|
|
@ -30,42 +30,48 @@ import java.net.HttpURLConnection; |
|
|
|
import java.net.URL; |
|
|
|
import java.net.URL; |
|
|
|
import java.net.URLConnection; |
|
|
|
import java.net.URLConnection; |
|
|
|
import java.text.MessageFormat; |
|
|
|
import java.text.MessageFormat; |
|
|
|
|
|
|
|
import java.util.Locale; |
|
|
|
import java.util.Objects; |
|
|
|
import java.util.Objects; |
|
|
|
import java.util.regex.Matcher; |
|
|
|
import java.util.regex.Matcher; |
|
|
|
import java.util.regex.Pattern; |
|
|
|
import java.util.regex.Pattern; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
import me.ag2s.epublib.util.IOUtil; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Character stream that handles all the necessary Voodo to figure out the |
|
|
|
* Character stream that handles all the necessary Voodoo to figure out the |
|
|
|
* charset encoding of the XML document within the stream. |
|
|
|
* charset encoding of the XML document within the stream. |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* IMPORTANT: This class is not related in any way to the org.xml.sax.XMLReader. |
|
|
|
* IMPORTANT: This class is not related in any way to the org.xml.sax.XMLReader. |
|
|
|
* This one IS a character stream. |
|
|
|
* This one IS a character stream. |
|
|
|
|
|
|
|
* </p> |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* All this has to be done without consuming characters from the stream, if not |
|
|
|
* All this has to be done without consuming characters from the stream, if not |
|
|
|
* the XML parser will not recognized the document as a valid XML. This is not |
|
|
|
* the XML parser will not recognized the document as a valid XML. This is not |
|
|
|
* 100% true, but it's close enough (UTF-8 BOM is not handled by all parsers |
|
|
|
* 100% true, but it's close enough (UTF-8 BOM is not handled by all parsers |
|
|
|
* right now, XmlStreamReader handles it and things work in all parsers). |
|
|
|
* right now, XmlStreamReader handles it and things work in all parsers). |
|
|
|
|
|
|
|
* </p> |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* The XmlStreamReader class handles the charset encoding of XML documents in |
|
|
|
* The XmlStreamReader class handles the charset encoding of XML documents in |
|
|
|
* Files, raw streams and HTTP streams by offering a wide set of constructors. |
|
|
|
* Files, raw streams and HTTP streams by offering a wide set of constructors. |
|
|
|
|
|
|
|
* </p> |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* By default the charset encoding detection is lenient, the constructor with |
|
|
|
* By default the charset encoding detection is lenient, the constructor with |
|
|
|
* the lenient flag can be used for an script (following HTTP MIME and XML |
|
|
|
* the lenient flag can be used for a script (following HTTP MIME and XML |
|
|
|
* specifications). All this is nicely explained by Mark Pilgrim in his blog, <a |
|
|
|
* specifications). All this is nicely explained by Mark Pilgrim in his blog, <a |
|
|
|
* href="http://diveintomark.org/archives/2004/02/13/xml-media-types"> |
|
|
|
* href="http://diveintomark.org/archives/2004/02/13/xml-media-types"> |
|
|
|
* Determining the character encoding of a feed</a>. |
|
|
|
* Determining the character encoding of a feed</a>. |
|
|
|
|
|
|
|
* </p> |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* Originally developed for <a href="http://rome.dev.java.net">ROME</a> under |
|
|
|
* Originally developed for <a href="http://rome.dev.java.net">ROME</a> under |
|
|
|
* Apache License 2.0. |
|
|
|
* Apache License 2.0. |
|
|
|
|
|
|
|
* </p> |
|
|
|
* |
|
|
|
* |
|
|
|
* @author Alejandro Abdelnur |
|
|
|
* //@seerr XmlStreamWriter
|
|
|
|
* @version $Id: XmlStreamReader.java 1052161 2010-12-23 03:12:09Z niallp $ |
|
|
|
* @since 2.0 |
|
|
|
* @see "org.apache.commons.io.output.XmlStreamWriter" |
|
|
|
|
|
|
|
* @since Commons IO 2.0 |
|
|
|
|
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public class XmlStreamReader extends Reader { |
|
|
|
public class XmlStreamReader extends Reader { |
|
|
|
private static final int BUFFER_SIZE = 4096; |
|
|
|
private static final int BUFFER_SIZE = IOUtil.DEFAULT_BUFFER_SIZE; |
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_8 = "UTF-8"; |
|
|
|
private static final String UTF_8 = "UTF-8"; |
|
|
|
|
|
|
|
|
|
|
@ -75,23 +81,36 @@ public class XmlStreamReader extends Reader { |
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_16LE = "UTF-16LE"; |
|
|
|
private static final String UTF_16LE = "UTF-16LE"; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_32BE = "UTF-32BE"; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_32LE = "UTF-32LE"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_16 = "UTF-16"; |
|
|
|
private static final String UTF_16 = "UTF-16"; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
private static final String UTF_32 = "UTF-32"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String EBCDIC = "CP1047"; |
|
|
|
private static final String EBCDIC = "CP1047"; |
|
|
|
|
|
|
|
|
|
|
|
private static final ByteOrderMark[] BOMS = new ByteOrderMark[] { |
|
|
|
private static final ByteOrderMark[] BOMS = new ByteOrderMark[] { |
|
|
|
ByteOrderMark.UTF_8, |
|
|
|
ByteOrderMark.UTF_8, |
|
|
|
ByteOrderMark.UTF_16BE, |
|
|
|
ByteOrderMark.UTF_16BE, |
|
|
|
ByteOrderMark.UTF_16LE |
|
|
|
ByteOrderMark.UTF_16LE, |
|
|
|
|
|
|
|
ByteOrderMark.UTF_32BE, |
|
|
|
|
|
|
|
ByteOrderMark.UTF_32LE |
|
|
|
}; |
|
|
|
}; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// UTF_16LE and UTF_32LE have the same two starting BOM bytes.
