gpt4 book ai didi

java - 我如何在 JXTA-JXSE 2.6 中发现点并发送消息?

转载 作者:搜寻专家 更新时间:2023-11-01 01:37:21 27 4
gpt4 key购买 nike

使用来自 http://jxse.kenai.com/ 的 JXTA 2.6我想创建可以在一台或多台主机上运行多个对等点的应用程序。对等点应该能够在组中找到彼此并发送直接消息和传播消息。

满足这些要求的简单的 hello world 类型的应用程序应该是什么样的?

我创建这个问题的目的是提供一个类似答案的教程,两个月前我开始为一个 uni 项目查看 JXTA 时非常努力地找到一个答案。随意添加您自己的答案或改进我的答案。我会等几天,然后接受最好的。

最佳答案

JXTA 2.6 点发现和管道消息传递简介

我希望 2 个月前就有的指南 =)

在大学类(class) build 中花费了大量时间之后一个 JXTA p2p 应用程序 我感到很多挫折和我经历的困惑本可以通过良好的方式避免起点。

您需要的 jar 文件可以在这里找到:

https://oss.sonatype.org/content/repositories/comkenaijxse-057/com/kenai/jxse/jxse/2.6/jxse-2.6.jar

http://sourceforge.net/projects/practicaljxta/files/lib-dependencies-2.6.zip/download

把它们扔进Yourproject/lib,打开eclipse,新建一个项目“Yourproject”,应该就可以搞定了为您导入库。

您很快就会意识到网络上几乎所有信息都已过时,非常过时。您还会遇到很多非常令人困惑的错误消息,其中大部分都可以通过以下方式避免仔细检查此 list 。

  • 您的防火墙是否已关闭或至少为您使用的端口打开?您可以在 Fedora 下使用“sudo service iptables stop”禁用 iptables。

  • 检查拼写!很多时候在加入群组或尝试发送拼写错误的群组名称或未使用的消息时在寻找对等点和服务或打开管道时完全相同的广告会导致非常困惑的消息。当我发现组名是“Net info”和“Net_info”时,我试图弄清楚为什么我的管道连接超时。

  • 您使用的是 JXTA 主目录吗?您在同一台计算机上运行的每个对等点一个?

  • 您真的使用唯一的点 ID 吗?提供给 IDFactory 的种子需要足够长,否则你会得到重复的。

  • 关闭 SELinux。我在开发过程中关闭了 SELinux,但可以想象它会导致错误。

虽然将所有字段组合在一起很常见,但我会在展示需要它们的位置时介绍它们。

注意:这在 2.7 中不起作用。我认为 PSE 成员资格存在一些问题。

public class Hello implements DiscoveryListener, PipeMsgListener {

// When developing you should handle these exceptions, I don't to lessen the clutter of start()
public static void main(String[] args) throws PeerGroupException, IOException {

// JXTA logs a lot, you can configure it setting level here
Logger.getLogger("net.jxta").setLevel(Level.ALL);


// Randomize a port to use with a number over 1000 (for non root on unix)
// JXTA uses TCP for incoming connections which will conflict if more than
// one Hello runs at the same time on one computer.
int port = 9000 + new Random().nextInt(100);

Hello hello = new Hello(port);
hello.start();
hello.fetch_advertisements();
}


private String peer_name;
private PeerID peer_id;
private File conf;
private NetworkManager manager;

public Hello(int port) {
// Add a random number to make it easier to identify by name, will also make sure the ID is unique
peer_name = "Peer " + new Random().nextInt(1000000);

// This is what you will be looking for in Wireshark instead of an IP, hint: filter by "jxta"
peer_id = IDFactory.newPeerID(PeerGroupID.defaultNetPeerGroupID, peer_name.getBytes());

// Here the local peer cache will be saved, if you have multiple peers this must be unique
conf = new File("." + System.getProperty("file.separator") + peer_name);

