Name: skT45625 Date: 08/28/2000


Java Hotspot Client VM (build 1.3.0beta-b07, mixed mode)

Certain control characters are interpreted mistakenly by JVM. In writing a
telnet-based server, it is essential to be able to use these characters as
they're part of the protcol. These characters are:
- IAC (interpret as command character, ASCII value 255) which is interpreted
with the Java as an exotic Syriac character with the unicode value of 0729
instead of the correct unicode value of 009B (i think)
- EL (erase line, ASCII value 248) is interpreted as a greek letter with the
unicode value 0345
- GA (go ahead character, ASCII value 249) is interpreted as character with a
unicode value of 367

There are most probobly many more, but these are tested. Note that the output is
consistent - the exact same erronous values are produced at any given time

Here's a code snippet that produces the incorrect values:

BufferedReader reader = new BufferedReader(
new InputStreamReader(socket.getInputStream()));

void doRead() throws Exception
{
int readOne;
readOne = reader.read();
// And now here's the erronous output
System.out.println("Char: " + (char) readOne + " / value: " +
readOne);
}

Please be advised that i'm certain that the characters sent to the stream for
reading are the correct values, since i used a standard telnet application that
comes with any Unix system to send them. Besides on a server that uses the exact
same protocol like mine (but is written in C) the sent characters were correctly
interpreted.
(Review ID: 108987)
======================================================================