|
|
|
|
private static final ByteOrderMark[] XML_GUESS_BYTES = new ByteOrderMark[] { |
|
|
|
private static final ByteOrderMark[] XML_GUESS_BYTES = new ByteOrderMark[] { |
|
|
|
new ByteOrderMark(UTF_8, 0x3C, 0x3F, 0x78, 0x6D), |
|
|
|
new ByteOrderMark(UTF_8, 0x3C, 0x3F, 0x78, 0x6D), |
|
|
|
new ByteOrderMark(UTF_16BE, 0x00, 0x3C, 0x00, 0x3F), |
|
|
|
new ByteOrderMark(UTF_16BE, 0x00, 0x3C, 0x00, 0x3F), |
|
|
|
new ByteOrderMark(UTF_16LE, 0x3C, 0x00, 0x3F, 0x00), |
|
|
|
new ByteOrderMark(UTF_16LE, 0x3C, 0x00, 0x3F, 0x00), |
|
|
|
new ByteOrderMark(EBCDIC, 0x4C, 0x6F, 0xA7, 0x94) |
|
|
|
new ByteOrderMark(UTF_32BE, 0x00, 0x00, 0x00, 0x3C, |
|
|
|
|
|
|
|
0x00, 0x00, 0x00, 0x3F, 0x00, 0x00, 0x00, 0x78, 0x00, 0x00, 0x00, 0x6D), |
|
|
|
|
|
|
|
new ByteOrderMark(UTF_32LE, 0x3C, 0x00, 0x00, 0x00, |
|
|
|
|
|
|
|
0x3F, 0x00, 0x00, 0x00, 0x78, 0x00, 0x00, 0x00, 0x6D, 0x00, 0x00, 0x00), |
|
|
|
|
|
|
|
new ByteOrderMark(EBCDIC, 0x4C, 0x6F, 0xA7, 0x94) |
|
|
|
}; |
|
|
|
}; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
private final Reader reader; |
|
|
|
private final Reader reader; |
|
|
|
|
|
|
|
|
|
|
|
private final String encoding; |
|
|
|
private final String encoding; |
|
|
@ -106,7 +125,6 @@ public class XmlStreamReader extends Reader { |
|
|
|
* |
|
|
|
* |
|
|
|
* @return the default encoding to use. |
|
|
|
* @return the default encoding to use. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
@SuppressWarnings("unused") |
|
|
|
|
|
|
|
public String getDefaultEncoding() { |
|
|
|
public String getDefaultEncoding() { |
|
|
|
return defaultEncoding; |
|
|
|
return defaultEncoding; |
|
|
|
} |
|
|
|
} |
|
|
@ -124,8 +142,8 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws IOException thrown if there is a problem reading the file. |
|
|
|
* @throws IOException thrown if there is a problem reading the file. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
@SuppressWarnings("unused") |
|
|
|
@SuppressWarnings("unused") |
|
|
|
public XmlStreamReader(File file) throws IOException { |
|
|
|
public XmlStreamReader(final File file) throws IOException { |
|
|
|
this(new FileInputStream(file)); |
|
|
|
this(new FileInputStream(Objects.requireNonNull(file))); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
@ -136,11 +154,11 @@ public class XmlStreamReader extends Reader { |
|
|
|
* It does a lenient charset encoding detection, check the constructor with |
|
|
|
* It does a lenient charset encoding detection, check the constructor with |
|
|
|
* the lenient parameter for details. |
|
|
|
* the lenient parameter for details. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create a Reader from. |
|
|
|
* @param inputStream InputStream to create a Reader from. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is) throws IOException { |
|
|
|
public XmlStreamReader(final InputStream inputStream) throws IOException { |
|
|
|
this(is, true); |
|
|
|
this(inputStream, true); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
@ -163,15 +181,15 @@ public class XmlStreamReader extends Reader { |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* thrown. |
|
|
|
* thrown. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create a Reader from. |
|
|
|
* @param inputStream InputStream to create a Reader from. |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* relaxed. |
|
|
|
* relaxed. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* be determined according to the specs. |
|
|
|
* be determined according to the specs. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is, boolean lenient) throws IOException { |
|
|
|
public XmlStreamReader(final InputStream inputStream, final boolean lenient) throws IOException { |
|
|
|
this(is, lenient, null); |
|
|
|
this(inputStream, lenient, null); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
@ -194,7 +212,7 @@ public class XmlStreamReader extends Reader { |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* thrown. |
|
|
|
* thrown. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create a Reader from. |
|
|
|
* @param inputStream InputStream to create a Reader from. |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* relaxed. |
|
|
|
* relaxed. |
|
|
|
* @param defaultEncoding The default encoding |
|
|
|
* @param defaultEncoding The default encoding |
|
|
@ -202,10 +220,12 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* be determined according to the specs. |
|
|
|
* be determined according to the specs. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is, boolean lenient, String defaultEncoding) throws IOException { |
|
|
|
public XmlStreamReader(final InputStream inputStream, final boolean lenient, final String defaultEncoding) |
|
|
|
|
|
|
|
throws IOException { |
|
|
|
|
|
|
|
Objects.requireNonNull(inputStream, "inputStream"); |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
BOMInputStream bom = new BOMInputStream(new BufferedInputStream(is, BUFFER_SIZE), false, BOMS); |
|
|
|
final BOMInputStream bom = new BOMInputStream(new BufferedInputStream(inputStream, BUFFER_SIZE), false, BOMS); |
|
|
|
BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
final BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
