cziegeler 01/10/23 05:04:35
Modified: src/org/apache/cocoon/generation StreamGenerator.java
src/org/apache/cocoon/util PostInputStream.java
Log:
Patch for Bug 4124 StreamGenerator does not preserve XML encoding PI
Submitted by: kingadziembowska [[EMAIL PROTECTED]]Reviewed by:
Revision Changes Path
1.8 +99 -19
xml-cocoon2/src/org/apache/cocoon/generation/StreamGenerator.java
Index: StreamGenerator.java
===================================================================
RCS file:
/home/cvs/xml-cocoon2/src/org/apache/cocoon/generation/StreamGenerator.java,v
retrieving revision 1.7
retrieving revision 1.8
diff -u -r1.7 -r1.8
--- StreamGenerator.java 2001/10/11 07:28:21 1.7
+++ StreamGenerator.java 2001/10/23 12:04:35 1.8
@@ -13,8 +13,9 @@
import org.apache.cocoon.ProcessingException;
import org.apache.cocoon.ResourceNotFoundException;
import org.apache.cocoon.components.parser.Parser;
-import org.apache.cocoon.environment.Request;
+import org.apache.avalon.excalibur.pool.Poolable;
import org.apache.cocoon.environment.http.HttpEnvironment;
+import javax.servlet.http.HttpServletRequest;
import org.apache.cocoon.util.PostInputStream;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;
@@ -43,9 +44,10 @@
* number of bytes read is equal to the getContentLength() value.
*
* @author <a href="mailto:[EMAIL PROTECTED]";>Kinga Dziembowski</a>
- * @version $Revision: 1.7 $ $Date: 2001/10/11 07:28:21 $
+ * @version $Revision: 1.8 $ $Date: 2001/10/23 12:04:35 $
*/
-public class StreamGenerator extends ComposerGenerator {
+public class StreamGenerator extends ComposerGenerator implements Poolable
+{
public static final String CLASS = StreamGenerator.class.getName();
/** The parameter holding the name associated with the xml data **/
@@ -70,31 +72,27 @@
/**
* Generate XML data out of request InputStream.
*/
- public void generate() throws IOException, SAXException, ProcessingException {
+ public void generate() throws IOException, SAXException, ProcessingException
+ {
Parser parser = null;
String parameter = parameters.getParameter(StreamGenerator.FORM_NAME, null);
int len = 0;
-
- try {
- Request request = (Request)objectModel.get(Constants.REQUEST_OBJECT);
- if
(request.getContentType().equals("application/x-www-form-urlencoded")) {
+ String contentType = null;
+ try
+ {
+ HttpServletRequest request = (HttpServletRequest)
objectModel.get(HttpEnvironment.HTTP_REQUEST_OBJECT);
+ contentType = request.getContentType();
+ if (contentType.startsWith("application/x-www-form-urlencoded")) {
String sXml = request.getParameter(parameter);
inputSource = new InputSource(new StringReader(sXml));
- } else if (request.getContentType().equals("text/plain") ||
- request.getContentType().equals("text/xml") ||
- request.getContentType().equals("application/xml")) {
+ } else if (contentType.startsWith("text/plain") ||
+ contentType.startsWith("text/xml") ||
+ contentType.startsWith("application/xml")) {
len = request.getContentLength();
if (len > 0) {
- // we have hopefully an http request here
- javax.servlet.http.HttpServletRequest httpRequest =
-
(javax.servlet.http.HttpServletRequest)objectModel.get(HttpEnvironment.HTTP_REQUEST_OBJECT);
- if (httpRequest != null) {
- PostInputStream anStream = new
PostInputStream(httpRequest.getInputStream(), len);
+ PostInputStream anStream = new
PostInputStream(request.getInputStream(), len);
inputSource = new InputSource(anStream);
- } else {
- throw new IOException("No http request object found");
- }
} else {
throw new IOException("getContentLen() == 0");
}
@@ -105,6 +103,11 @@
if (getLogger().isDebugEnabled()) {
getLogger().debug("processing stream ContentType= " +
request.getContentType() + "ContentLen= " + len);
}
+ String charset = getCharacterEncoding(request, contentType) ;
+ if( charset != null)
+ {
+ this.inputSource.setEncoding(charset);
+ }
parser = (Parser)this.manager.lookup(Parser.ROLE);
parser.setContentHandler(super.contentHandler);
parser.setLexicalHandler(super.lexicalHandler);
@@ -124,5 +127,82 @@
}
}
}
+
+ /**
+ * Content type HTTP header can contains character encodinf info
+ * for ex. Content-Type: text/xml; charset=UTF-8
+ * If the servlet is following spec 2.3 and higher the servlet API can be used
to retrieve character encoding part of
+ * Content-Type header. Some containers can choose to not unpack charset info -
the spec is not strong about it.
