javax.mail.internet.MimeUtility Maven / Gradle / Ivy
Show all versions of aem-sdk-api Show documentation
/*
* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS HEADER.
*
* Copyright (c) 1997-2018 Oracle and/or its affiliates. All rights reserved.
*
* The contents of this file are subject to the terms of either the GNU
* General Public License Version 2 only ("GPL") or the Common Development
* and Distribution License("CDDL") (collectively, the "License"). You
* may not use this file except in compliance with the License. You can
* obtain a copy of the License at
* https://oss.oracle.com/licenses/CDDL+GPL-1.1
* or LICENSE.txt. See the License for the specific
* language governing permissions and limitations under the License.
*
* When distributing the software, include this License Header Notice in each
* file and include the License file at LICENSE.txt.
*
* GPL Classpath Exception:
* Oracle designates this particular file as subject to the "Classpath"
* exception as provided by Oracle in the GPL Version 2 section of the License
* file that accompanied this code.
*
* Modifications:
* If applicable, add the following below the License Header, with the fields
* enclosed by brackets [] replaced by your own identifying information:
* "Portions Copyright [year] [name of copyright owner]"
*
* Contributor(s):
* If you wish your version of this file to be governed by only the CDDL or
* only the GPL Version 2, indicate your decision by adding "[Contributor]
* elects to include this software in this distribution under the [CDDL or GPL
* Version 2] license." If you don't indicate a single choice of license, a
* recipient has the option to distribute your version of this file under
* either the CDDL, the GPL Version 2 or to extend the choice of license to
* its licensees as provided above. However, if you add GPL Version 2 code
* and therefore, elected the GPL Version 2 license, then the option applies
* only if the new code is made subject to such option by the copyright
* holder.
*/
package javax.mail.internet;
import javax.mail.MessagingException;
import javax.mail.EncodingAware;
import javax.activation.*;
import java.util.*;
import java.io.*;
import java.nio.charset.Charset;
import com.sun.mail.util.PropUtil;
import com.sun.mail.util.ASCIIUtility;
import com.sun.mail.util.BASE64DecoderStream;
import com.sun.mail.util.BASE64EncoderStream;
import com.sun.mail.util.BEncoderStream;
import com.sun.mail.util.LineInputStream;
import com.sun.mail.util.LineOutputStream;
import com.sun.mail.util.LogOutputStream;
import com.sun.mail.util.QDecoderStream;
import com.sun.mail.util.QEncoderStream;
import com.sun.mail.util.QPDecoderStream;
import com.sun.mail.util.QPEncoderStream;
import com.sun.mail.util.UUDecoderStream;
import com.sun.mail.util.UUEncoderStream;
/**
* This is a utility class that provides various MIME related
* functionality.
*
* There are a set of methods to encode and decode MIME headers as
* per RFC 2047. Note that, in general, these methods are
* not needed when using methods such as
* setSubject and setRecipients; JavaMail
* will automatically encode and decode data when using these "higher
* level" methods. The methods below are only needed when maniuplating
* raw MIME headers using setHeader and getHeader
* methods. A brief description on handling such headers is given below:
*
* RFC 822 mail headers must contain only US-ASCII
* characters. Headers that contain non US-ASCII characters must be
* encoded so that they contain only US-ASCII characters. Basically,
* this process involves using either BASE64 or QP to encode certain
* characters. RFC 2047 describes this in detail.
*
* In Java, Strings contain (16 bit) Unicode characters. ASCII is a
* subset of Unicode (and occupies the range 0 - 127). A String
* that contains only ASCII characters is already mail-safe. If the
* String contains non US-ASCII characters, it must be encoded. An
* additional complexity in this step is that since Unicode is not
* yet a widely used charset, one might want to first charset-encode
* the String into another charset and then do the transfer-encoding.
*
* Note that to get the actual bytes of a mail-safe String (say,
* for sending over SMTP), one must do
*
*
* byte[] bytes = string.getBytes("iso-8859-1");
*
*
*
* The setHeader and addHeader methods
* on MimeMessage and MimeBodyPart assume that the given header values
* are Unicode strings that contain only US-ASCII characters. Hence
* the callers of those methods must insure that the values they pass
* do not contain non US-ASCII characters. The methods in this class
* help do this.
*
* The getHeader family of methods on MimeMessage and
* MimeBodyPart return the raw header value. These might be encoded
* as per RFC 2047, and if so, must be decoded into Unicode Strings.
* The methods in this class help to do this.
*
* Several System properties control strict conformance to the MIME
* spec. Note that these are not session properties but must be set
* globally as System properties.
*
* The mail.mime.decodetext.strict property controls
* decoding of MIME encoded words. The MIME spec requires that encoded
* words start at the beginning of a whitespace separated word. Some
* mailers incorrectly include encoded words in the middle of a word.
* If the mail.mime.decodetext.strict System property is
* set to "false", an attempt will be made to decode these
* illegal encoded words. The default is true.
*
* The mail.mime.encodeeol.strict property controls the
* choice of Content-Transfer-Encoding for MIME parts that are not of
* type "text". Often such parts will contain textual data for which
* an encoding that allows normal end of line conventions is appropriate.
* In rare cases, such a part will appear to contain entirely textual
* data, but will require an encoding that preserves CR and LF characters
* without change. If the mail.mime.encodeeol.strict
* System property is set to "true", such an encoding will
* be used when necessary. The default is false.
*
* In addition, the mail.mime.charset System property can
* be used to specify the default MIME charset to use for encoded words
* and text parts that don't otherwise specify a charset. Normally, the
* default MIME charset is derived from the default Java charset, as
* specified in the file.encoding System property. Most
* applications will have no need to explicitly set the default MIME
* charset. In cases where the default MIME charset to be used for
* mail messages is different than the charset used for files stored on
* the system, this property should be set.
*
* The current implementation also supports the following System property.
*
* The mail.mime.ignoreunknownencoding property controls
* whether unknown values in the Content-Transfer-Encoding
* header, as passed to the decode method, cause an exception.
* If set to "true", unknown values are ignored and 8bit
* encoding is assumed. Otherwise, unknown values cause a MessagingException
* to be thrown.
*
* @author John Mani
* @author Bill Shannon
*/
public class MimeUtility {
// This class cannot be instantiated
private MimeUtility() { }
public static final int ALL = -1;
// cached map of whether a charset is compatible with ASCII
// Map
private static final Map nonAsciiCharsetMap
= new HashMap<>();
private static final boolean decodeStrict =
PropUtil.getBooleanSystemProperty("mail.mime.decodetext.strict", true);
private static final boolean encodeEolStrict =
PropUtil.getBooleanSystemProperty("mail.mime.encodeeol.strict", false);
private static final boolean ignoreUnknownEncoding =
PropUtil.getBooleanSystemProperty(
"mail.mime.ignoreunknownencoding", false);
private static final boolean allowUtf8 =
PropUtil.getBooleanSystemProperty("mail.mime.allowutf8", false);
/*
* The following two properties allow disabling the fold()
* and unfold() methods and reverting to the previous behavior.