this.encoding = doRawStream(bom, pis, lenient); |
|
|
|
this.encoding = doRawStream(bom, pis, lenient); |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
} |
|
|
|
} |
|
|
@ -228,8 +248,8 @@ public class XmlStreamReader extends Reader { |
|
|
|
* the URL. |
|
|
|
* the URL. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
@SuppressWarnings("unused") |
|
|
|
@SuppressWarnings("unused") |
|
|
|
public XmlStreamReader(URL url) throws IOException { |
|
|
|
public XmlStreamReader(final URL url) throws IOException { |
|
|
|
this(url.openConnection(), null); |
|
|
|
this(Objects.requireNonNull(url, "url").openConnection(), null); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
@ -251,24 +271,24 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws IOException thrown if there is a problem reading the stream of |
|
|
|
* @throws IOException thrown if there is a problem reading the stream of |
|
|
|
* the URLConnection. |
|
|
|
* the URLConnection. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(URLConnection conn, String defaultEncoding) throws IOException { |
|
|
|
public XmlStreamReader(final URLConnection conn, final String defaultEncoding) throws IOException { |
|
|
|
|
|
|
|
Objects.requireNonNull(conn, "conm"); |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
@SuppressWarnings("unused") |
|
|
|
final boolean lenient = true; |
|
|
|
boolean lenient = true; |
|
|
|
final String contentType = conn.getContentType(); |
|
|
|
String contentType = conn.getContentType(); |
|
|
|
final InputStream inputStream = conn.getInputStream(); |
|
|
|
InputStream is = conn.getInputStream(); |
|
|
|
final BOMInputStream bom = new BOMInputStream(new BufferedInputStream(inputStream, BUFFER_SIZE), false, BOMS); |
|
|
|
BOMInputStream bom = new BOMInputStream(new BufferedInputStream(is, BUFFER_SIZE), false, BOMS); |
|
|
|
final BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
|
|
|
|
if (conn instanceof HttpURLConnection || contentType != null) { |
|
|
|
if (conn instanceof HttpURLConnection || contentType != null) { |
|
|
|
this.encoding = doHttpStream(bom, pis, contentType, true); |
|
|
|
this.encoding = processHttpStream(bom, pis, contentType, lenient); |
|
|
|
} else { |
|
|
|
} else { |
|
|
|
this.encoding = doRawStream(bom, pis, true); |
|
|
|
this.encoding = doRawStream(bom, pis, lenient); |
|
|
|
} |
|
|
|
} |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Creates a Reader using an InputStream an the associated content-type |
|
|
|
* Creates a Reader using an InputStream and the associated content-type |
|
|
|
* header. |
|
|
|
* header. |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
@ -279,18 +299,18 @@ public class XmlStreamReader extends Reader { |
|
|
|
* It does a lenient charset encoding detection, check the constructor with |
|
|
|
* It does a lenient charset encoding detection, check the constructor with |
|
|
|
* the lenient parameter for details. |
|
|
|
* the lenient parameter for details. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create the reader from. |
|
|
|
* @param inputStream InputStream to create the reader from. |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* the charset encoding. |
|
|
|
* the charset encoding. |
|
|
|
* @throws IOException thrown if there is a problem reading the file. |
|
|
|
* @throws IOException thrown if there is a problem reading the file. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is, String httpContentType) |
|
|
|
public XmlStreamReader(final InputStream inputStream, final String httpContentType) |
|
|
|
throws IOException { |
|
|
|
throws IOException { |
|
|
|
this(is, httpContentType, true); |
|
|
|
this(inputStream, httpContentType, true); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Creates a Reader using an InputStream an the associated content-type |
|
|
|
* Creates a Reader using an InputStream and the associated content-type |
|
|
|
* header. This constructor is lenient regarding the encoding detection. |
|
|
|
* header. This constructor is lenient regarding the encoding detection. |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
@ -313,7 +333,7 @@ public class XmlStreamReader extends Reader { |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* thrown. |
|
|
|
* thrown. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create the reader from. |
|
|
|
* @param inputStream InputStream to create the reader from. |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* the charset encoding. |
|
|
|
* the charset encoding. |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
@ -323,17 +343,18 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* be determined according to the specs. |
|
|
|
* be determined according to the specs. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is, String httpContentType, |
|
|
|
public XmlStreamReader(final InputStream inputStream, final String httpContentType, |
|
|
|
boolean lenient, String defaultEncoding) throws IOException { |
|
|
|
final boolean lenient, final String defaultEncoding) throws IOException { |
|
|
|
|
|
|
|
Objects.requireNonNull(inputStream, "inputStream"); |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
