Bug 8745 - Disallow rebuild_zebra.pl from executing, when run by root user.

[koha.git] / misc / migration_tools / rebuild_zebra.pl

diff --git a/misc/migration_tools/rebuild_zebra.pl b/misc/migration_tools/rebuild_zebra.pl
index 1621e84..14fea3c 100755 (executable)
--- a/misc/migration_tools/rebuild_zebra.pl
+++ b/misc/migration_tools/rebuild_zebra.pl
@@ -10,8 +10,9 @@ use File::Path;
 use C4::Biblio;
 use C4::AuthoritiesMarc;
 use C4::Items;
+use Koha::RecordProcessor;
+use XML::LibXML;
 
-# 
 # script that checks zebradir structure & create directories & mandatory files if needed
 #
 #
@@ -38,6 +39,9 @@ my $do_not_clear_zebraqueue;
 my $length;
 my $where;
 my $offset;
+my $run_as_root;
+my $run_user = (getpwuid($<))[0];
+
 my $verbose_logging = 0;
 my $zebraidx_log_opt = " -v none,fatal,warn ";
 my $result = GetOptions(
@@ -45,7 +49,7 @@ my $result = GetOptions(
     'r|reset'       => \$reset,
     's'             => \$skip_export,
     'k'             => \$keep_export,
-    'I|skip-index'    => \$skip_index,
+    'I|skip-index'  => \$skip_index,
     'nosanitize'    => \$nosanitize,
     'b'             => \$biblios,
     'noxml'         => \$noxml,
@@ -60,14 +64,21 @@ my $result = GetOptions(
     'length:i'        => \$length,
     'offset:i'      => \$offset,
     'v+'             => \$verbose_logging,
+    'run-as-root'    => \$run_as_root,
 );
 
-
 if (not $result or $want_help) {
     print_usage();
     exit 0;
 }
 
+if( not defined $run_as_root and $run_user eq 'root') {
+    my $msg = "Warning: You are running this script as the user 'root'.\n";
+    $msg   .= "If this is intentional you must explicitly specify this using the -run-as-root switch\n";
+    $msg   .= "Please do '$0 --help' to see usage.\n";
+    die $msg;
+}
+
 if (not $biblios and not $authorities) {
     my $msg = "Must specify -b or -a to reindex bibs or authorities\n";
     $msg   .= "Please do '$0 --help' to see usage.\n";
@@ -92,6 +103,10 @@ if ($process_zebraqueue and $do_not_clear_zebraqueue) {
     die $msg;
 }
 
+if ($reset) {
+    $noshadow = 1;
+}
+
 if ($noshadow) {
     $noshadow = ' -n ';
 }
@@ -135,6 +150,8 @@ if ($do_munge) {
     munge_config();
 }
 
+my $tester = XML::LibXML->new();
+
 if ($authorities) {
     index_records('authority', $directory, $skip_export, $skip_index, $process_zebraqueue, $as_xml, $noxml, $nosanitize, $do_not_clear_zebraqueue, $verbose_logging, $zebraidx_log_opt, $authorityserverdir);
 } else {
@@ -250,7 +267,7 @@ sub index_records {
         }
         my $record_fmt = ($as_xml) ? 'marcxml' : 'iso2709' ;
         if ($process_zebraqueue) {
-            do_indexing($record_type, 'delete', "$directory/del_$record_type", $reset, $noshadow, $record_fmt, $zebraidx_log_opt)
+            do_indexing($record_type, 'adelete', "$directory/del_$record_type", $reset, $noshadow, $record_fmt, $zebraidx_log_opt)
                 if %$records_deleted;
             do_indexing($record_type, 'update', "$directory/upd_$record_type", $reset, $noshadow, $record_fmt, $zebraidx_log_opt)
                 if $num_records_exported;
@@ -372,8 +389,18 @@ sub export_marc_records_from_sth {
                         substr($itemsxml, index($itemsxml, "</leader>\n", 0) + 10);
                 }
             }
+            # extra test to ensure that result is valid XML; otherwise
+            # Zebra won't parse it in DOM mode
+            eval {
+                my $doc = $tester->parse_string($marcxml);
+            };
+            if ($@) {
+                warn "Error exporting record $record_number ($record_type): $@\n";
+                next;
+            }
             if ( $marcxml ) {
-                print {$fh} $marcxml if $marcxml;
+                $marcxml =~ s!<\?xml version="1.0" encoding="UTF-8"\?>\n!!;
+                print {$fh} $marcxml;
                 $num_exported++;
             }
             next;
@@ -384,6 +411,12 @@ sub export_marc_records_from_sth {
                 my $rec;
                 if ($as_xml) {
                     $rec = $marc->as_xml_record(C4::Context->preference('marcflavour'));
+                    eval {
+                        my $doc = $tester->parse_string($rec);
+                    };
+                    if ($@) {
+                        die "invalid XML: $@";
+                    }
                     $rec =~ s!<\?xml version="1.0" encoding="UTF-8"\?>\n!!;
                 } else {
                     $rec = $marc->as_usmarc();
@@ -392,7 +425,8 @@ sub export_marc_records_from_sth {
                 $num_exported++;
             };
             if ($@) {
-              warn "Error exporting record $record_number ($record_type) ".($noxml ? "not XML" : "XML");
+                warn "Error exporting record $record_number ($record_type) ".($noxml ? "not XML" : "XML");
+                warn "... specific error is $@" if $verbose_logging;
             }
         }
     }
@@ -436,7 +470,6 @@ sub export_marc_records_from_list {
             if ($@) {
               warn "Error exporting record $record_number ($record_type) ".($noxml ? "not XML" : "XML");
             }
-            $num_exported++;
         }
     }
     print "\nRecords exported: $num_exported\n" if ( $verbose_logging );
@@ -497,6 +530,9 @@ sub get_corrected_marc_record {
         fix_leader($marc);
         if ($record_type eq 'authority') {
             fix_authority_id($marc, $record_number);
+        } elsif ($record_type eq 'biblio' && C4::Context->preference('IncludeSeeFromInSearches')) {
+            my $normalizer = Koha::RecordProcessor->new( { filters => 'EmbedSeeFromHeadings' } );
+            $marc = $normalizer->process($marc);
         }
         if (C4::Context->preference("marcflavour") eq "UNIMARC") {
             fix_unimarc_100($marc);
@@ -648,11 +684,10 @@ sub print_usage {
 $0: reindex MARC bibs and/or authorities in Zebra.
 
 Use this batch job to reindex all biblio or authority
-records in your Koha database.  This job is useful
-only if you are using Zebra; if you are using the 'NoZebra'
-mode, this job should not be used.
+records in your Koha database.
 
 Parameters:
+
     -b                      index bibliographic records
 
     -a                      index authority records
@@ -663,7 +698,7 @@ Parameters:
                             or -s.
 
     -r                      clear Zebra index before
-                            adding records to index
+                            adding records to index. Implies -w.
 
     -d                      Temporary directory for indexing.
                             If not specified, one is automatically
@@ -712,6 +747,9 @@ Parameters:
 
     --munge-config          Deprecated option to try
                             to fix Zebra config files.
+
+    --run-as-root           explicitily allow script to run as 'root' user
+
     --help or -h            show this message.
 _USAGE_
 }