* They should never need to be changed and are here only because
* of my paranoid concern with compatibility.
*/
private static final boolean foldEncodedWords =
PropUtil.getBooleanSystemProperty("mail.mime.foldencodedwords", false);
private static final boolean foldText =
PropUtil.getBooleanSystemProperty("mail.mime.foldtext", true);
/**
* Get the Content-Transfer-Encoding that should be applied
* to the input stream of this DataSource, to make it mail-safe.
*
* The algorithm used here is:
*
* -
* If the DataSource implements {@link EncodingAware}, ask it
* what encoding to use. If it returns non-null, return that value.
*
-
* If the primary type of this datasource is "text" and if all
* the bytes in its input stream are US-ASCII, then the encoding
* is "7bit". If more than half of the bytes are non-US-ASCII, then
* the encoding is "base64". If less than half of the bytes are
* non-US-ASCII, then the encoding is "quoted-printable".
*
-
* If the primary type of this datasource is not "text", then if
* all the bytes of its input stream are US-ASCII, the encoding
* is "7bit". If there is even one non-US-ASCII character, the
* encoding is "base64".
*
*
* @param ds the DataSource
* @return the encoding. This is either "7bit",
* "quoted-printable" or "base64"
*/
public static String getEncoding(DataSource ds) {
ContentType cType = null;
InputStream is = null;
String encoding = null;
if (ds instanceof EncodingAware) {
encoding = ((EncodingAware)ds).getEncoding();
if (encoding != null)
return encoding;
}
try {
cType = new ContentType(ds.getContentType());
is = ds.getInputStream();
boolean isText = cType.match("text/*");
// if not text, stop processing when we see non-ASCII
int i = checkAscii(is, ALL, !isText);
switch (i) {
case ALL_ASCII:
encoding = "7bit"; // all ASCII
break;
case MOSTLY_ASCII:
if (isText && nonAsciiCharset(cType))
encoding = "base64"; // charset isn't compatible with ASCII
else
encoding = "quoted-printable"; // mostly ASCII
break;
default:
encoding = "base64"; // mostly binary
break;
}
} catch (Exception ex) {
return "base64"; // what else ?!
} finally {
// Close the input stream
try {
if (is != null)
is.close();
} catch (IOException ioex) { }
}
return encoding;
}
/**
* Determine whether the charset in the Content-Type is compatible
* with ASCII or not. A charset is compatible with ASCII if the
* encoded byte stream representing the Unicode string "\r\n" is
* the ASCII characters CR and LF. For example, the utf-16be
* charset is not compatible with ASCII.
*
* For performance, we keep a static map that caches the results.
*/
private static boolean nonAsciiCharset(ContentType ct) {
String charset = ct.getParameter("charset");
if (charset == null)
return false;
charset = charset.toLowerCase(Locale.ENGLISH);
Boolean bool;
synchronized (nonAsciiCharsetMap) {
bool = nonAsciiCharsetMap.get(charset);
}
if (bool == null) {
try {
byte[] b = "\r\n".getBytes(charset);
bool = Boolean.valueOf(
b.length != 2 || b[0] != 015 || b[1] != 012);
} catch (UnsupportedEncodingException uex) {
bool = Boolean.FALSE; // a guess
} catch (RuntimeException ex) {
bool = Boolean.TRUE; // one of the weird ones?
}
synchronized (nonAsciiCharsetMap) {
nonAsciiCharsetMap.put(charset, bool);
}
}
return bool.booleanValue();
}
/**
* Same as getEncoding(DataSource) except that instead
* of reading the data from an InputStream it uses the
* writeTo method to examine the data. This is more
* efficient in the common case of a DataHandler
* created with an object and a MIME type (for example, a
* "text/plain" String) because all the I/O is done in this
* thread. In the case requiring an InputStream the
* DataHandler uses a thread, a pair of pipe streams,
* and the writeTo method to produce the data.
*
* @param dh the DataHandler
* @return the Content-Transfer-Encoding
* @since JavaMail 1.2
*/
public static String getEncoding(DataHandler dh) {
ContentType cType = null;
String encoding = null;
/*
* Try to pick the most efficient means of determining the
* encoding. If this DataHandler was created using a DataSource,
* the getEncoding(DataSource) method is typically faster. If
* the DataHandler was created with an object, this method is
* much faster. To distinguish the two cases, we use a heuristic.
* A DataHandler created with an object will always have a null name.
* A DataHandler created with a DataSource will usually have a
* non-null name.
*
* XXX - This is actually quite a disgusting hack, but it makes
* a common case run over twice as fast.
*/
if (dh.getName() != null)
return getEncoding(dh.getDataSource());
try {
cType = new ContentType(dh.getContentType());
} catch (Exception ex) {
return "base64"; // what else ?!
}
if (cType.match("text/*")) {
// Check all of the available bytes
AsciiOutputStream aos = new AsciiOutputStream(false, false);
try {
dh.writeTo(aos);
} catch (IOException ex) {
// ignore it, can't happen
}
switch (aos.getAscii()) {
case ALL_ASCII:
encoding = "7bit"; // all ascii
break;
case MOSTLY_ASCII:
encoding = "quoted-printable"; // mostly ascii
break;
default:
encoding = "base64"; // mostly binary
break;
}
} else { // not "text"
// Check all of available bytes, break out if we find
// at least one non-US-ASCII character
AsciiOutputStream aos =
new AsciiOutputStream(true, encodeEolStrict);
try {
dh.writeTo(aos);
} catch (IOException ex) { } // ignore it
if (aos.getAscii() == ALL_ASCII) // all ascii
encoding = "7bit";
else // found atleast one non-ascii character, use b64
encoding = "base64";
}
return encoding;
}
/**
* Decode the given input stream. The Input stream returned is
* the decoded input stream. All the encodings defined in RFC 2045
* are supported here. They include "base64", "quoted-printable",
* "7bit", "8bit", and "binary". In addition, "uuencode" is also
* supported.
*
* In the current implementation, if the
* mail.mime.ignoreunknownencoding system property is set to
* "true", unknown encoding values are ignored and the
* original InputStream is returned.
*
* @param is input stream
* @param encoding the encoding of the stream.
* @return decoded input stream.
