\@ARGV);
|
|---|
| 68 |
|
|---|
| 69 | my ($collection) = @ARGV;
|
|---|
| 70 |
|
|---|
| 71 | if (!defined $collection || $collection !~ /\w/) {
|
|---|
| 72 | print STDERR "You must specify a collection to build\n";
|
|---|
| 73 | &print_usage();
|
|---|
| 74 | die "\n";
|
|---|
| 75 | }
|
|---|
| 76 |
|
|---|
| 77 | if ($optionfile =~ /\w/) {
|
|---|
| 78 | open (OPTIONS, $optionfile) || die "Couldn't open $optionfile\n";
|
|---|
| 79 | my $line = [];
|
|---|
| 80 | my $options = [];
|
|---|
| 81 | while (defined ($line = &cfgread::read_cfg_line ('build::OPTIONS'))) {
|
|---|
| 82 | push (@$options, @$line);
|
|---|
| 83 | }
|
|---|
| 84 | close OPTIONS;
|
|---|
| 85 | &parse_args ($options);
|
|---|
| 86 | }
|
|---|
| 87 |
|
|---|
| 88 | if ($maxdocs == -1) {
|
|---|
| 89 | $maxdocs = "";
|
|---|
| 90 | } else {
|
|---|
| 91 | $maxdocs = "-maxdocs $maxdocs";
|
|---|
| 92 | }
|
|---|
| 93 |
|
|---|
| 94 | my $cdir = $collectdir;
|
|---|
| 95 | $cdir = &util::filename_cat ($ENV{'GSDLHOME'}, "collect") unless $collectdir =~ /\w/;
|
|---|
| 96 | my $importdir = &util::filename_cat ($cdir, $collection, "import");
|
|---|
| 97 | my $archivedir = &util::filename_cat ($cdir, $collection, "archives");
|
|---|
| 98 | my $buildingdir = &util::filename_cat ($cdir, $collection, "building");
|
|---|
| 99 | my $indexdir = &util::filename_cat ($cdir, $collection, "index");
|
|---|
| 100 | my $bindir = &util::filename_cat ($ENV{'GSDLHOME'}, "bin");
|
|---|
| 101 |
|
|---|
| 102 | my $use_out = 0;
|
|---|
| 103 | my $outfile = $out;
|
|---|
| 104 | if ($out !~ /^(STDERR|STDOUT)$/i) {
|
|---|
| 105 | open (OUT, ">$out") || die "Couldn't open output file $out\n";
|
|---|
| 106 | $out = "OUT";
|
|---|
| 107 |
|
|---|
| 108 | # delete any existing .final file
|
|---|
| 109 | &util::rm ("$outfile.final") if -e "$outfile.final";
|
|---|
| 110 |
|
|---|
| 111 | $use_out = 1;
|
|---|
| 112 | }
|
|---|
| 113 | $out->autoflush(1);
|
|---|
| 114 |
|
|---|
| 115 | # delete any .kill file left laying around from a previously aborted build
|
|---|
| 116 | if (-e &util::filename_cat ($cdir, $collection, ".kill")) {
|
|---|
| 117 | &util::rm (&util::filename_cat ($cdir, $collection, ".kill"));
|
|---|
| 118 | }
|
|---|
| 119 |
|
|---|
| 120 | # get maintainer email address from main.cfg
|
|---|
| 121 | my $maintainer = "NULL";
|
|---|
| 122 | my $main_cfg = &util::filename_cat ($ENV{'GSDLHOME'}, "etc", "main.cfg");
|
|---|
| 123 | my $cfgdata = &cfgread::read_cfg_file ($main_cfg, "maintainer");
|
|---|
| 124 | if (defined $cfgdata->{'maintainer'} && $cfgdata->{'maintainer'} =~ /\w/) {
|
|---|
| 125 | $maintainer = $cfgdata->{'maintainer'};
|
|---|
| 126 | }
|
|---|
| 127 | # if maintainer is "NULL" email_events should be disabled
|
|---|
| 128 | if ($maintainer =~ /^NULL$/i) {
|
|---|
| 129 | $email_events = "";
|
|---|
| 130 | }
|
|---|
| 131 |
|
|---|
| 132 | &main();
|
|---|
| 133 |
|
|---|
| 134 | close OUT if $use_out;
|
|---|
| 135 |
|
|---|
| 136 | sub print_usage {
|
|---|
| 137 | print STDERR "\n usage: $0 [options] collection-name\n\n";
|
|---|
| 138 | print STDERR " options:\n";
|
|---|
| 139 | print STDERR " -optionfile file Get options from file, useful on systems where\n";
|
|---|
| 140 | print STDERR " long command lines may cause problems\n";
|
|---|
| 141 | print STDERR " -append Add new files to existing collection\n";
|
|---|
