The member variable named "hash" in the gnu.gcj.convert.IOConverter class should have a map entry key named "EUC_JP".
The encoding alias "EUC_JP" is on the list of the Sun's document: "Supported Encoding" http://java.sun.com/j2se/1.4.2/docs/guide/intl/encoding.doc.html So GCJ should support this encoding alias. We need the encoding alias "EUC_JP" when we run a java program which connects to a PostgreSQL database whose encoding is EUC_JP through the PostgreSQL JDBC driver. If you look at IOConverter.java source flie, http://savannah.gnu.org/cgi-bin/viewcvs/gcc/gcc/libjava/gnu/gcj/convert/IOConverter.java#rev1.8 the variable named "hash" has two duplicated "euc-jp" entries. It must be a mistake that one of those should be "euc_jp", I think. Here is a patch to fix this problem: --- libjava/gnu/gcj/convert/IOConverter.java.orig 2005-10-27 12:19:32.739262400 +0900 +++ libjava/gnu/gcj/convert/IOConverter.java 2005-10-27 12:19:56.593563200 +0900 @@ -65,7 +65,7 @@ hash.put ("extended_unix_code_packed_format_for_japanese", "EUCJIS"); hash.put ("cseucpkdfmtjapanese", "EUCJIS"); hash.put ("euc-jp", "EUCJIS"); - hash.put ("euc-jp", "EUCJIS"); + hash.put ("euc_jp", "EUCJIS"); hash.put ("utf-16le", "UnicodeLittle"); hash.put ("utf-16be", "UnicodeBig"); iconv_byte_swap = iconv_init (); -- Summary: Encoding alias "EUC_JP" missing Product: gcc Version: 4.0.1 Status: UNCONFIRMED Severity: normal Priority: P2 Component: libgcj AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: hnakamur at good-day dot co dot jp http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24552