* @exception MessagingException if the encoding is unknown
*/
public static InputStream decode(InputStream is, String encoding)
throws MessagingException {
if (encoding.equalsIgnoreCase("base64"))
return new BASE64DecoderStream(is);
else if (encoding.equalsIgnoreCase("quoted-printable"))
return new QPDecoderStream(is);
else if (encoding.equalsIgnoreCase("uuencode") ||
encoding.equalsIgnoreCase("x-uuencode") ||
encoding.equalsIgnoreCase("x-uue"))
return new UUDecoderStream(is);
else if (encoding.equalsIgnoreCase("binary") ||
encoding.equalsIgnoreCase("7bit") ||
encoding.equalsIgnoreCase("8bit"))
return is;
else {
if (!ignoreUnknownEncoding)
throw new MessagingException("Unknown encoding: " + encoding);
return is;
}
}
/**
* Wrap an encoder around the given output stream.
* All the encodings defined in RFC 2045 are supported here.
* They include "base64", "quoted-printable", "7bit", "8bit" and
* "binary". In addition, "uuencode" is also supported.
*
* @param os output stream
* @param encoding the encoding of the stream.
* @return output stream that applies the
* specified encoding.
* @exception MessagingException if the encoding is unknown
*/
public static OutputStream encode(OutputStream os, String encoding)
throws MessagingException {
if (encoding == null)
return os;
else if (encoding.equalsIgnoreCase("base64"))
return new BASE64EncoderStream(os);
else if (encoding.equalsIgnoreCase("quoted-printable"))
return new QPEncoderStream(os);
else if (encoding.equalsIgnoreCase("uuencode") ||
encoding.equalsIgnoreCase("x-uuencode") ||
encoding.equalsIgnoreCase("x-uue"))
return new UUEncoderStream(os);
else if (encoding.equalsIgnoreCase("binary") ||
encoding.equalsIgnoreCase("7bit") ||
encoding.equalsIgnoreCase("8bit"))
return os;
else
throw new MessagingException("Unknown encoding: " +encoding);
}
/**
* Wrap an encoder around the given output stream.
* All the encodings defined in RFC 2045 are supported here.
* They include "base64", "quoted-printable", "7bit", "8bit" and
* "binary". In addition, "uuencode" is also supported.
* The filename parameter is used with the "uuencode"
* encoding and is included in the encoded output.
*
* @param os output stream
* @param encoding the encoding of the stream.
* @param filename name for the file being encoded (only used
* with uuencode)
* @return output stream that applies the
* specified encoding.
* @exception MessagingException for unknown encodings
* @since JavaMail 1.2
*/
public static OutputStream encode(OutputStream os, String encoding,
String filename)
throws MessagingException {
if (encoding == null)
return os;
else if (encoding.equalsIgnoreCase("base64"))
return new BASE64EncoderStream(os);
else if (encoding.equalsIgnoreCase("quoted-printable"))
return new QPEncoderStream(os);
else if (encoding.equalsIgnoreCase("uuencode") ||
encoding.equalsIgnoreCase("x-uuencode") ||
encoding.equalsIgnoreCase("x-uue"))
return new UUEncoderStream(os, filename);
else if (encoding.equalsIgnoreCase("binary") ||
encoding.equalsIgnoreCase("7bit") ||
encoding.equalsIgnoreCase("8bit"))
return os;
else
throw new MessagingException("Unknown encoding: " +encoding);
}
/**
* Encode a RFC 822 "text" token into mail-safe form as per
* RFC 2047.
*
* The given Unicode string is examined for non US-ASCII
* characters. If the string contains only US-ASCII characters,
* it is returned as-is. If the string contains non US-ASCII
* characters, it is first character-encoded using the platform's
* default charset, then transfer-encoded using either the B or
* Q encoding. The resulting bytes are then returned as a Unicode
* string containing only ASCII characters.
*
* Note that this method should be used to encode only
* "unstructured" RFC 822 headers.
*
* Example of usage:
*
*
* MimePart part = ...
* String rawvalue = "FooBar Mailer, Japanese version 1.1"
* try {
* // If we know for sure that rawvalue contains only US-ASCII
* // characters, we can skip the encoding part
* part.setHeader("X-mailer", MimeUtility.encodeText(rawvalue));
* } catch (UnsupportedEncodingException e) {
* // encoding failure
* } catch (MessagingException me) {
* // setHeader() failure
* }
*
*
*
* @param text Unicode string
* @return Unicode string containing only US-ASCII characters
* @exception UnsupportedEncodingException if the encoding fails
*/
public static String encodeText(String text)
throws UnsupportedEncodingException {
return encodeText(text, null, null);
}
/**
* Encode a RFC 822 "text" token into mail-safe form as per
* RFC 2047.
*
* The given Unicode string is examined for non US-ASCII
* characters. If the string contains only US-ASCII characters,
* it is returned as-is. If the string contains non US-ASCII
* characters, it is first character-encoded using the specified
* charset, then transfer-encoded using either the B or Q encoding.
* The resulting bytes are then returned as a Unicode string
* containing only ASCII characters.
*
* Note that this method should be used to encode only
* "unstructured" RFC 822 headers.
*
* @param text the header value
* @param charset the charset. If this parameter is null, the
* platform's default chatset is used.
* @param encoding the encoding to be used. Currently supported
* values are "B" and "Q". If this parameter is null, then
* the "Q" encoding is used if most of characters to be
* encoded are in the ASCII charset, otherwise "B" encoding
* is used.
* @return Unicode string containing only US-ASCII characters
* @exception UnsupportedEncodingException if the charset
* conversion failed.
*/
public static String encodeText(String text, String charset,
String encoding)
throws UnsupportedEncodingException {
return encodeWord(text, charset, encoding, false);
}
/**
* Decode "unstructured" headers, that is, headers that are defined
* as '*text' as per RFC 822.
*
* The string is decoded using the algorithm specified in
* RFC 2047, Section 6.1. If the charset-conversion fails
* for any sequence, an UnsupportedEncodingException is thrown.
* If the String is not an RFC 2047 style encoded header, it is
* returned as-is
*
* Example of usage:
*
*
* MimePart part = ...
* String rawvalue = null;
* String value = null;
* try {
* if ((rawvalue = part.getHeader("X-mailer")[0]) != null)
* value = MimeUtility.decodeText(rawvalue);
* } catch (UnsupportedEncodingException e) {
* // Don't care
* value = rawvalue;
* } catch (MessagingException me) { }
*
* return value;
*
*
*
* @param etext the possibly encoded value
* @return the decoded text
* @exception UnsupportedEncodingException if the charset
* conversion failed.
*/
public static String decodeText(String etext)
throws UnsupportedEncodingException {
/*
* We look for sequences separated by "linear-white-space".
* (as per RFC 2047, Section 6.1)
* RFC 822 defines "linear-white-space" as SPACE | HT | CR | NL.
*/
String lwsp = " \t\n\r";
StringTokenizer st;
/*
* First, lets do a quick run thru the string and check
* whether the sequence "=?" exists at all. If none exists,
* we know there are no encoded-words in here and we can just
* return the string as-is, without suffering thru the later
* decoding logic.