this.defaultEncoding = defaultEncoding; |
|
|
|
BOMInputStream bom = new BOMInputStream(new BufferedInputStream(is, BUFFER_SIZE), false, BOMS); |
|
|
|
final BOMInputStream bom = new BOMInputStream(new BufferedInputStream(inputStream, BUFFER_SIZE), false, BOMS); |
|
|
|
BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
final BOMInputStream pis = new BOMInputStream(bom, true, XML_GUESS_BYTES); |
|
|
|
this.encoding = doHttpStream(bom, pis, httpContentType, lenient); |
|
|
|
this.encoding = processHttpStream(bom, pis, httpContentType, lenient); |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
this.reader = new InputStreamReader(pis, encoding); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Creates a Reader using an InputStream an the associated content-type |
|
|
|
* Creates a Reader using an InputStream and the associated content-type |
|
|
|
* header. This constructor is lenient regarding the encoding detection. |
|
|
|
* header. This constructor is lenient regarding the encoding detection. |
|
|
|
* <p> |
|
|
|
* <p> |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
|
* First it checks if the stream has BOM. If there is not BOM checks the |
|
|
@ -356,7 +377,7 @@ public class XmlStreamReader extends Reader { |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* If lenient detection is indicated an XmlStreamReaderException is never |
|
|
|
* thrown. |
|
|
|
* thrown. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create the reader from. |
|
|
|
* @param inputStream InputStream to create the reader from. |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* @param httpContentType content-type header to use for the resolution of |
|
|
|
* the charset encoding. |
|
|
|
* the charset encoding. |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
|
* @param lenient indicates if the charset encoding detection should be |
|
|
@ -365,9 +386,9 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* @throws XmlStreamReaderException thrown if the charset encoding could not |
|
|
|
* be determined according to the specs. |
|
|
|
* be determined according to the specs. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
public XmlStreamReader(InputStream is, String httpContentType, |
|
|
|
public XmlStreamReader(final InputStream inputStream, final String httpContentType, |
|
|
|
boolean lenient) throws IOException { |
|
|
|
final boolean lenient) throws IOException { |
|
|
|
this(is, httpContentType, lenient, null); |
|
|
|
this(inputStream, httpContentType, lenient, null); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
@ -388,7 +409,7 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws IOException if an I/O error occurs |
|
|
|
* @throws IOException if an I/O error occurs |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
@Override |
|
|
|
@Override |
|
|
|
public int read(char[] buf, int offset, int len) throws IOException { |
|
|
|
public int read(final char[] buf, final int offset, final int len) throws IOException { |
|
|
|
return reader.read(buf, offset, len); |
|
|
|
return reader.read(buf, offset, len); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -412,19 +433,18 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @return the encoding to be used |
|
|
|
* @return the encoding to be used |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
private String doRawStream(BOMInputStream bom, BOMInputStream pis, boolean lenient) |
|
|
|
private String doRawStream(final BOMInputStream bom, final BOMInputStream pis, final boolean lenient) |
|
|
|
throws IOException { |
|
|
|
throws IOException { |
|
|
|
String bomEnc = bom.getBOMCharsetName(); |
|
|
|
final String bomEnc = bom.getBOMCharsetName(); |
|
|
|
String xmlGuessEnc = pis.getBOMCharsetName(); |
|
|
|
final String xmlGuessEnc = pis.getBOMCharsetName(); |
|
|
|
String xmlEnc = getXmlProlog(pis, xmlGuessEnc); |
|
|
|
final String xmlEnc = getXmlProlog(pis, xmlGuessEnc); |
|
|
|
try { |
|
|
|
try { |
|
|
|
return calculateRawEncoding(bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
return calculateRawEncoding(bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} catch (XmlStreamReaderException ex) { |
|
|
|
} catch (final XmlStreamReaderException ex) { |
|
|
|
if (lenient) { |
|
|
|
if (lenient) { |
|
|
|
return doLenientDetection(null, ex); |
|
|
|
return doLenientDetection(null, ex); |
|
|
|
} else { |
|
|
|
|
|
|
|
throw ex; |
|
|
|
|
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
throw ex; |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -439,20 +459,18 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @return the encoding to be used |
|
|
|
* @return the encoding to be used |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
private String doHttpStream(BOMInputStream bom, BOMInputStream pis, String httpContentType, |
|
|
|
private String processHttpStream(final BOMInputStream bom, final BOMInputStream pis, final String httpContentType, |
|
|
|
boolean lenient) throws IOException { |
|
|
|
final boolean lenient) throws IOException { |
|
|
|
String bomEnc = bom.getBOMCharsetName(); |
|
|
|
final String bomEnc = bom.getBOMCharsetName(); |
|
|