| 142 | print STDERR " -remove_archives Remove archives directory after successfully\n";
|
|---|
| 143 | print STDERR " building the collection.\n";
|
|---|
| 144 | print STDERR " -remove_import Remove import directory after successfully\n";
|
|---|
| 145 | print STDERR " importing the collection.\n";
|
|---|
| 146 | print STDERR " -buildtype build|import If 'build' attempt to build directly\n";
|
|---|
| 147 | print STDERR " from archives directory (bypassing import\n";
|
|---|
| 148 | print STDERR " stage). Defaults to 'import'\n";
|
|---|
| 149 | print STDERR " -maxdocs number Maximum number of documents to build\n";
|
|---|
| 150 | print STDERR " -download directory Directory (or file) to get import documents from.\n";
|
|---|
| 151 | print STDERR " There may be multiple download directories and they\n";
|
|---|
| 152 | print STDERR " may be of type http://, ftp://, or file://\n";
|
|---|
| 153 | print STDERR " Note that any existing import directory will be\n";
|
|---|
| 154 | print STDERR " deleted to make way for the downloaded data if\n";
|
|---|
| 155 | print STDERR " a -download option is supplied\n";
|
|---|
| 156 | print STDERR " -collectdir directory Collection directory (defaults to " .
|
|---|
| 157 | &util::filename_cat ($ENV{'GSDLHOME'}, "collect") . ")\n";
|
|---|
| 158 | print STDERR " -dontinstall Only applicable if -collectdir is set to something\n";
|
|---|
| 159 | print STDERR " other than the default. -dontinstall will suppress the\n";
|
|---|
| 160 | print STDERR " default behaviour which is to install the collection to\n";
|
|---|
| 161 | print STDERR " the gsdl/collect directory once it has been built.\n";
|
|---|
| 162 | print STDERR " -save_archives Create a copy of the existing archives directory called\n";
|
|---|
| 163 | print STDERR " archives.org\n";
|
|---|
| 164 | print STDERR " -out Filename or handle to print output status to.\n";
|
|---|
| 165 | print STDERR " The default is STDERR\n";
|
|---|
| 166 | print STDERR " -log_events Log important events (collection built successfully etc.)\n";
|
|---|
| 167 | print STDERR " to event_log_file\n";
|
|---|
| 168 | print STDERR " -event_log_file file File to append important events to (defaults to\n";
|
|---|
| 169 | print STDERR " " . &util::filename_cat ($ENV{'GSDLHOME'}, "etc", "events.txt") . "\n";
|
|---|
| 170 | print STDERR " -email_events addr Comma separated list of email addresses to mail details\n";
|
|---|
| 171 | print STDERR " of important collection building events\n";
|
|---|
| 172 | print STDERR " -mail_server server The outgoing (SMTP) mail server to be used by email_events.\n";
|
|---|
| 173 | print STDERR " email_events will be disabled if mail_server isn't set\n";
|
|---|
| 174 | print STDERR " -event_header file File containing a header to go on any event messages. If not\n";
|
|---|
| 175 | print STDERR " specified build will create a generic header\n\n";
|
|---|
| 176 | }
|
|---|
| 177 |
|
|---|
| 178 | sub main {
|
|---|
| 179 |
|
|---|
| 180 | if ($save_archives && -d $archivedir) {
|
|---|
| 181 | print $out "caching original archives to ${archivedir}.org\n";
|
|---|
| 182 | &util::cp_r ($archivedir, "${archivedir}.org");
|
|---|
| 183 | }
|
|---|
| 184 |
|
|---|
| 185 | # do the download thing if we have any -download options
|
|---|
| 186 | if (scalar (@download)) {
|
|---|
| 187 | # remove any existing import data
|
|---|
| 188 | if (&has_content ($importdir)) {
|
|---|