* This handles the most common case of unencoded headers
* efficiently.
*/
if (etext.indexOf("=?") == -1)
return etext;
// Encoded words found. Start decoding ...
st = new StringTokenizer(etext, lwsp, true);
StringBuilder sb = new StringBuilder(); // decode buffer
StringBuilder wsb = new StringBuilder(); // white space buffer
boolean prevWasEncoded = false;
while (st.hasMoreTokens()) {
char c;
String s = st.nextToken();
// If whitespace, append it to the whitespace buffer
if (((c = s.charAt(0)) == ' ') || (c == '\t') ||
(c == '\r') || (c == '\n'))
wsb.append(c);
else {
// Check if token is an 'encoded-word' ..
String word;
try {
word = decodeWord(s);
// Yes, this IS an 'encoded-word'.
if (!prevWasEncoded && wsb.length() > 0) {
// if the previous word was also encoded, we
// should ignore the collected whitespace. Else
// we include the whitespace as well.
sb.append(wsb);
}
prevWasEncoded = true;
} catch (ParseException pex) {
// This is NOT an 'encoded-word'.
word = s;
// possibly decode inner encoded words
if (!decodeStrict) {
String dword = decodeInnerWords(word);
if (dword != word) {
// if a different String object was returned,
// decoding was done.
if (prevWasEncoded && word.startsWith("=?")) {
// encoded followed by encoded,
// throw away whitespace between
} else {
// include collected whitespace ..
if (wsb.length() > 0)
sb.append(wsb);
}
// did original end with encoded?
prevWasEncoded = word.endsWith("?=");
word = dword;
} else {
// include collected whitespace ..
if (wsb.length() > 0)
sb.append(wsb);
prevWasEncoded = false;
}
} else {
// include collected whitespace ..
if (wsb.length() > 0)
sb.append(wsb);
prevWasEncoded = false;
}
}
sb.append(word); // append the actual word
wsb.setLength(0); // reset wsb for reuse
}
}
sb.append(wsb); // append trailing whitespace
return sb.toString();
}
/**
* Encode a RFC 822 "word" token into mail-safe form as per
* RFC 2047.
*
* The given Unicode string is examined for non US-ASCII
* characters. If the string contains only US-ASCII characters,
* it is returned as-is. If the string contains non US-ASCII
* characters, it is first character-encoded using the platform's
* default charset, then transfer-encoded using either the B or
* Q encoding. The resulting bytes are then returned as a Unicode
* string containing only ASCII characters.
*
* This method is meant to be used when creating RFC 822 "phrases".
* The InternetAddress class, for example, uses this to encode
* it's 'phrase' component.
*
* @param word Unicode string
* @return Array of Unicode strings containing only US-ASCII
* characters.
* @exception UnsupportedEncodingException if the encoding fails
*/
public static String encodeWord(String word)
throws UnsupportedEncodingException {
return encodeWord(word, null, null);
}
/**
* Encode a RFC 822 "word" token into mail-safe form as per
* RFC 2047.
*
* The given Unicode string is examined for non US-ASCII
* characters. If the string contains only US-ASCII characters,
* it is returned as-is. If the string contains non US-ASCII
* characters, it is first character-encoded using the specified
* charset, then transfer-encoded using either the B or Q encoding.
* The resulting bytes are then returned as a Unicode string
* containing only ASCII characters.
*
* @param word Unicode string
* @param charset the MIME charset
* @param encoding the encoding to be used. Currently supported
* values are "B" and "Q". If this parameter is null, then
* the "Q" encoding is used if most of characters to be
* encoded are in the ASCII charset, otherwise "B" encoding
* is used.
* @return Unicode string containing only US-ASCII characters
* @exception UnsupportedEncodingException if the encoding fails
*/
public static String encodeWord(String word, String charset,
String encoding)
throws UnsupportedEncodingException {
return encodeWord(word, charset, encoding, true);
}
/*
* Encode the given string. The parameter 'encodingWord' should
* be true if a RFC 822 "word" token is being encoded and false if a
* RFC 822 "text" token is being encoded. This is because the
* "Q" encoding defined in RFC 2047 has more restrictions when
* encoding "word" tokens. (Sigh)
*/
private static String encodeWord(String string, String charset,
String encoding, boolean encodingWord)
throws UnsupportedEncodingException {
// If 'string' contains only US-ASCII characters, just
// return it.
int ascii = checkAscii(string);
if (ascii == ALL_ASCII)
return string;
// Else, apply the specified charset conversion.
String jcharset;
if (charset == null) { // use default charset
jcharset = getDefaultJavaCharset(); // the java charset
charset = getDefaultMIMECharset(); // the MIME equivalent
} else // MIME charset -> java charset
jcharset = javaCharset(charset);
// If no transfer-encoding is specified, figure one out.
if (encoding == null) {
if (ascii != MOSTLY_NONASCII)
encoding = "Q";
else
encoding = "B";
}
boolean b64;
if (encoding.equalsIgnoreCase("B"))
b64 = true;
else if (encoding.equalsIgnoreCase("Q"))
b64 = false;
else
throw new UnsupportedEncodingException(
"Unknown transfer encoding: " + encoding);
StringBuilder outb = new StringBuilder(); // the output buffer
doEncode(string, b64, jcharset,
// As per RFC 2047, size of an encoded string should not
// exceed 75 bytes.
// 7 = size of "=?", '?', 'B'/'Q', '?', "?="
75 - 7 - charset.length(), // the available space
"=?" + charset + "?" + encoding + "?", // prefix
true, encodingWord, outb);
return outb.toString();
}
private static void doEncode(String string, boolean b64,
String jcharset, int avail, String prefix,
boolean first, boolean encodingWord, StringBuilder buf)
throws UnsupportedEncodingException {
// First find out what the length of the encoded version of
// 'string' would be.
byte[] bytes = string.getBytes(jcharset);
int len;
if (b64) // "B" encoding
len = BEncoderStream.encodedLength(bytes);
else // "Q"
len = QEncoderStream.encodedLength(bytes, encodingWord);
int size;
if ((len > avail) && ((size = string.length()) > 1)) {
// If the length is greater than 'avail', split 'string'
// into two and recurse.