|
String xmlGuessEnc = pis.getBOMCharsetName(); |
|
|
|
final String xmlGuessEnc = pis.getBOMCharsetName(); |
|
|
|
String xmlEnc = getXmlProlog(pis, xmlGuessEnc); |
|
|
|
final String xmlEnc = getXmlProlog(pis, xmlGuessEnc); |
|
|
|
try { |
|
|
|
try { |
|
|
|
return calculateHttpEncoding(httpContentType, bomEnc, |
|
|
|
return calculateHttpEncoding(httpContentType, bomEnc, xmlGuessEnc, xmlEnc, lenient); |
|
|
|
xmlGuessEnc, xmlEnc, lenient); |
|
|
|
} catch (final XmlStreamReaderException ex) { |
|
|
|
} catch (XmlStreamReaderException ex) { |
|
|
|
|
|
|
|
if (lenient) { |
|
|
|
if (lenient) { |
|
|
|
return doLenientDetection(httpContentType, ex); |
|
|
|
return doLenientDetection(httpContentType, ex); |
|
|
|
} else { |
|
|
|
|
|
|
|
throw ex; |
|
|
|
|
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
throw ex; |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -466,14 +484,14 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
private String doLenientDetection(String httpContentType, |
|
|
|
private String doLenientDetection(String httpContentType, |
|
|
|
XmlStreamReaderException ex) throws IOException { |
|
|
|
XmlStreamReaderException ex) throws IOException { |
|
|
|
if (httpContentType != null && httpContentType.startsWith("text/html")) { |
|
|
|
if (httpContentType != null && httpContentType.startsWith("text/html")) { |
|
|
|
httpContentType = httpContentType.substring("text/html".length()); |
|
|
|
httpContentType = httpContentType.substring("text/html".length()); |
|
|
|
httpContentType = "text/xml" + httpContentType; |
|
|
|
httpContentType = "text/xml" + httpContentType; |
|
|
|
try { |
|
|
|
try { |
|
|
|
return calculateHttpEncoding(httpContentType, ex.getBomEncoding(), |
|
|
|
return calculateHttpEncoding(httpContentType, ex.getBomEncoding(), |
|
|
|
ex.getXmlGuessEncoding(), ex.getXmlEncoding(), true); |
|
|
|
ex.getXmlGuessEncoding(), ex.getXmlEncoding(), true); |
|
|
|
} catch (XmlStreamReaderException ex2) { |
|
|
|
} catch (final XmlStreamReaderException ex2) { |
|
|
|
ex = ex2; |
|
|
|
ex = ex2; |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
@ -482,7 +500,7 @@ public class XmlStreamReader extends Reader { |
|
|
|
encoding = ex.getContentTypeEncoding(); |
|
|
|
encoding = ex.getContentTypeEncoding(); |
|
|
|
} |
|
|
|
} |
|
|
|
if (encoding == null) { |
|
|
|
if (encoding == null) { |
|
|
|
encoding = (defaultEncoding == null) ? UTF_8 : defaultEncoding; |
|
|
|
encoding = defaultEncoding == null ? UTF_8 : defaultEncoding; |
|
|
|
} |
|
|
|
} |
|
|
|
return encoding; |
|
|
|
return encoding; |
|
|
|
} |
|
|
|
} |
|
|
@ -496,16 +514,16 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @return the raw encoding |
|
|
|
* @return the raw encoding |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
String calculateRawEncoding(String bomEnc, String xmlGuessEnc, |
|
|
|
String calculateRawEncoding(final String bomEnc, final String xmlGuessEnc, |
|
|
|
String xmlEnc) throws IOException { |
|
|
|
final String xmlEnc) throws IOException { |
|
|
|
|
|
|
|
|
|
|
|
// BOM is Null
|
|
|
|
// BOM is Null
|
|
|
|
if (bomEnc == null) { |
|
|
|
if (bomEnc == null) { |
|
|
|
if (xmlGuessEnc == null || xmlEnc == null) { |
|
|
|
if (xmlGuessEnc == null || xmlEnc == null) { |
|
|
|
return (defaultEncoding == null ? UTF_8 : defaultEncoding); |
|
|
|
return defaultEncoding == null ? UTF_8 : defaultEncoding; |
|
|
|
} |
|
|
|
} |
|
|
|
if (xmlEnc.equals(UTF_16) && |
|
|
|
if (xmlEnc.equals(UTF_16) && |
|
|
|
(xmlGuessEnc.equals(UTF_16BE) || xmlGuessEnc.equals(UTF_16LE))) { |
|
|
|
(xmlGuessEnc.equals(UTF_16BE) || xmlGuessEnc.equals(UTF_16LE))) { |
|
|
|
return xmlGuessEnc; |
|
|
|
return xmlGuessEnc; |
|
|
|
} |
|
|
|
} |
|
|
|
return xmlEnc; |
|
|
|
return xmlEnc; |
|
|
@ -514,11 +532,11 @@ public class XmlStreamReader extends Reader { |
|
|
|
// BOM is UTF-8
|
|
|
|
// BOM is UTF-8
|
|
|
|
if (bomEnc.equals(UTF_8)) { |
|
|
|
if (bomEnc.equals(UTF_8)) { |
|
|
|
if (xmlGuessEnc != null && !xmlGuessEnc.equals(UTF_8)) { |
|
|
|
if (xmlGuessEnc != null && !xmlGuessEnc.equals(UTF_8)) { |
|
|
|
String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
if (xmlEnc != null && !xmlEnc.equals(UTF_8)) { |
|
|
|
if (xmlEnc != null && !xmlEnc.equals(UTF_8)) { |
|
|
|
String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
return bomEnc; |
|
|
|
return bomEnc; |
|
|
@ -527,18 +545,31 @@ public class XmlStreamReader extends Reader { |
|
|
|
// BOM is UTF-16BE or UTF-16LE
|
|
|
|
// BOM is UTF-16BE or UTF-16LE
|
|
|
|
if (bomEnc.equals(UTF_16BE) || bomEnc.equals(UTF_16LE)) { |
|
|
|
if (bomEnc.equals(UTF_16BE) || bomEnc.equals(UTF_16LE)) { |
|
|
|
if (xmlGuessEnc != null && !xmlGuessEnc.equals(bomEnc)) { |
|
|
|
if (xmlGuessEnc != null && !xmlGuessEnc.equals(bomEnc)) { |
|
|
|
String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
if (xmlEnc != null && !xmlEnc.equals(UTF_16) && !xmlEnc.equals(bomEnc)) { |
|
|
|
if (xmlEnc != null && !xmlEnc.equals(UTF_16) && !xmlEnc.equals(bomEnc)) { |
|
|
|
String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
return bomEnc; |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// BOM is UTF-32BE or UTF-32LE
|
|
|
|
|
|
|
|