| 189 | print $out "build: WARNING: removing contents of $importdir\n";
|
|---|
| 190 | &util::rm_r ($importdir);
|
|---|
| 191 | }
|
|---|
| 192 |
|
|---|
| 193 | foreach $download_dir (@download) {
|
|---|
| 194 |
|
|---|
| 195 | # remove any leading or trailing whitespace from filenames (just in case)
|
|---|
| 196 | $download_dir =~ s/^\s+//;
|
|---|
| 197 | $download_dir =~ s/\s+$//;
|
|---|
| 198 |
|
|---|
| 199 | if ($download_dir =~ /^(http|ftp):\/\//) {
|
|---|
| 200 | # use wget to mirror http or ftp urls
|
|---|
| 201 | # options used are:
|
|---|
| 202 | # -P = the directory to download documents to
|
|---|
| 203 | # -np = don't ascend to parent directories. this means that only documents
|
|---|
| 204 | # that live in the same directory or below on the same server as
|
|---|
| 205 | # the given url will be downloaded
|
|---|
| 206 | # -nv = not too verbose
|
|---|
| 207 | # -r = recursively mirror
|
|---|
| 208 | # -N = use time-stamping to see if an up-to-date local copy of each
|
|---|
| 209 | # file already exists. this may be useful if wget fails and
|
|---|
| 210 | # is restarted
|
|---|
| 211 | # -l inf = infinite recursion depth
|
|---|
| 212 | # -R "*\?*" = don't download cgi based urls
|
|---|
| 213 | # -o = the output file to write download status to (only used if the -out
|
|---|
| 214 | # option was given to build)
|
|---|
| 215 |
|
|---|
| 216 | my $download_cmd = "perl -S gsWget.pl -P \"$importdir\" -np -nv";
|
|---|
| 217 | $download_cmd .= " -r -N -l inf -R \"*\\?*\"";
|
|---|
| 218 | $download_cmd .= " -o \"$outfile.download\"" if $use_out;
|
|---|
| 219 | $download_cmd .= " \"$download_dir\"";
|
|---|
| 220 | system ($download_cmd);
|
|---|
| 221 |
|
|---|
| 222 | # note that wget obeys the robot rules. this means that it will have
|
|---|
| 223 | # downloaded a robots.txt file if one was present. since it's unlikely
|
|---|
| 224 | # anyone really wants to include it in a collection we'll delete it.
|
|---|
| 225 | # robots.txt shouldn't be more than two directories deep (I think it will
|
|---|
| 226 | # always be exactly two deep but will look for it in the top directory too)
|
|---|
| 227 | # so that's as far as we'll go looking for it.
|
|---|
| 228 | if (opendir (DIR, $importdir)) {
|
|---|
| 229 | my @files = readdir DIR;
|
|---|
| 230 | closedir DIR;
|
|---|
| 231 | foreach my $file (@files) {
|
|---|
| 232 | next if $file =~ /^\.\.?$/;
|
|---|
| 233 | if ($file =~ /^robots.txt$/i) {
|
|---|
| 234 | &util::rm (&util::filename_cat ($importdir, $file));
|
|---|
| 235 | last;
|
|---|
| 236 | } else {
|
|---|
| 237 | $file = &util::filename_cat ($importdir, $file);
|
|---|
| 238 | if (-d $file) {
|
|---|
| 239 | if (opendir (DIR, $file)) {
|
|---|
| 240 | my @files2 = readdir DIR;
|
|---|
| 241 | closedir DIR;
|
|---|
| 242 | foreach my $file2 (@files2) {
|
|---|
| 243 | if ($file2 =~ /^robots.txt$/i) {
|
|---|
| 244 | &util::rm (&util::filename_cat ($file, $file2));
|
|---|
| 245 | last;
|
|---|
| 246 | }
|
|---|
| 247 | }
|
|---|
| 248 | }
|
|---|
| 249 | }
|
|---|
| 250 | }
|
|---|
| 251 | }
|
|---|
| 252 | }
|
|---|
| 253 |
|
|---|
| 254 | # if using output directory append the file download output to it
|
|---|
| 255 | &append_file ($out, "$outfile.download");
|
|---|
| 256 |
|
|---|
| 257 | } else {
|
|---|
| 258 | # we assume anything not beginning with http:// or ftp://
|
|---|
| 259 | # is a file or directory on the local file system.
