X-Git-Url: http://git.indexdata.com/?p=irspy-moved-to-github.git;a=blobdiff_plain;f=lib%2FZOOM%2FIRSpy%2FUtils.pm;h=75a7bf654c3952cfd5c1adc591dae3ef8377d484;hp=e24fd6ca569c6478f602f7c74aa98bb87e4c98f8;hb=3500a3bcd22ae868037a7b203e7984eb2ca73fe7;hpb=d8931f76879e7d7b5d0cb8340291b7d2dac65c91

diff --git a/lib/ZOOM/IRSpy/Utils.pm b/lib/ZOOM/IRSpy/Utils.pm
index e24fd6c..75a7bf6 100644
--- a/lib/ZOOM/IRSpy/Utils.pm
+++ b/lib/ZOOM/IRSpy/Utils.pm
@@ -1,4 +1,4 @@
-# $Id: Utils.pm,v 1.27 2007-04-27 14:04:40 mike Exp $
+# $Id: Utils.pm,v 1.34 2007-06-27 11:08:08 mike Exp $
 
 package ZOOM::IRSpy::Utils;
 
@@ -7,7 +7,8 @@ use strict;
 use warnings;
 
 use Exporter 'import';
-our @EXPORT_OK = qw(isodate
+our @EXPORT_OK = qw(utf8param
+		    isodate
 		    xml_encode 
 		    cql_quote
 		    cql_target
@@ -21,11 +22,26 @@ our @EXPORT_OK = qw(isodate
 
 use XML::LibXML;
 use XML::LibXML::XPathContext;
+use Encode;
+use Encode qw(is_utf8);
+
 
 our $IRSPY_NS = 'http://indexdata.com/irspy/1.0';
 
 
 # Utility functions follow, exported for use of web UI
+sub utf8param {
+    my($r, $key, $value) = @_;
+    die "utf8param() called with value '$value'" if defined $value;
+
+    my $raw = $r->param($key);
+    return undef if !defined $raw;
+    my $cooked = decode_utf8($raw);
+    warn "converted '$raw' to '", $cooked, "'\n" if $cooked ne $raw;
+    return $cooked;
+}
+
+
 sub isodate {
     my($time) = @_;
 
@@ -70,7 +86,7 @@ sub xml_encode {
 sub cql_quote {
     my($term) = @_;
 
-    $term =~ s/([""\\])/\\$1/g;
+    $term =~ s/([""\\*?])/\\$1/g;
     $term = qq["$term"] if $term =~ /[\s""\/]/;
     return $term;
 }
@@ -181,11 +197,24 @@ sub irspy_record2identifier {
 # a null transform; now we have to be a bit cleverer.
 #
 sub irspy_identifier2target {
+    my $res = _irspy_identifier2target(@_);
+    #carp "converted ID '@_' to target '$res'";
+    return $res;
+}
+
+sub _irspy_identifier2target {
     my($id) = @_;
 
     my($protocol, $target) = ($id =~ /(.*?):(.*)/);
-    print STDERR "protocol='$protocol', target='$target'\n";
-    ### This assumes everything is Z39.50
+    if (uc($protocol) eq "Z39.50") {
+	return "tcp:$target";
+    } elsif (uc($protocol) eq "SRU") {
+	return "sru=get,http:$target";
+    } elsif (uc($protocol) eq "SRW") {
+	return "sru=srw,http:$target";
+    }
+
+    warn "unrecognised protocol '$protocol' in ID $id";
     return $target;
 }
 
@@ -219,21 +248,22 @@ sub modify_xml_document {
 		# we'll check whether the element is already
 		# canonical, to determine whether our change is a
 		# no-op.
-		my $old = "???";
+		my $old = "";
 		my @children = $node->childNodes();
 		if (@children == 1) {
 		    my $child = $node->firstChild();
 		    if (ref $child && ref $child eq "XML::LibXML::Text") {
 			$old = $child->getData();
-			next if $value eq $old;
+			#print STDERR "child='$child', old=", _renderchars($old), "\n" if $key eq "title";
 		    }
 		}
+		next if $value eq $old;
 
 		$node->removeChildNodes();
 		my $child = new XML::LibXML::Text($value);
 		$node->appendChild($child);
 		push @changes, $ref;
-		#print "Elem $key: '$old' -> '$value' ($xpath)<br/>\n";
+		#print STDERR "Elem $key ($xpath): ", _renderchars($old), " -> '", _renderchars($value), "\n";
 	    } else {
 		warn "unexpected node type $node";
 	    }
@@ -251,6 +281,13 @@ sub modify_xml_document {
 }
 
 
+sub _renderchars {
+    my($text) = @_;
+
+    return "'" . $text . "'", " (", join(" ", map {ord($_)} split //, $text), "), is_utf8=" , is_utf8($text);
+}
+
+
 sub dom_add_node {
     my($xc, $ppath, $selector, $value, @addAfter) = @_;