Beyond the Basics - TCP/IP Sockets in Java Practical Guide for Programmers - page 90

32

while ((bytesRead = in.read(buffer)) != −1)

33

out.write(buffer, 0, bytesRead);

34
35

out.finish(); // Flush bytes from GZIPOutputStream

36

} catch (IOException e) {

37

logger.writeEntry("Exception="+ e.getMessage());

38

}

39
40 try { // Close socket

41 clntSock.close();

42 } catch (IOException e) {

43 entry.add("Exception="+ e.getMessage());

44 }

45
46

logger.writeEntry(entry);

47 }

48 }

CompressProtocolFactory.java

1. Factory method for compression protocol: lines 9-15
createProtocol() returns an anonymous class instance that implements the Runnable
interface. The run() method of this instance simply calls the static method CompressPro−
tocolFactory.handleClient() , which implements the server-side compression protocol.
Note that we do not need a separate CompressProtocol class, because createProtocol()
returns the type of instance (one that implements Runnable ) that we need.
2. handleClient() : lines 17-38
Given a socket connected to the compression client, read the uncompressed bytes from
the client and write the compressed bytes back.
Get socket I/O streams: lines 26-27
The socket's output stream is wrapped in a GZIPOutputStream . The sequence of bytes
written to this stream is compressed, using the GZIP algorithm, before being written
to the underlying output stream.
Read uncompressed and write compressed bytes: lines 29-33
The while loop reads from the socket input stream and writes to the GZIPOutputStream ,
which in turn writes to the socket output stream, until the end-of-stream indication is
received.
Flush and close: lines 35-44
Calling finish() on the GZIPOutputStream is necessary to flush any bytes that may be
buffered by the compression algorithm.

A simple iterative version of the server can be found in CompressServer.java on the topic's
Web site.