|
|---|
| 260 | $download_dir =~ s/^file:(\/\/)?//;
|
|---|
| 261 | $download_dir =~ s/^\s+//; # may be whitespace between "file://" and the rest
|
|---|
| 262 |
|
|---|
| 263 | if (-e $download_dir) {
|
|---|
| 264 | # copy download_dir and all it contains to the import directory
|
|---|
| 265 | my $download_cmd = "perl -S filecopy.pl";
|
|---|
| 266 | $download_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
|
|---|
| 267 | $download_cmd .= " -out \"$outfile.download\"" if $use_out;
|
|---|
| 268 | $download_cmd .= " \"" . $download_dir . "\" " . $collection;
|
|---|
| 269 | system ($download_cmd);
|
|---|
| 270 | # if using output directory append the file download output to it
|
|---|
| 271 | &append_file ($out, "$outfile.download");
|
|---|
| 272 | } else {
|
|---|
| 273 | print $out "WARNING: '$download_dir' does not exist\n";
|
|---|
| 274 | }
|
|---|
| 275 | }
|
|---|
| 276 | }
|
|---|
| 277 | }
|
|---|
| 278 |
|
|---|
| 279 | if (-e &util::filename_cat ($archivedir, "archives.inf")) {
|
|---|
| 280 | if (&has_content ($importdir)) {
|
|---|
| 281 | if ($buildtype eq "build") {
|
|---|
| 282 | &gsdl_build();
|
|---|
| 283 | } else {
|
|---|
| 284 | &gsdl_import();
|
|---|
| 285 | &gsdl_build();
|
|---|
| 286 | }
|
|---|
| 287 | } else {
|
|---|
| 288 | # there are archives but no import, build directly from archives
|
|---|
| 289 | print $out "build: no import material was found, building directly\n";
|
|---|
| 290 | print $out " from archives\n";
|
|---|
| 291 | &gsdl_build();
|
|---|
| 292 | }
|
|---|
| 293 | } else {
|
|---|
| 294 | if (&has_content ($importdir)) {
|
|---|
| 295 | if ($buildtype eq "build") {
|
|---|
| 296 | print $out "build: can't build directly from archives as no\n";
|
|---|
| 297 | print $out " imported archives exist (did you forget to\n";
|
|---|
| 298 | print $out " move the contents of $collection/import to\n";
|
|---|
| 299 | print $out " collection/archives?)\n";
|
|---|
| 300 | }
|
|---|
| 301 | &gsdl_import();
|
|---|
| 302 | &gsdl_build();
|
|---|
| 303 | } else {
|
|---|
| 304 | # no import or archives
|
|---|
| 305 | my $msg = "build: ERROR: The $collection collection has no import or archives data.\n";
|
|---|
| 306 | print $out $msg;
|
|---|
| 307 | &log_event ($msg);
|
|---|
| 308 | &final_out (1) if $use_out;
|
|---|
| 309 | die "\n";
|
|---|
| 310 | }
|
|---|
| 311 | }
|
|---|
| 312 |
|
|---|
| 313 | if ($collectdir ne "" && !$dontinstall) {
|
|---|
| 314 | my $install_collectdir = &util::filename_cat ($ENV{'GSDLHOME'}, "collect");
|
|---|
| 315 | if (!&util::filenames_equal ($collectdir, $install_collectdir)) {
|
|---|
| 316 |
|
|---|
| 317 | # install collection to gsdl/collect
|
|---|
| 318 | print $out "installing the $collection collection\n";
|
|---|
| 319 | my $newdir = &util::filename_cat ($install_collectdir, $collection);
|
|---|
| 320 | my $olddir = &util::filename_cat ($collectdir, $collection);
|
|---|
| 321 | if (-d $newdir) {
|
|---|
| 322 | my $msg = "build: Could not install collection as $newdir\n" .
