From a7fe6ecf9c3cd707f743fdf0cffbf85a94d0b14c Mon Sep 17 00:00:00 2001
From: Tomas Cohen Arazi <tomascohen@theke.io>
Date: Fri, 13 May 2016 11:51:54 -0300
Subject: [PATCH] [SIGNED-OFF] Bug 16505: <collection> is missing the marc
 namespace and updates fail if -x is passed

Using rebuild_zerba.pl with the -x option switch, produces an incorrect output in
terms of what our XSLTs expect for indexing. This patch introduces the right namespace information
on the exported records so indexing succeeds.

To test:
- On current master, have some records on your db
- Run:
  $ sudo koha-shell kohadev
  $ cd kohaclone
  $ misc/migration_tools/rebuild_zebra.pl -r -b -k -x
=> you will get a message like this:

NOTHING cleaned : the export /tmp/NL5ufjUfpp has been kept.

- Run
  $ less /tmp/NL5ufjUfpp/biblio/exported_records
=> FAIL: The first line looks like this

<?xml version="1.0" encoding="UTF-8"?><collection><record

- Now run:
  $ xsltproc \
     /etc/koha/zebradb/marc_defs/marc21/biblios/biblio-zebra-indexdefs.xsl \
     /tmp/NL5ufjUfpp/biblio/exported_records
=> FAIL: No output
- Apply the patch
- Run:
  $ misc/migration_tools/rebuild_zebra.pl -r -b -k -x
- Take a look at the result file:
  $ less /tmp/asdiouqwiue/biblio/exported_records
=> SUCCESS: The start of the file looks like this:
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim">

- Run:
  $ xsltproc \
     /etc/koha/zebradb/marc_defs/marc21/biblios/biblio-zebra-indexdefs.xsl \
     /tmp/asdiouqwiue/biblio/exported_records
=> SUCCESS: There is actually indexing data :-D
- Sign off :-D

Edit: I changed qq{} for q{} as suggested by Jonathan.

Sponsored-by: American Numismatic Society

Signed-off-by: Bernardo Gonzalez Kriegel <bgkriegel@gmail.com>
Works as described following test plan
No errors
---
 misc/migration_tools/rebuild_zebra.pl | 45 +++++++++++++++++++++++------------
 1 file changed, 30 insertions(+), 15 deletions(-)

diff --git a/misc/migration_tools/rebuild_zebra.pl b/misc/migration_tools/rebuild_zebra.pl
index beaf4d3..45c074a 100755
--- a/misc/migration_tools/rebuild_zebra.pl
+++ b/misc/migration_tools/rebuild_zebra.pl
@@ -168,6 +168,14 @@ my $dbh = C4::Context->dbh;
 my ($biblionumbertagfield,$biblionumbertagsubfield) = &GetMarcFromKohaField("biblio.biblionumber","");
 my ($biblioitemnumbertagfield,$biblioitemnumbertagsubfield) = &GetMarcFromKohaField("biblioitems.biblioitemnumber","");
 
+my $marcxml_open = q{<?xml version="1.0" encoding="UTF-8"?>
+<collection xmlns="http://www.loc.gov/MARC21/slim">
+};
+
+my $marcxml_close = q{
+</collection>
+};
+
 # Protect again simultaneous update of the zebra index by using a lock file.
 # Create our own lock directory if its missing.  This shouild be created
 # by koha-zebra-ctl.sh or at system installation.  If the desired directory
@@ -478,10 +486,10 @@ sub export_marc_records_from_sth {
 
     my $num_exported = 0;
     open my $fh, '>:encoding(UTF-8) ', "$directory/exported_records" or die $!;
-    if (include_xml_wrapper($as_xml, $record_type)) {
-        # include XML declaration and root element
-        print {$fh} '<?xml version="1.0" encoding="UTF-8"?><collection>';
-    }
+
+    print {$fh} $marcxml_open
+        if include_xml_wrapper($as_xml, $record_type);
+
     my $i = 0;
     my ( $itemtag, $itemsubfield ) = GetMarcFromKohaField("items.itemnumber",'');
     while (my ($record_number) = $sth->fetchrow_array) {
@@ -550,7 +558,8 @@ sub export_marc_records_from_sth {
         }
     }
     print "\nRecords exported: $num_exported\n" if ( $verbose_logging );
-    print {$fh} '</collection>' if (include_xml_wrapper($as_xml, $record_type));
+    print {$fh} $marcxml_close
+        if include_xml_wrapper($as_xml, $record_type);
     close $fh;
     return $num_exported;
 }
@@ -560,10 +569,10 @@ sub export_marc_records_from_list {
 
     my $num_exported = 0;
     open my $fh, '>:encoding(UTF-8)', "$directory/exported_records" or die $!;
-    if (include_xml_wrapper($as_xml, $record_type)) {
-        # include XML declaration and root element
-        print {$fh} '<?xml version="1.0" encoding="UTF-8"?><collection>';
-    }
+
+    print {$fh} $marcxml_open
+        if include_xml_wrapper($as_xml, $record_type);
+
     my $i = 0;
 
     # Skip any deleted records. We check for this anyway, but this reduces error spam
@@ -592,7 +601,10 @@ sub export_marc_records_from_list {
         }
     }
     print "\nRecords exported: $num_exported\n" if ( $verbose_logging );
-    print {$fh} '</collection>' if (include_xml_wrapper($as_xml, $record_type));
+
+    print {$fh} $marcxml_close
+        if include_xml_wrapper($as_xml, $record_type);
+
     close $fh;
     return $num_exported;
 }
@@ -602,10 +614,10 @@ sub generate_deleted_marc_records {
 
     my $records_deleted = {};
     open my $fh, '>:encoding(UTF-8)', "$directory/exported_records" or die $!;
-    if (include_xml_wrapper($as_xml, $record_type)) {
-        # include XML declaration and root element
-        print {$fh} '<?xml version="1.0" encoding="UTF-8"?><collection>';
-    }
+
+    print {$fh} $marcxml_open
+        if include_xml_wrapper($as_xml, $record_type);
+
     my $i = 0;
     foreach my $record_number (map { $_->{biblio_auth_number} } @$entries ) {
         print "\r$i" unless ($i++ %100 or !$verbose_logging);
@@ -633,7 +645,10 @@ sub generate_deleted_marc_records {
         $records_deleted->{$record_number} = 1;
     }
     print "\nRecords exported: $i\n" if ( $verbose_logging );
-    print {$fh} '</collection>' if (include_xml_wrapper($as_xml, $record_type));
+
+    print {$fh} $marcxml_close
+        if include_xml_wrapper($as_xml, $record_type);
+
     close $fh;
     return $records_deleted;
 
-- 
1.9.1