1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/org/sonews/storage/Article.java Wed Jul 22 14:04:05 2009 +0200
1.3 @@ -0,0 +1,336 @@
1.4 +/*
1.5 + * SONEWS News Server
1.6 + * see AUTHORS for the list of contributors
1.7 + *
1.8 + * This program is free software: you can redistribute it and/or modify
1.9 + * it under the terms of the GNU General Public License as published by
1.10 + * the Free Software Foundation, either version 3 of the License, or
1.11 + * (at your option) any later version.
1.12 + *
1.13 + * This program is distributed in the hope that it will be useful,
1.14 + * but WITHOUT ANY WARRANTY; without even the implied warranty of
1.15 + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1.16 + * GNU General Public License for more details.
1.17 + *
1.18 + * You should have received a copy of the GNU General Public License
1.19 + * along with this program. If not, see <http://www.gnu.org/licenses/>.
1.20 + */
1.21 +
1.22 +package org.sonews.storage;
1.23 +
1.24 +import java.io.ByteArrayInputStream;
1.25 +import java.io.ByteArrayOutputStream;
1.26 +import java.io.IOException;
1.27 +import java.io.InputStream;
1.28 +import java.nio.charset.Charset;
1.29 +import java.security.MessageDigest;
1.30 +import java.security.NoSuchAlgorithmException;
1.31 +import java.util.UUID;
1.32 +import java.util.ArrayList;
1.33 +import java.util.Enumeration;
1.34 +import java.util.List;
1.35 +import javax.mail.Header;
1.36 +import javax.mail.Message;
1.37 +import javax.mail.MessagingException;
1.38 +import javax.mail.Multipart;
1.39 +import javax.mail.internet.InternetHeaders;
1.40 +import org.sonews.config.Config;
1.41 +import org.sonews.util.Log;
1.42 +
1.43 +/**
1.44 + * Represents a newsgroup article.
1.45 + * @author Christian Lins
1.46 + * @author Denis Schwerdel
1.47 + * @since n3tpd/0.1
1.48 + */
1.49 +public class Article extends ArticleHead
1.50 +{
1.51 +
1.52 + /**
1.53 + * Loads the Article identified by the given ID from the JDBCDatabase.
1.54 + * @param messageID
1.55 + * @return null if Article is not found or if an error occurred.
1.56 + */
1.57 + public static Article getByMessageID(final String messageID)
1.58 + {
1.59 + try
1.60 + {
1.61 + return StorageManager.current().getArticle(messageID);
1.62 + }
1.63 + catch(StorageBackendException ex)
1.64 + {
1.65 + ex.printStackTrace();
1.66 + return null;
1.67 + }
1.68 + }
1.69 +
1.70 + private byte[] body = new byte[0];
1.71 +
1.72 + /**
1.73 + * Default constructor.
1.74 + */
1.75 + public Article()
1.76 + {
1.77 + }
1.78 +
1.79 + /**
1.80 + * Creates a new Article object using the date from the given
1.81 + * raw data.
1.82 + */
1.83 + public Article(String headers, byte[] body)
1.84 + {
1.85 + try
1.86 + {
1.87 + this.body = body;
1.88 +
1.89 + // Parse the header
1.90 + this.headers = new InternetHeaders(
1.91 + new ByteArrayInputStream(headers.getBytes()));
1.92 +
1.93 + this.headerSrc = headers;
1.94 + }
1.95 + catch(MessagingException ex)
1.96 + {
1.97 + ex.printStackTrace();
1.98 + }
1.99 + }
1.100 +
1.101 + /**
1.102 + * Creates an Article instance using the data from the javax.mail.Message
1.103 + * object.
1.104 + * @see javax.mail.Message
1.105 + * @param msg
1.106 + * @throws IOException
1.107 + * @throws MessagingException
1.108 + */
1.109 + public Article(final Message msg)
1.110 + throws IOException, MessagingException
1.111 + {
1.112 + this.headers = new InternetHeaders();
1.113 +
1.114 + for(Enumeration e = msg.getAllHeaders() ; e.hasMoreElements();)
1.115 + {
1.116 + final Header header = (Header)e.nextElement();
1.117 + this.headers.addHeader(header.getName(), header.getValue());
1.118 + }
1.119 +
1.120 + // The "content" of the message can be a String if it's a simple text/plain
1.121 + // message, a Multipart object or an InputStream if the content is unknown.