|
|---|
| 323 | " already exists. Collection will remain at\n$olddir\n";
|
|---|
| 324 |
|
|---|
| 325 | print $out $msg;
|
|---|
| 326 | &log_event ($msg);
|
|---|
| 327 | &final_out (4) if $use_out;
|
|---|
| 328 | die "\n";
|
|---|
| 329 | }
|
|---|
| 330 | if (!&File::Copy::move ($olddir, $newdir)) {
|
|---|
| 331 | my $msg = "build: Failed to install collection to $newdir\n" .
|
|---|
| 332 | " Collection will remain at $olddir\n";
|
|---|
| 333 | print $out $msg;
|
|---|
| 334 | &log_event ($msg);
|
|---|
| 335 | &final_out (5) if $use_out;
|
|---|
| 336 | die "\n";
|
|---|
| 337 | }
|
|---|
| 338 | }
|
|---|
| 339 | }
|
|---|
| 340 |
|
|---|
| 341 | &log_event ("The $collection collection was built successfully\n");
|
|---|
| 342 | &final_out (0) if $use_out;
|
|---|
| 343 | }
|
|---|
| 344 |
|
|---|
| 345 | sub gsdl_import {
|
|---|
| 346 |
|
|---|
| 347 | print $out "importing the $collection collection\n\n";
|
|---|
| 348 |
|
|---|
| 349 | my $import_cmd = "perl -S import.pl";
|
|---|
| 350 | $import_cmd .= " -out \"$outfile.import\"" if $use_out;
|
|---|
| 351 | $import_cmd .= " -removeold" unless $append;
|
|---|
| 352 | $import_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
|
|---|
| 353 | $import_cmd .= " $maxdocs $collection";
|
|---|
| 354 | system ($import_cmd);
|
|---|
| 355 | # if using output directory append the import output to it
|
|---|
| 356 | &append_file ($out, "$outfile.import");
|
|---|
| 357 |
|
|---|
| 358 | if (-e &util::filename_cat ($archivedir, "archives.inf")) {
|
|---|
| 359 | print $out "$collection collection imported successfully\n\n";
|
|---|
| 360 | if ($remove_import) {
|
|---|
| 361 | print $out "removing import directory ($importdir)\n";
|
|---|
| 362 | &util::rm_r ($importdir);
|
|---|
| 363 | }
|
|---|
| 364 | } else {
|
|---|
| 365 | my $msg = "build: ERROR: import.pl failed\n";
|
|---|
| 366 | print $out "\n$msg";
|
|---|
| 367 | &log_event ($msg);
|
|---|
| 368 | &final_out (2) if $use_out;
|
|---|
| 369 | die "\n";
|
|---|
| 370 | }
|
|---|
| 371 | }
|
|---|
| 372 |
|
|---|
| 373 | sub gsdl_build {
|
|---|
| 374 |
|
|---|
| 375 | print $out "building the $collection collection\n\n";
|
|---|
| 376 |
|
|---|
| 377 | my $build_cmd = "perl -S buildcol.pl";
|
|---|
| 378 | $build_cmd .= " -out \"$outfile.build\"" if $use_out;
|
|---|
| 379 | $build_cmd .= " -collectdir \"$collectdir\"" if $collectdir =~ /\w/;
|
|---|
| 380 | $build_cmd .= " $maxdocs $collection";
|
|---|
| 381 | system ($build_cmd);
|
|---|
| 382 | # if using output directory append the buildcol output to it
|
|---|
| 383 | &append_file ($out, "$outfile.build");
|
|---|
| 384 |
|
|---|
| 385 | if (-e &util::filename_cat ($buildingdir, "text", "$collection.ldb") ||
|
|---|
| 386 | -e &util::filename_cat ($buildingdir, "text", "$collection.bdb")) {
|
|---|
| 387 | print $out "$collection collection built successfully\n\n";