// Have to make sure not to split a Unicode surrogate pair.
int split = size / 2;
if (Character.isHighSurrogate(string.charAt(split-1)))
split--;
if (split > 0)
doEncode(string.substring(0, split), b64, jcharset,
avail, prefix, first, encodingWord, buf);
doEncode(string.substring(split, size), b64, jcharset,
avail, prefix, false, encodingWord, buf);
} else {
// length <= than 'avail'. Encode the given string
ByteArrayOutputStream os = new ByteArrayOutputStream();
OutputStream eos; // the encoder
if (b64) // "B" encoding
eos = new BEncoderStream(os);
else // "Q" encoding
eos = new QEncoderStream(os, encodingWord);
try { // do the encoding
eos.write(bytes);
eos.close();
} catch (IOException ioex) { }
byte[] encodedBytes = os.toByteArray(); // the encoded stuff
// Now write out the encoded (all ASCII) bytes into our
// StringBuilder
if (!first) // not the first line of this sequence
if (foldEncodedWords)
buf.append("\r\n "); // start a continuation line
else
buf.append(" "); // line will be folded later
buf.append(prefix);
for (int i = 0; i < encodedBytes.length; i++)
buf.append((char)encodedBytes[i]);
buf.append("?="); // terminate the current sequence
}
}
/**
* The string is parsed using the rules in RFC 2047 and RFC 2231 for
* parsing an "encoded-word". If the parse fails, a ParseException is
* thrown. Otherwise, it is transfer-decoded, and then
* charset-converted into Unicode. If the charset-conversion
* fails, an UnsupportedEncodingException is thrown.
*
* @param eword the encoded value
* @return the decoded word
* @exception ParseException if the string is not an
* encoded-word as per RFC 2047 and RFC 2231.
* @exception UnsupportedEncodingException if the charset
* conversion failed.
*/
public static String decodeWord(String eword)
throws ParseException, UnsupportedEncodingException {
if (!eword.startsWith("=?")) // not an encoded word
throw new ParseException(
"encoded word does not start with \"=?\": " + eword);
// get charset
int start = 2; int pos;
if ((pos = eword.indexOf('?', start)) == -1)
throw new ParseException(
"encoded word does not include charset: " + eword);
String charset = eword.substring(start, pos);
int lpos = charset.indexOf('*'); // RFC 2231 language specified?
if (lpos >= 0) // yes, throw it away
charset = charset.substring(0, lpos);
charset = javaCharset(charset);
// get encoding
start = pos+1;
if ((pos = eword.indexOf('?', start)) == -1)
throw new ParseException(
"encoded word does not include encoding: " + eword);
String encoding = eword.substring(start, pos);
// get encoded-sequence
start = pos+1;
if ((pos = eword.indexOf("?=", start)) == -1)
throw new ParseException(
"encoded word does not end with \"?=\": " + eword);
/*
* XXX - should include this, but leaving it out for compatibility...
*
if (decodeStrict && pos != eword.length() - 2)
throw new ParseException(
"encoded word does not end with \"?=\": " + eword););
*/
String word = eword.substring(start, pos);
try {
String decodedWord;
if (word.length() > 0) {
// Extract the bytes from word
ByteArrayInputStream bis =
new ByteArrayInputStream(ASCIIUtility.getBytes(word));
// Get the appropriate decoder
InputStream is;
if (encoding.equalsIgnoreCase("B"))
is = new BASE64DecoderStream(bis);
else if (encoding.equalsIgnoreCase("Q"))
is = new QDecoderStream(bis);
else
throw new UnsupportedEncodingException(
"unknown encoding: " + encoding);
// For b64 & q, size of decoded word <= size of word. So
// the decoded bytes must fit into the 'bytes' array. This
// is certainly more efficient than writing bytes into a
// ByteArrayOutputStream and then pulling out the byte[]
// from it.
int count = bis.available();
byte[] bytes = new byte[count];
// count is set to the actual number of decoded bytes
count = is.read(bytes, 0, count);
// Finally, convert the decoded bytes into a String using
// the specified charset
decodedWord = count <= 0 ? "" :
new String(bytes, 0, count, charset);
} else {
// no characters to decode, return empty string
decodedWord = "";
}
if (pos + 2 < eword.length()) {
// there's still more text in the string
String rest = eword.substring(pos + 2);
if (!decodeStrict)
rest = decodeInnerWords(rest);
decodedWord += rest;
}
return decodedWord;
} catch (UnsupportedEncodingException uex) {
// explicitly catch and rethrow this exception, otherwise
// the below IOException catch will swallow this up!
throw uex;
} catch (IOException ioex) {
// Shouldn't happen.
throw new ParseException(ioex.toString());
} catch (IllegalArgumentException iex) {
/* An unknown charset of the form ISO-XXX-XXX, will cause
* the JDK to throw an IllegalArgumentException ... Since the
* JDK will attempt to create a classname using this string,
* but valid classnames must not contain the character '-',
* and this results in an IllegalArgumentException, rather than
* the expected UnsupportedEncodingException. Yikes
*/
throw new UnsupportedEncodingException(charset);
}
}
/**
* Look for encoded words within a word. The MIME spec doesn't
* allow this, but many broken mailers, especially Japanese mailers,
* produce such incorrect encodings.
*/
private static String decodeInnerWords(String word)
throws UnsupportedEncodingException {
int start = 0, i;
StringBuilder buf = new StringBuilder();
while ((i = word.indexOf("=?", start)) >= 0) {
buf.append(word.substring(start, i));
// find first '?' after opening '=?' - end of charset
int end = word.indexOf('?', i + 2);
if (end < 0)
break;
// find next '?' after that - end of encoding
end = word.indexOf('?', end + 1);
if (end < 0)
break;
// find terminating '?='
end = word.indexOf("?=", end + 1);
if (end < 0)
break;
String s = word.substring(i, end + 2);
try {
s = decodeWord(s);
} catch (ParseException pex) {
// ignore it, just use the original string
}
buf.append(s);
start = end + 2;
}
if (start == 0)
return word;
if (start < word.length())
buf.append(word.substring(start));
return buf.toString();
}
/**
* A utility method to quote a word, if the word contains any
* characters from the specified 'specials' list.
*
* The HeaderTokenizer class defines two special
* sets of delimiters - MIME and RFC 822.
*
* This method is typically used during the generation of
* RFC 822 and MIME header fields.
*
* @param word word to be quoted
* @param specials the set of special characters
* @return the possibly quoted word
* @see javax.mail.internet.HeaderTokenizer#MIME
* @see javax.mail.internet.HeaderTokenizer#RFC822
*/
public static String quote(String word, String specials) {
int len = word == null ? 0 : word.length();
if (len == 0)
return "\"\""; // an empty string is handled specially
/*
* Look for any "bad" characters, Escape and
* quote the entire string if necessary.
