Strings display problems after converting java source files to utf-8 and setting eclipse to utf-8

In order to adapt to new testing tools, I had to convert all my java source files to utf-8 (mostly from windows1252 or iso-8859-1) and changed Eclipse configuration to use utf-8 by default. But the conversion resulted in problems on some strings containing accents.

These strings are read from a database (NLS_CHARACTERSET : WE8MSWIN1252) then sent to a Delphi program using a socket. Neither the database nor the delphi program have been modified.

The strings are retrieved from the database using :

ArrayList<String> menus = new ArrayList<String>(); 
String query = "SELECT ITEM FROM menus ...";
psmt = con.prepareStatement( query );
rs = psmt.executeQuery();
while( rs.next() ) {
    if( rs.getString( "ITEM" ) == null ) continue;
	String s = rs.getString( "ITEM" );
	menus.add( s );
}
return menus;

Then they are sent to the other program using a socket and printwriter

Socket socket = new Socket( getTcpIPAddress(), getTcpCommandPort() );
PrintWriter pred = new PrintWriter(new BufferedWriter(new OutputStreamWriter(socket.getOutputStream())), true);

String str = "ADD:";
str = str.concat( menus.get( 0 ) );
pred.println(str);

I've tried a number of different conversions to create string to send but I still get strange characters instead of accents

String s = rs.getString( "ITEM" );
String m1 = new String( s.getBytes("UTF-8") );
String m2 = new String( s.getBytes("UTF-8"), "ISO-8859-1" );
String m3 = new String( s.getBytes("ISO-8859-1") );
String m4 = new String( s.getBytes("ISO-8859-1"), "UTF-8" );
String m5 = new String( s.getBytes(), "ISO-8859-1" );
String m6 = new String( s.getBytes(), "UTF-8" );

byte[] ba = rs.getBytes( "ITEM" );
String b1 = new String( ba ); 
String b2 = new String( ba, "ISO-8859-1" ); 
String b3 = new String( ba, "UTF-8" ); 
String b4 = new String( ba, "windows-1252" ); 
String b5 = new String( ba, "US-ASCII" ); 

Any idea how to get my accents back, apart from converting the source files back and resetting the default configuration for Eclipse?

Encoding always comes into play when converting between bits and characters and back. The #getBytes() call itself is converting the characters in the string into bits according to your platform's runtime default charset. There are versions of #getBytes() that take character set information to help avoid that. You should specify a charset there, as well as when you instantiate the OutputStreamWriter, to avoid these unintended changes.

The culprit was the Eclipse configuration although I don't understand why.

Setting the options back to default (Cp1252) in Window -> Preferences, General -> Workspace -> "Text file encoding" fixed this issue.