|
|---|
| 388 | if ($remove_archives) {
|
|---|
| 389 | print $out "removing archives directory ($archivedir)\n";
|
|---|
| 390 | &util::rm_r ($archivedir);
|
|---|
| 391 | }
|
|---|
| 392 | } else {
|
|---|
| 393 | my $msg = "build: ERROR: buildcol.pl failed\n";
|
|---|
| 394 | print $out "\n$msg";
|
|---|
| 395 | &log_event ($msg);
|
|---|
| 396 | &final_out (3) if $use_out;
|
|---|
| 397 | die "\n";
|
|---|
| 398 | }
|
|---|
| 399 |
|
|---|
| 400 | # replace old indexes with new ones
|
|---|
| 401 | if (&has_content ($indexdir)) {
|
|---|
| 402 | print $out "removing old indexes\n";
|
|---|
| 403 | &util::rm_r ($indexdir);
|
|---|
| 404 | }
|
|---|
| 405 | rmdir ($indexdir) if -d $indexdir;
|
|---|
| 406 | &File::Copy::move ($buildingdir, $indexdir);
|
|---|
| 407 |
|
|---|
| 408 | # remove the cached arhives
|
|---|
| 409 | if ($save_archives && -d "${archivedir}.org") {
|
|---|
| 410 | &util::rm_r ("${archivedir}.org");
|
|---|
| 411 | }
|
|---|
| 412 | }
|
|---|
| 413 |
|
|---|
| 414 | sub has_content {
|
|---|
| 415 | my ($dir) = @_;
|
|---|
| 416 |
|
|---|
| 417 | if (!-d $dir) {return 0;}
|
|---|
| 418 |
|
|---|
| 419 | opendir (DIR, $dir) || return 0;
|
|---|
| 420 | my @files = readdir DIR;
|
|---|
| 421 | close DIR;
|
|---|
| 422 |
|
|---|
| 423 | foreach my $file (@files) {
|
|---|
| 424 | if ($file !~ /^\.{1,2}$/) {
|
|---|
| 425 | return 1;
|
|---|
| 426 | }
|
|---|
| 427 | }
|
|---|
| 428 | return 0;
|
|---|
| 429 | }
|
|---|
| 430 |
|
|---|
| 431 | sub append_file {
|
|---|
| 432 | my ($handle, $file) = @_;
|
|---|
| 433 |
|
|---|
| 434 | open (FILE, $file) || return;
|
|---|
| 435 | undef $/;
|
|---|
| 436 | print $handle <FILE>;
|
|---|
| 437 | $/ = "\n";
|
|---|
| 438 | close FILE;
|
|---|
| 439 | &util::rm ($file);
|
|---|
| 440 | }
|
|---|
| 441 |
|
|---|
| 442 | # creates a file called $outfile.final (should only be called if -out option
|
|---|
| 443 | # is used and isn't STDERR or STDOUT) and writes an output code to it.
|
|---|
| 444 | # An output code of 0 specifies that there was no error
|
|---|
| 445 | sub final_out {
|
|---|
| 446 | my ($exit_code) = @_;
|
|---|
| 447 |
|
|---|
| 448 | if (open (FINAL, ">$outfile.final")) {
|
|---|
| 449 | print FINAL $exit_code;
|
|---|
| 450 | close FINAL;
|
|---|
| 451 | }
|
|---|
| 452 | }
|
|---|
| 453 |
|
|---|
| 454 | sub log_event {
|
|---|
| 455 | my ($msg) = @_;
|
|---|
| 456 |
|
|---|
| 457 | return unless ($log_events || $email_events);
|
|---|
| 458 |
|
|---|
| 459 | # get the event header
|
|---|
| 460 | my $eheader = "[Build Event]\n";
|
|---|
| 461 | $eheader .= "Date: " . scalar localtime() . "\n";
|
|---|
| 462 | if ($event_header ne "" && open (HEADER, $event_header)) {
|
|---|
| 463 | undef $/;
|
|---|
| 464 | $eheader .= <HEADER>;
|
|---|
| 465 | $/ = "\n";
|
|---|
| 466 | close HEADER;
|
|---|
| 467 | } else {
