# HG changeset patch
# User František Kučera <franta-hg@frantovo.cz>
# Date 1320447969 -3600
# Node ID e5bfc969d41fd5e0fd11c38a560d9da86dbd83eb
# Parent 311502c83ff83c15a80dd7773045ba02147240cd
SMTP: correct unescaping of posted messages containing lines with single dot.
diff -r 311502c83ff8 -r e5bfc969d41f src/org/sonews/daemon/command/PostCommand.java
--- a/src/org/sonews/daemon/command/PostCommand.java Sun Oct 30 22:15:49 2011 +0100
+++ b/src/org/sonews/daemon/command/PostCommand.java Sat Nov 05 00:06:09 2011 +0100
@@ -37,6 +37,7 @@
import org.sonews.storage.StorageManager;
import org.sonews.feed.FeedManager;
import org.sonews.util.Stats;
+import org.sonews.util.io.SMTPInputStream;
/**
* Implementation of the POST command. This command requires multiple lines
@@ -103,6 +104,7 @@
if ("".equals(line) || ".".equals(line)) {
// we finally met the blank line
// separating headers from body
+ // WTF: "."
try {
// Parse the header using the InternetHeader class from JavaMail API
@@ -124,6 +126,7 @@
state = PostState.ReadingBody;
+ // WTF: do we need articles without bodies?
if (".".equals(line)) {
// Post an article without body
postArticle(conn, article);
@@ -138,7 +141,7 @@
headers.setHeader(Headers.LINES, Integer.toString(lineCount));
headers.setHeader(Headers.BYTES, Long.toString(bodySize));
- byte[] body = bufBody.toByteArray();
+ byte[] body = unescapeDots(bufBody.toByteArray());
if (body.length >= 2) {
// Remove trailing CRLF
body = Arrays.copyOf(body, body.length - 2);
@@ -213,7 +216,7 @@
if (conn.getUser() != null && conn.getUser().isAuthenticated()) {
article.setAuthenticatedUser(conn.getUser().getUserName());
}
-
+
if (article.getHeader(Headers.CONTROL)[0].length() > 0) {
controlMessage(conn, article);
} else if (article.getHeader(Headers.SUPERSEDES)[0].length() > 0) {
@@ -273,4 +276,27 @@
}
}
}
+
+ /**
+ * TODO: rework, integrate into NNTPConnection
+ *
+ * @param body message body with doubled dots
+ * @return message body with unescaped dots (.. → .)
+ */
+ private static byte[] unescapeDots(byte[] body) throws IOException {
+ byte[] result = new byte[body.length];
+ int resultLength = 0;
+
+ ByteArrayInputStream escapedInput = new ByteArrayInputStream(body);
+ SMTPInputStream unescapedInput = new SMTPInputStream(escapedInput);
+
+ int ch = unescapedInput.read();
+ while (ch >= 0) {
+ result[resultLength] = (byte) ch;
+ resultLength++;
+ ch = unescapedInput.read();
+ }
+
+ return Arrays.copyOfRange(result, 0, resultLength);
+ }
}
diff -r 311502c83ff8 -r e5bfc969d41f src/org/sonews/util/io/SMTPInputStream.java
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/src/org/sonews/util/io/SMTPInputStream.java Sat Nov 05 00:06:09 2011 +0100
@@ -0,0 +1,110 @@
+/*
+ * SONEWS News Server
+ * see AUTHORS for the list of contributors
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+package org.sonews.util.io;
+
+import java.io.FilterInputStream;
+import java.io.IOException;
+import java.io.InputStream;
+
+/**
+ * Filter input stream for reading from SMTP (or NNTP or similar) socket
+ * where lines containing single dot have special meaning – end of message.
+ *
+ * @author František Kučera (frantovo.cz)
+ */
+public class SMTPInputStream extends FilterInputStream {
+
+ public static final int CR = 0x0d;
+ public static final int LF = 0x0a;
+ public static final int DOT = 0x2e;
+ protected int last;
+
+ public SMTPInputStream(InputStream in) {
+ super(in);
+ }
+
+ /**
+ * @return one byte as expected
+ * or -2 if there was line with single dot (which means end of message)
+ * @throws IOException
+ */
+ @Override
+ public int read() throws IOException {
+ // read current character
+ int ch = super.read();
+
+ if (ch == DOT) {
+ if (last == LF) {
+ int next = super.read();
+
+ if (next == CR || next == LF) { // There should be CRLF, but we may accept also just LF or CR with missing LF. Or should we be more strict?
+ // <CRLF>.<CRLF> → end of current message
+ ch = -2;
+ } else {
+ // <CRLF>.… → eat one dot and return next character
+ ch = next;
+ }
+ }
+ }
+
+ last = ch;
+ return ch;
+ }
+
+ /**
+ * @param buffer
+ * @param offset
+ * @param length
+ * @return See {@link FilterInputStream#read(byte[], int, int)} or -2 (then see {@link #read(byte[])})
+ * @throws IOException
+ */
+ @Override
+ public int read(byte[] buffer, int offset, int length) throws IOException {
+ if (buffer == null) {
+ throw new NullPointerException("Byte array should not be null.");
+ } else if ((offset < 0) || (offset > buffer.length) || (length < 0) || ((offset + length) > buffer.length) || ((offset + length) < 0)) {
+ throw new IndexOutOfBoundsException("Invalid offset or length.");
+ } else if (length == 0) {
+ return 0;
+ }
+
+ int ch = read();
+
+ if (ch == -1 || ch == -2) {
+ return ch;
+ }
+
+ buffer[offset] = (byte) ch;
+
+ int readCounter = 1;
+
+ for (; readCounter < length; readCounter++) {
+ ch = read();
+
+ if (ch == -1 || ch == -2) {
+ break;
+ }
+
+ if (buffer != null) {
+ buffer[offset + readCounter] = (byte) ch;
+ }
+ }
+
+ return readCounter;
+ }
+}