if (bomEnc.equals(UTF_32BE) || bomEnc.equals(UTF_32LE)) { |
|
|
|
|
|
|
|
if (xmlGuessEnc != null && !xmlGuessEnc.equals(bomEnc)) { |
|
|
|
|
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
if (xmlEnc != null && !xmlEnc.equals(UTF_32) && !xmlEnc.equals(bomEnc)) { |
|
|
|
|
|
|
|
final String msg = MessageFormat.format(RAW_EX_1, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
return bomEnc; |
|
|
|
return bomEnc; |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
// BOM is something else
|
|
|
|
// BOM is something else
|
|
|
|
String msg = MessageFormat.format(RAW_EX_2, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(RAW_EX_2, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -555,9 +586,9 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @return the HTTP encoding |
|
|
|
* @return the HTTP encoding |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
String calculateHttpEncoding(String httpContentType, |
|
|
|
String calculateHttpEncoding(final String httpContentType, |
|
|
|
String bomEnc, String xmlGuessEnc, String xmlEnc, |
|
|
|
final String bomEnc, final String xmlGuessEnc, final String xmlEnc, |
|
|
|
boolean lenient) throws IOException { |
|
|
|
final boolean lenient) throws IOException { |
|
|
|
|
|
|
|
|
|
|
|
// Lenient and has XML encoding
|
|
|
|
// Lenient and has XML encoding
|
|
|
|
if (lenient && xmlEnc != null) { |
|
|
|
if (lenient && xmlEnc != null) { |
|
|
@ -565,14 +596,14 @@ public class XmlStreamReader extends Reader { |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
// Determine mime/encoding content types from HTTP Content Type
|
|
|
|
// Determine mime/encoding content types from HTTP Content Type
|
|
|
|
String cTMime = getContentTypeMime(httpContentType); |
|
|
|
final String cTMime = getContentTypeMime(httpContentType); |
|
|
|
String cTEnc = getContentTypeEncoding(httpContentType); |
|
|
|
final String cTEnc = getContentTypeEncoding(httpContentType); |
|
|
|
boolean appXml = isAppXml(cTMime); |
|
|
|
final boolean appXml = isAppXml(cTMime); |
|
|
|
boolean textXml = isTextXml(cTMime); |
|
|
|
final boolean textXml = isTextXml(cTMime); |
|
|
|
|
|
|
|
|
|
|
|
// Mime type NOT "application/xml" or "text/xml"
|
|
|
|
// Mime type NOT "application/xml" or "text/xml"
|
|
|
|
if (!appXml && !textXml) { |
|
|
|
if (!appXml && !textXml) { |
|
|
|
String msg = MessageFormat.format(HTTP_EX_3, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(HTTP_EX_3, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -580,15 +611,14 @@ public class XmlStreamReader extends Reader { |
|
|
|
if (cTEnc == null) { |
|
|
|
if (cTEnc == null) { |
|
|
|
if (appXml) { |
|
|
|
if (appXml) { |
|
|
|
return calculateRawEncoding(bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
return calculateRawEncoding(bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} else { |
|
|
|
|
|
|
|
return (defaultEncoding == null) ? US_ASCII : defaultEncoding; |
|
|
|
|
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
return defaultEncoding == null ? US_ASCII : defaultEncoding; |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
// UTF-16BE or UTF-16LE content type encoding
|
|
|
|
// UTF-16BE or UTF-16LE content type encoding
|
|
|
|
if (cTEnc.equals(UTF_16BE) || cTEnc.equals(UTF_16LE)) { |
|
|
|
if (cTEnc.equals(UTF_16BE) || cTEnc.equals(UTF_16LE)) { |
|
|
|
if (bomEnc != null) { |
|
|
|
if (bomEnc != null) { |
|
|
|
String msg = MessageFormat.format(HTTP_EX_1, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(HTTP_EX_1, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
return cTEnc; |
|
|
|
return cTEnc; |
|
|
@ -599,7 +629,25 @@ public class XmlStreamReader extends Reader { |
|
|
|
if (bomEnc != null && bomEnc.startsWith(UTF_16)) { |
|
|
|
if (bomEnc != null && bomEnc.startsWith(UTF_16)) { |
|
|
|
return bomEnc; |
|
|
|
return bomEnc; |
|
|
|
} |
|
|
|
} |
|
|
|
String msg = MessageFormat.format(HTTP_EX_2, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
final String msg = MessageFormat.format(HTTP_EX_2, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// UTF-32BE or UTF-132E content type encoding
|
|
|
|
|
|
|
|
if (cTEnc.equals(UTF_32BE) || cTEnc.equals(UTF_32LE)) { |
|
|
|
|
|
|
|
if (bomEnc != null) { |
|
|
|
|
|
|
|
final String msg = MessageFormat.format(HTTP_EX_1, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
return cTEnc; |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
// UTF-32 content type encoding
|
|
|
|
|
|
|
|
if (cTEnc.equals(UTF_32)) { |
|
|
|
|
|
|
|
if (bomEnc != null && bomEnc.startsWith(UTF_32)) { |
|
|
|
|
|
|
|
return bomEnc; |
|
|
|
|
|
|
|
} |
|
|
|
|
|
|
|
final String msg = MessageFormat.format(HTTP_EX_2, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
throw new XmlStreamReaderException(msg, cTMime, cTEnc, bomEnc, xmlGuessEnc, xmlEnc); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
@ -612,10 +660,10 @@ public class XmlStreamReader extends Reader { |
|
|
|
* @param httpContentType the HTTP content type |
|
|
|
* @param httpContentType the HTTP content type |
|
|
|
* @return The mime content type |
|
|
|
* @return The mime content type |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