*/
boolean needQuoting = false;
for (int i = 0; i < len; i++) {
char c = word.charAt(i);
if (c == '"' || c == '\\' || c == '\r' || c == '\n') {
// need to escape them and then quote the whole string
StringBuilder sb = new StringBuilder(len + 3);
sb.append('"');
sb.append(word.substring(0, i));
int lastc = 0;
for (int j = i; j < len; j++) {
char cc = word.charAt(j);
if ((cc == '"') || (cc == '\\') ||
(cc == '\r') || (cc == '\n'))
if (cc == '\n' && lastc == '\r')
; // do nothing, CR was already escaped
else
sb.append('\\'); // Escape the character
sb.append(cc);
lastc = cc;
}
sb.append('"');
return sb.toString();
} else if (c < 040 || (c >= 0177 && !allowUtf8) ||
specials.indexOf(c) >= 0)
// These characters cause the string to be quoted
needQuoting = true;
}
if (needQuoting) {
StringBuilder sb = new StringBuilder(len + 2);
sb.append('"').append(word).append('"');
return sb.toString();
} else
return word;
}
/**
* Fold a string at linear whitespace so that each line is no longer
* than 76 characters, if possible. If there are more than 76
* non-whitespace characters consecutively, the string is folded at
* the first whitespace after that sequence. The parameter
* used indicates how many characters have been used in
* the current line; it is usually the length of the header name.
*
* Note that line breaks in the string aren't escaped; they probably
* should be.
*
* @param used characters used in line so far
* @param s the string to fold
* @return the folded string
* @since JavaMail 1.4
*/
public static String fold(int used, String s) {
if (!foldText)
return s;
int end;
char c;
// Strip trailing spaces and newlines
for (end = s.length() - 1; end >= 0; end--) {
c = s.charAt(end);
if (c != ' ' && c != '\t' && c != '\r' && c != '\n')
break;
}
if (end != s.length() - 1)
s = s.substring(0, end + 1);
// if the string fits now, just return it
if (used + s.length() <= 76)
return makesafe(s);
// have to actually fold the string
StringBuilder sb = new StringBuilder(s.length() + 4);
char lastc = 0;
while (used + s.length() > 76) {
int lastspace = -1;
for (int i = 0; i < s.length(); i++) {
if (lastspace != -1 && used + i > 76)
break;
c = s.charAt(i);
if (c == ' ' || c == '\t')
if (!(lastc == ' ' || lastc == '\t'))
lastspace = i;
lastc = c;
}
if (lastspace == -1) {
// no space, use the whole thing
sb.append(s);
s = "";
used = 0;
break;
}
sb.append(s.substring(0, lastspace));
sb.append("\r\n");
lastc = s.charAt(lastspace);
sb.append(lastc);
s = s.substring(lastspace + 1);
used = 1;
}
sb.append(s);
return makesafe(sb);
}
/**
* If the String or StringBuilder has any embedded newlines,
* make sure they're followed by whitespace, to prevent header
* injection errors.
*/
private static String makesafe(CharSequence s) {
int i;
for (i = 0; i < s.length(); i++) {
char c = s.charAt(i);
if (c == '\r' || c == '\n')
break;
}
if (i == s.length()) // went through whole string with no CR or LF
return s.toString();
// read the lines in the string and reassemble them,
// eliminating blank lines and inserting whitespace as necessary
StringBuilder sb = new StringBuilder(s.length() + 1);
BufferedReader r = new BufferedReader(new StringReader(s.toString()));
String line;
try {
while ((line = r.readLine()) != null) {
if (line.trim().length() == 0)
continue; // ignore empty lines
if (sb.length() > 0) {
sb.append("\r\n");
assert line.length() > 0; // proven above
char c = line.charAt(0);
if (c != ' ' && c != '\t')
sb.append(' ');
}
sb.append(line);
}
} catch (IOException ex) {
// XXX - should never happen when reading from a string
return s.toString();
}
return sb.toString();
}
/**
* Unfold a folded header. Any line breaks that aren't escaped and
* are followed by whitespace are removed.
*
* @param s the string to unfold
* @return the unfolded string
* @since JavaMail 1.4
*/
public static String unfold(String s) {
if (!foldText)
return s;
StringBuilder sb = null;
int i;
while ((i = indexOfAny(s, "\r\n")) >= 0) {
int start = i;
int slen = s.length();
i++; // skip CR or NL
if (i < slen && s.charAt(i - 1) == '\r' && s.charAt(i) == '\n')
i++; // skip LF
if (start > 0 && s.charAt(start - 1) == '\\') {
// there's a backslash before the line break
// strip it out, but leave in the line break
if (sb == null)
sb = new StringBuilder(s.length());
sb.append(s.substring(0, start - 1));
sb.append(s.substring(start, i));
s = s.substring(i);
} else {
char c;
// if next line starts with whitespace,
// or at the end of the string, remove the line break
// XXX - next line should always start with whitespace
if (i >= slen || (c = s.charAt(i)) == ' ' || c == '\t') {
if (sb == null)
sb = new StringBuilder(s.length());
sb.append(s.substring(0, start));
s = s.substring(i);
} else {
// it's not a continuation line, just leave in the newline
if (sb == null)
sb = new StringBuilder(s.length());
sb.append(s.substring(0, i));
s = s.substring(i);
}
}
}
if (sb != null) {
sb.append(s);
return sb.toString();
} else
return s;
}
/**
* Return the first index of any of the characters in "any" in "s",
* or -1 if none are found.
*
* This should be a method on String.
*/
private static int indexOfAny(String s, String any) {
return indexOfAny(s, any, 0);
}
private static int indexOfAny(String s, String any, int start) {
try {
int len = s.length();
for (int i = start; i < len; i++) {
if (any.indexOf(s.charAt(i)) >= 0)
return i;
}
return -1;
} catch (StringIndexOutOfBoundsException e) {
return -1;
}
}
/**
* Convert a MIME charset name into a valid Java charset name.
*
* @param charset the MIME charset name
* @return the Java charset equivalent. If a suitable mapping is
* not available, the passed in charset is itself returned.
*/
public static String javaCharset(String charset) {
if (mime2java == null || charset == null)
// no mapping table, or charset parameter is null
return charset;
String alias = mime2java.get(charset.toLowerCase(Locale.ENGLISH));
if (alias != null) {
// verify that the mapped name is valid before trying to use it
try {
Charset.forName(alias);
} catch (Exception ex) {
alias = null; // charset alias not valid, use original name
}
}
return alias == null ? charset : alias;
}
/**
* Convert a java charset into its MIME charset name.
*
* Note that a future version of JDK (post 1.2) might provide
* this functionality, in which case, we may deprecate this
* method then.
*
* @param charset the JDK charset
* @return the MIME/IANA equivalent. If a mapping
* is not possible, the passed in charset itself
* is returned.
* @since JavaMail 1.1
*/
public static String mimeCharset(String charset) {
if (java2mime == null || charset == null)
// no mapping table or charset param is null
return charset;
String alias = java2mime.get(charset.toLowerCase(Locale.ENGLISH));
return alias == null ? charset : alias;
}
private static String defaultJavaCharset;
private static String defaultMIMECharset;
/**
* Get the default charset corresponding to the system's current
* default locale. If the System property mail.mime.charset
* is set, a system charset corresponding to this MIME charset will be
* returned.