1.122 + final Object content = msg.getContent();
1.123 + if(content instanceof String)
1.124 + {
1.125 + this.body = ((String)content).getBytes();
1.126 + }
1.127 + else if(content instanceof Multipart) // probably subclass MimeMultipart
1.128 + {
1.129 + // We're are not interested in the different parts of the MultipartMessage,
1.130 + // so we simply read in all data which *can* be huge.
1.131 + InputStream in = msg.getInputStream();
1.132 + this.body = readContent(in);
1.133 + }
1.134 + else if(content instanceof InputStream)
1.135 + {
1.136 + // The message format is unknown to the Message class, but we can
1.137 + // simply read in the whole message data.
1.138 + this.body = readContent((InputStream)content);
1.139 + }
1.140 + else
1.141 + {
1.142 + // Unknown content is probably a malformed mail we should skip.
1.143 + // On the other hand we produce an inconsistent mail mirror, but no
1.144 + // mail system must transport invalid content.
1.145 + Log.msg("Skipping message due to unknown content. Throwing exception...", true);
1.146 + throw new MessagingException("Unknown content: " + content);
1.147 + }
1.148 +
1.149 + // Validate headers
1.150 + validateHeaders();
1.151 + }
1.152 +
1.153 + /**
1.154 + * Reads from the given InputString into a byte array.
1.155 + * TODO: Move this generalized method to org.sonews.util.io.Resource.
1.156 + * @param in
1.157 + * @return
1.158 + * @throws IOException
1.159 + */
1.160 + private byte[] readContent(InputStream in)
1.161 + throws IOException
1.162 + {
1.163 + ByteArrayOutputStream out = new ByteArrayOutputStream();
1.164 +
1.165 + int b = in.read();
1.166 + while(b >= 0)
1.167 + {
1.168 + out.write(b);
1.169 + b = in.read();
1.170 + }
1.171 +
1.172 + return out.toByteArray();
1.173 + }
1.174 +
1.175 + /**
1.176 + * Removes the header identified by the given key.
1.177 + * @param headerKey
1.178 + */
1.179 + public void removeHeader(final String headerKey)
1.180 + {
1.181 + this.headers.removeHeader(headerKey);
1.182 + this.headerSrc = null;
1.183 + }
1.184 +
1.185 + /**
1.186 + * Generates a message id for this article and sets it into
1.187 + * the header object. You have to update the JDBCDatabase manually to make this
1.188 + * change persistent.
1.189 + * Note: a Message-ID should never be changed and only generated once.
1.190 + */
1.191 + private String generateMessageID()
1.192 + {
1.193 + String randomString;
1.194 + MessageDigest md5;
1.195 + try
1.196 + {
1.197 + md5 = MessageDigest.getInstance("MD5");
1.198 + md5.reset();
1.199 + md5.update(getBody());
1.200 + md5.update(getHeader(Headers.SUBJECT)[0].getBytes());
1.201 + md5.update(getHeader(Headers.FROM)[0].getBytes());
1.202 + byte[] result = md5.digest();
1.203 + StringBuffer hexString = new StringBuffer();
1.204 + for (int i = 0; i < result.length; i++)
1.205 + {
1.206 + hexString.append(Integer.toHexString(0xFF & result[i]));
1.207 + }
1.208 + randomString = hexString.toString();
1.209 + }
1.210 + catch (NoSuchAlgorithmException e)
1.211 + {
1.212 + e.printStackTrace();
1.213 + randomString = UUID.randomUUID().toString();
1.214 + }
1.215 + String msgID = "<" + randomString + "@"
1.216 + + Config.inst().get(Config.HOSTNAME, "localhost") + ">";
1.217 +
1.218 + this.headers.setHeader(Headers.MESSAGE_ID, msgID);
1.219 +
1.220 + return msgID;
1.221 + }
1.222 +
1.223 + /**
1.224 + * Returns the body string.