// Most documentation you will find use a deprecated network manager setup, use this one instead
// ADHOC is usually a good starting point, other alternatives include Edge and Rendezvous
try {
manager = new NetworkManager(
NetworkManager.ConfigMode.ADHOC,
peer_name, conf.toURI());
}
catch (IOException e) {
// Will be thrown if you specify an invalid directory in conf
e.printStackTrace();
}

NetworkConfigurator configurator;
try {
// Settings Configuration
configurator = manager.getConfigurator();
configurator.setTcpPort(port);
configurator.setTcpEnabled(true);
configurator.setTcpIncoming(true);
configurator.setTcpOutgoing(true);
configurator.setUseMulticast(true);
configurator.setPeerID(peer_id);
}
catch (IOException e) {
// Never caught this one but let me know if you do =)
e.printStackTrace();
}
}

private static final String subgroup_name = "Make sure this is spelled the same everywhere";
private static final String subgroup_desc = "...";
private static final PeerGroupID subgroup_id = IDFactory.newPeerGroupID(PeerGroupID.defaultNetPeerGroupID, subgroup_name.getBytes());

private static final String unicast_name = "This must be spelled the same too";
private static final String multicast_name = "Or else you will get the wrong PipeID";

private static final String service_name = "And dont forget it like i did a million times";

private PeerGroup subgroup;
private PipeService pipe_service;
private PipeID unicast_id;
private PipeID multicast_id;
private PipeID service_id;
private DiscoveryService discovery;
private ModuleSpecAdvertisement mdadv;

public void start() throws PeerGroupException, IOException {
// Launch the missiles, if you have logging on and see no exceptions
// after this is ran, then you probably have at least the jars setup correctly.
PeerGroup net_group = manager.startNetwork();

// Connect to our subgroup (all groups are subgroups of Netgroup)
// If the group does not exist, it will be automatically created
// Note this is suggested deprecated, not sure what the better way is
ModuleImplAdvertisement mAdv = null;
try {
mAdv = net_group.getAllPurposePeerGroupImplAdvertisement();
} catch (Exception ex) {
System.err.println(ex.toString());
}
subgroup = net_group.newGroup(subgroup_id, mAdv, subgroup_name, subgroup_desc);

// A simple check to see if connecting to the group worked
if (Module.START_OK != subgroup.startApp(new String[0]))
System.err.println("Cannot start child peergroup");

// We will spice things up to a more interesting level by sending unicast and multicast messages
// In order to be able to do that we will create to listeners that will listen for
// unicast and multicast advertisements respectively. All messages will be handled by Hello in the
// pipeMsgEvent method.

unicast_id = IDFactory.newPipeID(subgroup.getPeerGroupID(), unicast_name.getBytes());
multicast_id = IDFactory.newPipeID(subgroup.getPeerGroupID(), multicast_name.getBytes());

pipe_service = subgroup.getPipeService();
pipe_service.createInputPipe(get_advertisement(unicast_id, false), this);
pipe_service.createInputPipe(get_advertisement(multicast_id, true), this);

// In order to for other peers to find this one (and say hello) we will
// advertise a Hello Service.
discovery = subgroup.getDiscoveryService();
discovery.addDiscoveryListener(this);

ModuleClassAdvertisement mcadv = (ModuleClassAdvertisement)
AdvertisementFactory.newAdvertisement(ModuleClassAdvertisement.getAdvertisementType());

mcadv.setName("STACK-OVERFLOW:HELLO");
mcadv.setDescription("Tutorial example to use JXTA module advertisement Framework");

ModuleClassID mcID = IDFactory.newModuleClassID();

mcadv.setModuleClassID(mcID);