|
|---|
| 468 | $eheader .= "Collection: $collection\n";
|
|---|
| 469 | $eheader .= "GSDLHOME: $ENV{'GSDLHOME'}\n";
|
|---|
| 470 | $eheader .= "Build Location: $collectdir\n";
|
|---|
| 471 | }
|
|---|
| 472 |
|
|---|
| 473 | if ($log_events) {
|
|---|
| 474 | my $fail = 0;
|
|---|
| 475 | # append the event to the event log file
|
|---|
| 476 | if ($event_log_file eq "" || !open (LOG, ">>$event_log_file")) {
|
|---|
| 477 | # log file defaults to $GSDLHOME/etc/events.txt
|
|---|
| 478 | $event_log_file = &util::filename_cat ($ENV{'GSDLHOME'}, "etc", "events.txt");
|
|---|
| 479 | if (!open (LOG, ">>$event_log_file")) {
|
|---|
| 480 | print $out "build: ERROR: Couldn't open event log file $event_log_file\n";
|
|---|
| 481 | $fail = 1;
|
|---|
| 482 | }
|
|---|
| 483 | }
|
|---|
| 484 | if (!$fail) {
|
|---|
| 485 | print LOG $eheader;
|
|---|
| 486 | print LOG $msg;
|
|---|
| 487 | print LOG "\n";
|
|---|
| 488 | close LOG;
|
|---|
| 489 | }
|
|---|
| 490 | }
|
|---|
| 491 |
|
|---|
| 492 | if ($email_events) {
|
|---|
| 493 | # if mail_server isn't set email_events does nothing
|
|---|
| 494 | if ($mail_server eq "") {
|
|---|
| 495 | print $out "build: WARNING: mail_server was not set - email_events option was ignored\n";
|
|---|
| 496 | return;
|
|---|
| 497 | }
|
|---|
| 498 |
|
|---|
| 499 | my %mail = ('SMTP' => $mail_server,
|
|---|
| 500 | 'To' => $email_events,
|
|---|
| 501 | 'From' => $maintainer,
|
|---|
| 502 | 'Subject' => 'Greenstone Build Event'
|
|---|
| 503 | );
|
|---|
| 504 | $mail{'Message'} = $eheader . $msg;
|
|---|
| 505 |
|
|---|
| 506 | if (!sendmail %mail) {
|
|---|
| 507 | print $out "build: ERROR sending mail to $email_events\n";
|
|---|
| 508 | print $out "'$Mail::Sendmail::error'\n";
|
|---|
| 509 | }
|
|---|
| 510 | }
|
|---|
| 511 | }
|
|---|
| 512 |
|
|---|
| 513 |
|
|---|
| 514 | sub parse_args {
|
|---|
| 515 | my ($argref) = @_;
|
|---|
| 516 |
|
|---|
| 517 | if (!parsargv::parse($argref,
|
|---|
| 518 | 'optionfile/.*/', \$optionfile,
|
|---|
| 519 | 'append', \$append,
|
|---|
| 520 | 'remove_archives', \$remove_archives,
|
|---|
| 521 | 'remove_import', \$remove_import,
|
|---|
| 522 | 'buildtype/^(build|import)$/import', \$buildtype,
|
|---|
| 523 | 'maxdocs/^\-?\d+/-1', \$maxdocs,
|
|---|
| 524 | 'download/.+', \@download,
|
|---|
| 525 | 'collectdir/.*/', \$collectdir,
|
|---|
| 526 | 'dontinstall', \$dontinstall,
|
|---|
| 527 | 'save_archives', \$save_archives,
|
|---|
| 528 | 'out/.*/STDERR', \$out,
|
|---|
| 529 | 'log_events', \$log_events,
|
|---|
| 530 | 'event_log_file/.*/', \$event_log_file,
|
|---|
| 531 | 'email_events/.*/', \$email_events,
|
|---|
| 532 | 'mail_server/.*/', \$mail_server,
|
|---|
| 533 | 'event_header/.*/', \$event_header)) {
|
|---|
| 534 |
|
|---|
| 535 | &print_usage();
|
|---|
| 536 | die "\n";
|
|---|
| 537 | }
|
|---|
| 538 | }
|
|---|
