Bug #2165 for MARC-Record: Incorrect record/field lengths for records with UTF-8 characters

Additional information on MARC::Record and UTF-8 characters (what works and what doesn't): The procedure I was using to produce UTF-8 encoded MARC records was to go through a batch of MARC records, passing each subfield to_utf8() of MARC::Charset to get the UTF-8 encoding, and replacing each field of the record as the subfields were finished. The result, as I mentioned, was an incorrect record length for any record with characters outside the ASCII range, and incorrect field lengths for any field with characters outside the ASCII range. Trying something different, I started with a text format (OCLC-like) of the MARC records. I passed each line of the text file to_utf8() to produce perfectly formatted UTF-8 text. I then wrote a Perl program to read in the lines, making subfields and fields of the lines, and appending the resulting fields to a new MARC record. The result: a MARC record with no errors. All the lengths are recorded properly in the leader/directory. Thinking that perhaps a subtle difference between replacing and appending fields resulted in the incorrect lengths, I revised the original program that read through a MARC record, and instead of replacing the field with the content converted to UTF-8 encoding, I deleted the original field, and appended the converted field. The result: a MARC record with incorrect lengths, as before. Thus far stands my research on the issue. Richard Rev. Richard A. Lammert e-mail: lammertra@mail.ctsfw.edu Technical Services Librarian mail: 6600 N. Clinton St. Walther Library Fort Wayne, IN 46825-4996 Concordia Theological Seminary phone: 260-452-3148