// Let the group know of this service "module" / collection
discovery.publish(mcadv);
discovery.remotePublish(mcadv);

mdadv = (ModuleSpecAdvertisement)
AdvertisementFactory.newAdvertisement(ModuleSpecAdvertisement.getAdvertisementType());
mdadv.setName("STACK-OVERFLOW:HELLO");
mdadv.setVersion("Version 1.0");
mdadv.setCreator("sun.com");
mdadv.setModuleSpecID(IDFactory.newModuleSpecID(mcID));
mdadv.setSpecURI("http://www.jxta.org/Ex1");

service_id = IDFactory.newPipeID(subgroup.getPeerGroupID(), service_name.getBytes());
PipeAdvertisement pipeadv = get_advertisement(service_id, false);
mdadv.setPipeAdvertisement(pipeadv);

// Let the group know of the service
discovery.publish(mdadv);
discovery.remotePublish(mdadv);

// Start listening for discovery events, received by the discoveryEvent method
pipe_service.createInputPipe(pipeadv, this);
}

private static PipeAdvertisement get_advertisement(PipeID id, boolean is_multicast) {
PipeAdvertisement adv = (PipeAdvertisement )AdvertisementFactory.
newAdvertisement(PipeAdvertisement.getAdvertisementType());
adv.setPipeID(id);
if (is_multicast)
adv.setType(PipeService.PropagateType);
else
adv.setType(PipeService.UnicastType);
adv.setName("This however");
adv.setDescription("does not really matter");
return adv;
}

@Override public void discoveryEvent(DiscoveryEvent event) {
// Found another peer! Let's say hello shall we!
// Reformatting to create a real peer id string
String found_peer_id = "urn:jxta:" + event.getSource().toString().substring(7);
send_to_peer("Hello", found_peer_id);
}


private void send_to_peer(String message, String found_peer_id) {
// This is where having the same ID is important or else we wont be
// able to open a pipe and send messages
PipeAdvertisement adv = get_advertisement(unicast_id, false);

// Send message to all peers in "ps", just one in our case
Set<PeerID> ps = new HashSet<PeerID>();
try {
ps.add((PeerID)IDFactory.fromURI(new URI(found_peer_id)));
}
catch (URISyntaxException e) {
// The JXTA peer ids need to be formatted as proper urns
e.printStackTrace();
}

// A pipe we can use to send messages with
OutputPipe sender = null;
try {
sender = pipe_service.createOutputPipe(adv, ps, 10000);
}
catch (IOException e) {
// Thrown if there was an error opening the connection, check firewall settings
e.printStackTrace();
}

Message msg = new Message();
MessageElement fromElem = null;
MessageElement msgElem = null;
try {
fromElem = new ByteArrayMessageElement("From", null, peer_id.toString().getBytes("ISO-8859-1"), null);
msgElem = new ByteArrayMessageElement("Msg", null, message.getBytes("ISO-8859-1"), null);
} catch (UnsupportedEncodingException e) {
// Yepp, you want to spell ISO-8859-1 correctly
e.printStackTrace();
}


msg.addMessageElement(fromElem);
msg.addMessageElement(msgElem);

try {
sender.send(msg);
} catch (IOException e) {
// Check, firewall, settings.
e.printStackTrace();
}
}

@Override public void pipeMsgEvent(PipeMsgEvent event) {
// Someone is sending us a message!
try {
Message msg = event.getMessage();
byte[] msgBytes = msg.getMessageElement("Msg").getBytes(true);
byte[] fromBytes = msg.getMessageElement("From").getBytes(true);
String from = new String(fromBytes);
String message = new String(msgBytes);
System.out.println(message + " says " + from);
}
catch (Exception e) {
// You will notice that JXTA is not very specific with exceptions...
e.printStackTrace();
}
}

/**
* We will not find anyone if we are not regularly looking
*/
private void fetch_advertisements() {
new Thread("fetch advertisements thread") {
public void run() {
while(true) {
discovery.getRemoteAdvertisements(null, DiscoveryService.ADV, "Name", "STACK-OVERFLOW:HELLO", 1, null);
try {
sleep(10000);

}
catch(InterruptedException e) {}
}
}
}.start();
}
}

关于java - 我如何在 JXTA-JXSE 2.6 中发现点并发送消息?,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/8910929/

27 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com