*
* @return the default charset of the system's default locale,
* as a Java charset. (NOT a MIME charset)
* @since JavaMail 1.1
*/
public static String getDefaultJavaCharset() {
if (defaultJavaCharset == null) {
/*
* If mail.mime.charset is set, it controls the default
* Java charset as well.
*/
String mimecs = null;
try {
mimecs = System.getProperty("mail.mime.charset");
} catch (SecurityException ex) { } // ignore it
if (mimecs != null && mimecs.length() > 0) {
defaultJavaCharset = javaCharset(mimecs);
return defaultJavaCharset;
}
try {
defaultJavaCharset = System.getProperty("file.encoding",
"8859_1");
} catch (SecurityException sex) {
class NullInputStream extends InputStream {
@Override
public int read() {
return 0;
}
}
InputStreamReader reader =
new InputStreamReader(new NullInputStream());
defaultJavaCharset = reader.getEncoding();
if (defaultJavaCharset == null)
defaultJavaCharset = "8859_1";
}
}
return defaultJavaCharset;
}
/*
* Get the default MIME charset for this locale.
*/
static String getDefaultMIMECharset() {
if (defaultMIMECharset == null) {
try {
defaultMIMECharset = System.getProperty("mail.mime.charset");
} catch (SecurityException ex) { } // ignore it
}
if (defaultMIMECharset == null)
defaultMIMECharset = mimeCharset(getDefaultJavaCharset());
return defaultMIMECharset;
}
// Tables to map MIME charset names to Java names and vice versa.
// XXX - Should eventually use J2SE 1.4 java.nio.charset.Charset
private static Map mime2java;
private static Map java2mime;
static {
java2mime = new HashMap<>(40);
mime2java = new HashMap<>(14);
try {
// Use this class's classloader to load the mapping file
// XXX - we should use SecuritySupport, but it's in another package
InputStream is =
javax.mail.internet.MimeUtility.class.getResourceAsStream(
"/META-INF/javamail.charset.map");
if (is != null) {
try {
is = new LineInputStream(is);
// Load the JDK-to-MIME charset mapping table
loadMappings((LineInputStream)is, java2mime);
// Load the MIME-to-JDK charset mapping table
loadMappings((LineInputStream)is, mime2java);
} finally {
try {
is.close();
} catch (Exception cex) {
// ignore
}
}
}
} catch (Exception ex) { }
// If we didn't load the tables, e.g., because we didn't have
// permission, load them manually. The entries here should be
// the same as the default javamail.charset.map.
if (java2mime.isEmpty()) {
java2mime.put("8859_1", "ISO-8859-1");
java2mime.put("iso8859_1", "ISO-8859-1");
java2mime.put("iso8859-1", "ISO-8859-1");
java2mime.put("8859_2", "ISO-8859-2");
java2mime.put("iso8859_2", "ISO-8859-2");
java2mime.put("iso8859-2", "ISO-8859-2");
java2mime.put("8859_3", "ISO-8859-3");
java2mime.put("iso8859_3", "ISO-8859-3");
java2mime.put("iso8859-3", "ISO-8859-3");
java2mime.put("8859_4", "ISO-8859-4");
java2mime.put("iso8859_4", "ISO-8859-4");
java2mime.put("iso8859-4", "ISO-8859-4");
java2mime.put("8859_5", "ISO-8859-5");
java2mime.put("iso8859_5", "ISO-8859-5");
java2mime.put("iso8859-5", "ISO-8859-5");
java2mime.put("8859_6", "ISO-8859-6");
java2mime.put("iso8859_6", "ISO-8859-6");
java2mime.put("iso8859-6", "ISO-8859-6");
java2mime.put("8859_7", "ISO-8859-7");
java2mime.put("iso8859_7", "ISO-8859-7");
java2mime.put("iso8859-7", "ISO-8859-7");
java2mime.put("8859_8", "ISO-8859-8");
java2mime.put("iso8859_8", "ISO-8859-8");
java2mime.put("iso8859-8", "ISO-8859-8");
java2mime.put("8859_9", "ISO-8859-9");
java2mime.put("iso8859_9", "ISO-8859-9");
java2mime.put("iso8859-9", "ISO-8859-9");
java2mime.put("sjis", "Shift_JIS");
java2mime.put("jis", "ISO-2022-JP");
java2mime.put("iso2022jp", "ISO-2022-JP");
java2mime.put("euc_jp", "euc-jp");
java2mime.put("koi8_r", "koi8-r");
java2mime.put("euc_cn", "euc-cn");
java2mime.put("euc_tw", "euc-tw");
java2mime.put("euc_kr", "euc-kr");
}
if (mime2java.isEmpty()) {
mime2java.put("iso-2022-cn", "ISO2022CN");
mime2java.put("iso-2022-kr", "ISO2022KR");
mime2java.put("utf-8", "UTF8");
mime2java.put("utf8", "UTF8");
mime2java.put("ja_jp.iso2022-7", "ISO2022JP");
mime2java.put("ja_jp.eucjp", "EUCJIS");
mime2java.put("euc-kr", "KSC5601");
mime2java.put("euckr", "KSC5601");
mime2java.put("us-ascii", "ISO-8859-1");
mime2java.put("x-us-ascii", "ISO-8859-1");
mime2java.put("gb2312", "GB18030");
mime2java.put("cp936", "GB18030");
mime2java.put("ms936", "GB18030");
mime2java.put("gbk", "GB18030");
}
}
private static void loadMappings(LineInputStream is,
Map table) {
String currLine;
while (true) {
try {
currLine = is.readLine();
} catch (IOException ioex) {
break; // error in reading, stop
}
if (currLine == null) // end of file, stop
break;
if (currLine.startsWith("--") && currLine.endsWith("--"))
// end of this table
break;
// ignore empty lines and comments
if (currLine.trim().length() == 0 || currLine.startsWith("#"))
continue;
// A valid entry is of the form
// where, := SPACE | HT. Parse this
StringTokenizer tk = new StringTokenizer(currLine, " \t");
try {
String key = tk.nextToken();
String value = tk.nextToken();
table.put(key.toLowerCase(Locale.ENGLISH), value);
} catch (NoSuchElementException nex) { }
}
}
static final int ALL_ASCII = 1;
static final int MOSTLY_ASCII = 2;
static final int MOSTLY_NONASCII = 3;
/**
* Check if the given string contains non US-ASCII characters.
* @param s string
* @return ALL_ASCII if all characters in the string
* belong to the US-ASCII charset. MOSTLY_ASCII
* if more than half of the available characters
* are US-ASCII characters. Else MOSTLY_NONASCII.