static String getContentTypeMime(String httpContentType) { |
|
|
|
static String getContentTypeMime(final String httpContentType) { |
|
|
|
String mime = null; |
|
|
|
String mime = null; |
|
|
|
if (httpContentType != null) { |
|
|
|
if (httpContentType != null) { |
|
|
|
int i = httpContentType.indexOf(";"); |
|
|
|
final int i = httpContentType.indexOf(";"); |
|
|
|
if (i >= 0) { |
|
|
|
if (i >= 0) { |
|
|
|
mime = httpContentType.substring(0, i); |
|
|
|
mime = httpContentType.substring(0, i); |
|
|
|
} else { |
|
|
|
} else { |
|
|
@ -634,22 +682,25 @@ public class XmlStreamReader extends Reader { |
|
|
|
* httpContentType is NULL. |
|
|
|
* httpContentType is NULL. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param httpContentType the HTTP content type |
|
|
|
* @param httpContentType the HTTP content type |
|
|
|
* @return The content type encoding |
|
|
|
* @return The content type encoding (upcased) |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
static String getContentTypeEncoding(String httpContentType) { |
|
|
|
static String getContentTypeEncoding(final String httpContentType) { |
|
|
|
String encoding = null; |
|
|
|
String encoding = null; |
|
|
|
if (httpContentType != null) { |
|
|
|
if (httpContentType != null) { |
|
|
|
int i = httpContentType.indexOf(";"); |
|
|
|
final int i = httpContentType.indexOf(";"); |
|
|
|
if (i > -1) { |
|
|
|
if (i > -1) { |
|
|
|
String postMime = httpContentType.substring(i + 1); |
|
|
|
final String postMime = httpContentType.substring(i + 1); |
|
|
|
Matcher m = CHARSET_PATTERN.matcher(postMime); |
|
|
|
final Matcher m = CHARSET_PATTERN.matcher(postMime); |
|
|
|
encoding = (m.find()) ? m.group(1) : null; |
|
|
|
encoding = m.find() ? m.group(1) : null; |
|
|
|
encoding = (encoding != null) ? encoding.toUpperCase() : null; |
|
|
|
encoding = encoding != null ? encoding.toUpperCase(Locale.ROOT) : null; |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
return encoding; |
|
|
|
return encoding; |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
|
|
|
* Pattern capturing the encoding of the "xml" processing instruction. |
|
|
|
|
|
|
|
*/ |
|
|
|
public static final Pattern ENCODING_PATTERN = Pattern.compile( |
|
|
|
public static final Pattern ENCODING_PATTERN = Pattern.compile( |
|
|
|
"<\\?xml.*encoding[\\s]*=[\\s]*((?:\".[^\"]*\")|(?:'.[^']*'))", |
|
|
|
"<\\?xml.*encoding[\\s]*=[\\s]*((?:\".[^\"]*\")|(?:'.[^']*'))", |
|
|
|
Pattern.MULTILINE); |
|
|
|
Pattern.MULTILINE); |
|
|
@ -657,52 +708,50 @@ public class XmlStreamReader extends Reader { |
|
|
|
/** |
|
|
|
/** |
|
|
|
* Returns the encoding declared in the <?xml encoding=...?>, NULL if none. |
|
|
|
* Returns the encoding declared in the <?xml encoding=...?>, NULL if none. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param is InputStream to create the reader from. |
|
|
|
* @param inputStream InputStream to create the reader from. |
|
|
|
* @param guessedEnc guessed encoding |
|
|
|
* @param guessedEnc guessed encoding |
|
|
|
* @return the encoding declared in the <?xml encoding=...?> |
|
|
|
* @return the encoding declared in the <?xml encoding=...?> |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
* @throws IOException thrown if there is a problem reading the stream. |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
private static String getXmlProlog(InputStream is, String guessedEnc) |
|
|
|
private static String getXmlProlog(final InputStream inputStream, final String guessedEnc) |
|
|
|
throws IOException { |
|
|
|
throws IOException { |
|
|
|
String encoding = null; |
|
|
|
String encoding = null; |
|
|
|
if (guessedEnc != null) { |
|
|
|
if (guessedEnc != null) { |
|
|
|
byte[] bytes = new byte[BUFFER_SIZE]; |
|
|
|
final byte[] bytes = new byte[BUFFER_SIZE]; |
|
|
|
is.mark(BUFFER_SIZE); |
|
|
|
inputStream.mark(BUFFER_SIZE); |
|
|
|
int offset = 0; |
|
|
|
int offset = 0; |
|
|
|
int max = BUFFER_SIZE; |
|
|
|
int max = BUFFER_SIZE; |
|
|
|
int c = is.read(bytes, offset, max); |
|
|
|
int c = inputStream.read(bytes, offset, max); |
|
|
|
int firstGT = -1; |
|
|
|
int firstGT = -1; |
|
|
|
String xmlProlog = null; |
|
|
|
String xmlProlog = ""; // avoid possible NPE warning (cannot happen; this just silences the warning)
|
|
|
|
while (c != -1 && firstGT == -1 && offset < BUFFER_SIZE) { |
|
|
|
while (c != -1 && firstGT == -1 && offset < BUFFER_SIZE) { |
|
|
|
offset += c; |
|
|
|
offset += c; |
|
|
|
max -= c; |
|
|
|
max -= c; |
|
|
|
c = is.read(bytes, offset, max); |
|
|
|
c = inputStream.read(bytes, offset, max); |
|
|
|
xmlProlog = new String(bytes, 0, offset, guessedEnc); |
|
|
|
xmlProlog = new String(bytes, 0, offset, guessedEnc); |
|
|
|
firstGT = xmlProlog.indexOf('>'); |
|
|
|
firstGT = xmlProlog.indexOf('>'); |
|
|
|
} |
|
|
|
} |
|
|
|
if (firstGT == -1) { |
|
|
|
if (firstGT == -1) { |
|
|
|
if (c == -1) { |
|
|
|
if (c == -1) { |
|
|
|
throw new IOException("Unexpected end of XML stream"); |
|
|
|
throw new IOException("Unexpected end of XML stream"); |
|
|
|
} else { |
|
|
|
|
|
|
|
throw new IOException( |
|
|
|
|
|
|
|
"XML prolog or ROOT element not found on first " |
|
|
|
|
|
|
|
+ offset + " bytes"); |
|
|
|
|