1.225 + */
1.226 + public byte[] getBody()
1.227 + {
1.228 + return body;
1.229 + }
1.230 +
1.231 + /**
1.232 + * @return Charset of the body text
1.233 + */
1.234 + private Charset getBodyCharset()
1.235 + {
1.236 + // We espect something like
1.237 + // Content-Type: text/plain; charset=ISO-8859-15
1.238 + String contentType = getHeader(Headers.CONTENT_TYPE)[0];
1.239 + int idxCharsetStart = contentType.indexOf("charset=") + "charset=".length();
1.240 + int idxCharsetEnd = contentType.indexOf(";", idxCharsetStart);
1.241 +
1.242 + String charsetName = "UTF-8";
1.243 + if(idxCharsetStart >= 0 && idxCharsetStart < contentType.length())
1.244 + {
1.245 + if(idxCharsetEnd < 0)
1.246 + {
1.247 + charsetName = contentType.substring(idxCharsetStart);
1.248 + }
1.249 + else
1.250 + {
1.251 + charsetName = contentType.substring(idxCharsetStart, idxCharsetEnd);
1.252 + }
1.253 + }
1.254 +
1.255 + // Sometimes there are '"' around the name
1.256 + if(charsetName.length() > 2 &&
1.257 + charsetName.charAt(0) == '"' && charsetName.endsWith("\""))
1.258 + {
1.259 + charsetName = charsetName.substring(1, charsetName.length() - 2);
1.260 + }
1.261 +
1.262 + // Create charset
1.263 + Charset charset = Charset.forName("UTF-8"); // This MUST be supported by JVM
1.264 + try
1.265 + {
1.266 + charset = Charset.forName(charsetName);
1.267 + }
1.268 + catch(Exception ex)
1.269 + {
1.270 + Log.msg(ex.getMessage(), false);
1.271 + Log.msg("Article.getBodyCharset(): Unknown charset: " + charsetName, false);
1.272 + }
1.273 + return charset;
1.274 + }
1.275 +
1.276 + /**
1.277 + * @return Numerical IDs of the newsgroups this Article belongs to.
1.278 + */
1.279 + public List<Group> getGroups()
1.280 + {
1.281 + String[] groupnames = getHeader(Headers.NEWSGROUPS)[0].split(",");
1.282 + ArrayList<Group> groups = new ArrayList<Group>();
1.283 +
1.284 + try
1.285 + {
1.286 + for(String newsgroup : groupnames)
1.287 + {
1.288 + newsgroup = newsgroup.trim();
1.289 + Group group = StorageManager.current().getGroup(newsgroup);
1.290 + if(group != null && // If the server does not provide the group, ignore it
1.291 + !groups.contains(group)) // Yes, there may be duplicates
1.292 + {
1.293 + groups.add(group);
1.294 + }
1.295 + }
1.296 + }
1.297 + catch(StorageBackendException ex)
1.298 + {
1.299 + ex.printStackTrace();
1.300 + return null;
1.301 + }
1.302 + return groups;
1.303 + }
1.304 +
1.305 + public void setBody(byte[] body)
1.306 + {
1.307 + this.body = body;
1.308 + }
1.309 +
1.310 + /**
1.311 + *
1.312 + * @param groupname Name(s) of newsgroups
1.313 + */
1.314 + public void setGroup(String groupname)
1.315 + {
1.316 + this.headers.setHeader(Headers.NEWSGROUPS, groupname);
1.317 + }
1.318 +
1.319 + /**
1.320 + * Returns the Message-ID of this Article. If the appropriate header
1.321 + * is empty, a new Message-ID is created.
1.322 + * @return Message-ID of this Article.
1.323 + */
1.324 + public String getMessageID()
1.325 + {
1.326 + String[] msgID = getHeader(Headers.MESSAGE_ID);
1.327 + return msgID[0].equals("") ? generateMessageID() : msgID[0];
1.328 + }
1.329 +
1.330 + /**
1.331 + * @return String containing the Message-ID.
1.332 + */
1.333 + @Override
1.334 + public String toString()
1.335 + {
1.336 + return getMessageID();
1.337 + }
1.338 +
1.339 +}