*/
static int checkAscii(String s) {
int ascii = 0, non_ascii = 0;
int l = s.length();
for (int i = 0; i < l; i++) {
if (nonascii((int)s.charAt(i))) // non-ascii
non_ascii++;
else
ascii++;
}
if (non_ascii == 0)
return ALL_ASCII;
if (ascii > non_ascii)
return MOSTLY_ASCII;
return MOSTLY_NONASCII;
}
/**
* Check if the given byte array contains non US-ASCII characters.
* @param b byte array
* @return ALL_ASCII if all characters in the string
* belong to the US-ASCII charset. MOSTLY_ASCII
* if more than half of the available characters
* are US-ASCII characters. Else MOSTLY_NONASCII.
*
* XXX - this method is no longer used
*/
static int checkAscii(byte[] b) {
int ascii = 0, non_ascii = 0;
for (int i=0; i < b.length; i++) {
// The '&' operator automatically causes b[i] to be promoted
// to an int, and we mask out the higher bytes in the int
// so that the resulting value is not a negative integer.
if (nonascii(b[i] & 0xff)) // non-ascii
non_ascii++;
else
ascii++;
}
if (non_ascii == 0)
return ALL_ASCII;
if (ascii > non_ascii)
return MOSTLY_ASCII;
return MOSTLY_NONASCII;
}
/**
* Check if the given input stream contains non US-ASCII characters.
* Upto max bytes are checked. If max is
* set to ALL, then all the bytes available in this
* input stream are checked. If breakOnNonAscii is true
* the check terminates when the first non-US-ASCII character is
* found and MOSTLY_NONASCII is returned. Else, the check continues
* till max bytes or till the end of stream.
*
* @param is the input stream
* @param max maximum bytes to check for. The special value
* ALL indicates that all the bytes in this input
* stream must be checked.
* @param breakOnNonAscii if true, then terminate the
* the check when the first non-US-ASCII character
* is found.
* @return ALL_ASCII if all characters in the string
* belong to the US-ASCII charset. MOSTLY_ASCII
* if more than half of the available characters
* are US-ASCII characters. Else MOSTLY_NONASCII.
*/
static int checkAscii(InputStream is, int max, boolean breakOnNonAscii) {
int ascii = 0, non_ascii = 0;
int len;
int block = 4096;
int linelen = 0;
boolean longLine = false, badEOL = false;
boolean checkEOL = encodeEolStrict && breakOnNonAscii;
byte buf[] = null;
if (max != 0) {
block = (max == ALL) ? 4096 : Math.min(max, 4096);
buf = new byte[block];
}
while (max != 0) {
try {
if ((len = is.read(buf, 0, block)) == -1)
break;
int lastb = 0;
for (int i = 0; i < len; i++) {
// The '&' operator automatically causes b[i] to
// be promoted to an int, and we mask out the higher
// bytes in the int so that the resulting value is
// not a negative integer.
int b = buf[i] & 0xff;
if (checkEOL &&
((lastb == '\r' && b != '\n') ||
(lastb != '\r' && b == '\n')))
badEOL = true;
if (b == '\r' || b == '\n')
linelen = 0;
else {
linelen++;
if (linelen > 998) // 1000 - CRLF
longLine = true;
}
if (nonascii(b)) { // non-ascii
if (breakOnNonAscii) // we are done
return MOSTLY_NONASCII;
else
non_ascii++;
} else
ascii++;
lastb = b;
}
} catch (IOException ioex) {
break;
}
if (max != ALL)
max -= len;
}
if (max == 0 && breakOnNonAscii)
// We have been told to break on the first non-ascii character.
// We haven't got any non-ascii character yet, but then we
// have not checked all of the available bytes either. So we
// cannot say for sure that this input stream is ALL_ASCII,
// and hence we must play safe and return MOSTLY_NONASCII
return MOSTLY_NONASCII;
if (non_ascii == 0) { // no non-us-ascii characters so far
// If we're looking at non-text data, and we saw CR without LF
// or vice versa, consider this mostly non-ASCII so that it
// will be base64 encoded (since the quoted-printable encoder
// doesn't encode this case properly).
if (badEOL)
return MOSTLY_NONASCII;
// if we've seen a long line, we degrade to mostly ascii
else if (longLine)
return MOSTLY_ASCII;
else
return ALL_ASCII;
}
if (ascii > non_ascii) // mostly ascii
return MOSTLY_ASCII;
return MOSTLY_NONASCII;
}
static final boolean nonascii(int b) {
return b >= 0177 || (b < 040 && b != '\r' && b != '\n' && b != '\t');
}
}
/**
* An OutputStream that determines whether the data written to
* it is all ASCII, mostly ASCII, or mostly non-ASCII.
*/
class AsciiOutputStream extends OutputStream {
private boolean breakOnNonAscii;
private int ascii = 0, non_ascii = 0;
private int linelen = 0;
private boolean longLine = false;
private boolean badEOL = false;
private boolean checkEOL = false;
private int lastb = 0;
private int ret = 0;
public AsciiOutputStream(boolean breakOnNonAscii, boolean encodeEolStrict) {
this.breakOnNonAscii = breakOnNonAscii;
checkEOL = encodeEolStrict && breakOnNonAscii;
}
@Override
public void write(int b) throws IOException {
check(b);
}
@Override
public void write(byte b[]) throws IOException {
write(b, 0, b.length);
}
@Override
public void write(byte b[], int off, int len) throws IOException {
len += off;
for (int i = off; i < len ; i++)
check(b[i]);
}
private final void check(int b) throws IOException {
b &= 0xff;
if (checkEOL &&
((lastb == '\r' && b != '\n') || (lastb != '\r' && b == '\n')))
badEOL = true;
if (b == '\r' || b == '\n')
linelen = 0;
else {
linelen++;
if (linelen > 998) // 1000 - CRLF
longLine = true;
}
if (MimeUtility.nonascii(b)) { // non-ascii
non_ascii++;
if (breakOnNonAscii) { // we are done
ret = MimeUtility.MOSTLY_NONASCII;
throw new EOFException();
}
} else
ascii++;
lastb = b;
}
/**
* Return ASCII-ness of data stream.
*/
public int getAscii() {
if (ret != 0)
return ret;
// If we're looking at non-text data, and we saw CR without LF
// or vice versa, consider this mostly non-ASCII so that it
// will be base64 encoded (since the quoted-printable encoder
// doesn't encode this case properly).
if (badEOL)
return MimeUtility.MOSTLY_NONASCII;
else if (non_ascii == 0) { // no non-us-ascii characters so far
// if we've seen a long line, we degrade to mostly ascii
if (longLine)
return MimeUtility.MOSTLY_ASCII;
else
return MimeUtility.ALL_ASCII;
}
if (ascii > non_ascii) // mostly ascii
return MimeUtility.MOSTLY_ASCII;
return MimeUtility.MOSTLY_NONASCII;
}
}