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
throw new IOException( |
|
|
|
|
|
|
|
"XML prolog or ROOT element not found on first " |
|
|
|
|
|
|
|
+ offset + " bytes"); |
|
|
|
} |
|
|
|
} |
|
|
|
int bytesRead = offset; |
|
|
|
final int bytesRead = offset; |
|
|
|
if (bytesRead > 0) { |
|
|
|
if (bytesRead > 0) { |
|
|
|
is.reset(); |
|
|
|
inputStream.reset(); |
|
|
|
BufferedReader bReader = new BufferedReader(new StringReader( |
|
|
|
final BufferedReader bReader = new BufferedReader(new StringReader( |
|
|
|
xmlProlog.substring(0, firstGT + 1))); |
|
|
|
xmlProlog.substring(0, firstGT + 1))); |
|
|
|
StringBuilder prolog = new StringBuilder(); |
|
|
|
final StringBuffer prolog = new StringBuffer(); |
|
|
|
String line = bReader.readLine(); |
|
|
|
String line; |
|
|
|
while (line != null) { |
|
|
|
while ((line = bReader.readLine()) != null) { |
|
|
|
prolog.append(line); |
|
|
|
prolog.append(line); |
|
|
|
line = bReader.readLine(); |
|
|
|
|
|
|
|
} |
|
|
|
} |
|
|
|
Matcher m = ENCODING_PATTERN.matcher(prolog); |
|
|
|
final Matcher m = ENCODING_PATTERN.matcher(prolog); |
|
|
|
if (m.find()) { |
|
|
|
if (m.find()) { |
|
|
|
encoding = Objects.requireNonNull(m.group(1)).toUpperCase(); |
|
|
|
encoding = Objects.requireNonNull(m.group(1)).toUpperCase(Locale.ROOT); |
|
|
|
encoding = encoding.substring(1, encoding.length() - 1); |
|
|
|
encoding = encoding.substring(1, encoding.length() - 1); |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
|
} |
|
|
@ -712,46 +761,46 @@ public class XmlStreamReader extends Reader { |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Indicates if the MIME type belongs to the APPLICATION XML family. |
|
|
|
* Indicates if the MIME type belongs to the APPLICATION XML family. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param mime The mime type |
|
|
|
* @param mime The mime type |
|
|
|
* @return true if the mime type belongs to the APPLICATION XML family, |
|
|
|
* @return true if the mime type belongs to the APPLICATION XML family, |
|
|
|
* otherwise false |
|
|
|
* otherwise false |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
static boolean isAppXml(String mime) { |
|
|
|
static boolean isAppXml(final String mime) { |
|
|
|
return mime != null && |
|
|
|
return mime != null && |
|
|
|
(mime.equals("application/xml") || |
|
|
|
(mime.equals("application/xml") || |
|
|
|
mime.equals("application/xml-dtd") || |
|
|
|
mime.equals("application/xml-dtd") || |
|
|
|
mime.equals("application/xml-external-parsed-entity") || |
|
|
|
mime.equals("application/xml-external-parsed-entity") || |
|
|
|
(mime.startsWith("application/") && mime.endsWith("+xml"))); |
|
|
|
mime.startsWith("application/") && mime.endsWith("+xml")); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
/** |
|
|
|
* Indicates if the MIME type belongs to the TEXT XML family. |
|
|
|
* Indicates if the MIME type belongs to the TEXT XML family. |
|
|
|
* |
|
|
|
* |
|
|
|
* @param mime The mime type |
|
|
|
* @param mime The mime type |
|
|
|
* @return true if the mime type belongs to the TEXT XML family, |
|
|
|
* @return true if the mime type belongs to the TEXT XML family, |
|
|
|
* otherwise false |
|
|
|
* otherwise false |
|
|
|
*/ |
|
|
|
*/ |
|
|
|
static boolean isTextXml(String mime) { |
|
|
|
static boolean isTextXml(final String mime) { |
|
|
|
return mime != null && |
|
|
|
return mime != null && |
|
|
|
(mime.equals("text/xml") || |
|
|
|
(mime.equals("text/xml") || |
|
|
|
mime.equals("text/xml-external-parsed-entity") || |
|
|
|
mime.equals("text/xml-external-parsed-entity") || |
|
|
|
(mime.startsWith("text/") && mime.endsWith("+xml"))); |
|
|
|
mime.startsWith("text/") && mime.endsWith("+xml")); |
|
|
|
} |
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
private static final String RAW_EX_1 = |
|
|
|
private static final String RAW_EX_1 = |
|
|
|
"Invalid encoding, BOM [{0}] XML guess [{1}] XML prolog [{2}] encoding mismatch"; |
|
|
|
"Invalid encoding, BOM [{0}] XML guess [{1}] XML prolog [{2}] encoding mismatch"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String RAW_EX_2 = |
|
|
|
private static final String RAW_EX_2 = |
|
|
|
"Invalid encoding, BOM [{0}] XML guess [{1}] XML prolog [{2}] unknown BOM"; |
|
|
|
"Invalid encoding, BOM [{0}] XML guess [{1}] XML prolog [{2}] unknown BOM"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String HTTP_EX_1 = |
|
|
|
private static final String HTTP_EX_1 = |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], BOM must be NULL"; |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], BOM must be NULL"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String HTTP_EX_2 = |
|
|
|
private static final String HTTP_EX_2 = |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], encoding mismatch"; |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], encoding mismatch"; |
|
|
|
|
|
|
|
|
|
|
|
private static final String HTTP_EX_3 = |
|
|
|
private static final String HTTP_EX_3 = |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], Invalid MIME"; |
|
|
|
"Invalid encoding, CT-MIME [{0}] CT-Enc [{1}] BOM [{2}] XML guess [{3}] XML prolog [{4}], Invalid MIME"; |
|
|
|
|
|
|
|
|
|
|
|
} |
|
|
|
} |