Changeset 27283 for main/trunk

Timestamp:

2013-05-01T18:03:33+12:00 (11 years ago)

Author:

ak19

Message:

1. Fixed an encoding bug that Diego helpfully discovered. Metadata extracted from downloaded .oai files were double encoded. This was because we forgot to add a call to decode() after reading in the contents of the oai file, something we had remembered to do in many other plugins. 2. John Thompson recommended the very helpful Dump(dollar-var) function of Devel::Peek which dumps information about any var to stderr, including encoding information. For strings marked utf8 it also prints out hex and octal values for chars outside of Basic Latin.

File:

• main/trunk/greenstone2/perllib/plugins/OAIPlugin.pm (modified) (2 diffs)

main/trunk/greenstone2/perllib/plugins/OAIPlugin.pm

@@ -27 +27 @@
 package OAIPlugin;
 
+# Devel::Peek's Dump($var) function is useful for debugging encoding issues. 
+#use Devel::Peek;
+use Encode;
 use extrametautil;
 use unicode;
@@ -500 +503 @@
     $self->open_prettyprint_metadata_table();
 
+    # need to decode the string, else it will be double-encoded at this point
+    $$textref = decode("utf-8",$$textref);
+
+# Debugging encoding issues with Devel::Peek's Dump() which prints octal and hexcode
+#    print STDERR "#### text ref: $$textref\n";
+#    print STDERR "\n@@@\n";
+#    Dump($$textref);
+#    print STDERR "\n";
+
     if ($$textref =~ m/<metadata\s*>(.*?)<\/metadata\s*>/s)
     {