+ * in any case this method can be used as a latest resource to retrieve the
passed charset value.
+ * <code>null</code> is returned.
+ * It is very common mistake to send : Content-Type: text/xml; charset="UTF-8".
+ * Some containers are not filtering this mistake and the processing results in
exception..
+ * The getCharacterEncoding() compensates for above mistake.
+ *
+ * @param contentType value associated with Content-Type HTTP header.
+ */
+ public String getCharacterEncoding(HttpServletRequest req, String contentType)
+ {
+ String charencoding = null;
+ String charset = "charset=";
+ if (contentType == null)
+ {
+ return (null);
+ }
+ int idx = contentType.indexOf(charset);
+ if (idx == -1)
+ {
+ return (null);
+ }
+ try
+ {
+ charencoding = req.getCharacterEncoding();
+
+ if ( charencoding != null)
+ {
+ getLogger().debug("charset from container: " +
charencoding);
+ charencoding = charencoding.trim();
+ if ((charencoding.length() > 2) &&
(charencoding.startsWith("\""))&& (charencoding.endsWith("\"")))
+ {
+ charencoding = charencoding.substring(1,
charencoding.length() - 1);
+ }
+ getLogger().debug("charset from container clean: " +
charencoding);
+ return (charencoding);
+ }
+ else
+ {
+
+ return extractCharset( contentType, idx );
+ }
+ }
+ catch(Throwable e)
+ {
+ // We will be there if the container do not implement
getCharacterEncoding() method
+ return extractCharset( contentType, idx );
+ }
+ }
+
+
+ protected String extractCharset(String contentType, int idx)
+ {
+ String charencoding = null;
+ String charset = "charset=";
+
+ getLogger().debug("charset from extractCharset");
+ charencoding = contentType.substring(idx + charset.length());
+ int idxEnd = charencoding.indexOf(";");
+ if (idxEnd != -1)
+ {
+ charencoding = charencoding.substring(0, idxEnd);
+ }
+ charencoding = charencoding.trim();
+ if ((charencoding.length() > 2) && (charencoding.startsWith("\""))&&
(charencoding.endsWith("\"")))
+ {
+ charencoding = charencoding.substring(1, charencoding.length()
- 1);
+ }
+ getLogger().debug("charset from extractCharset: " + charencoding);
+ return (charencoding.trim());
+
+ }
}
1.6 +13 -4 xml-cocoon2/src/org/apache/cocoon/util/PostInputStream.java
Index: PostInputStream.java
===================================================================
RCS file: /home/cvs/xml-cocoon2/src/org/apache/cocoon/util/PostInputStream.java,v
retrieving revision 1.5
retrieving revision 1.6
diff -u -r1.5 -r1.6
--- PostInputStream.java 2001/10/11 07:28:25 1.5
+++ PostInputStream.java 2001/10/23 12:04:35 1.6
@@ -14,7 +14,7 @@
* It allows to control read operation, restricting the number of bytes read to the
value returned by getContentLen() method.
*
* @author <a href="mailto:[EMAIL PROTECTED]";>Kinga Dziembowski</a>
- * @version $Id: PostInputStream.java,v 1.5 2001/10/11 07:28:25 cziegeler Exp $
+ * @version $Id: PostInputStream.java,v 1.6 2001/10/23 12:04:35 cziegeler Exp $
*/
public class PostInputStream extends InputStream {
@@ -151,7 +151,8 @@
if (m_bytesRead == m_contentLen) {
return -1;
}
- int num = m_inputStream.read(buffer, offset, len);
+ int available = Math.min( available(), len );
+ int num = m_inputStream.read( buffer, offset, available );
if (num > 0) {
m_bytesRead += num;
}
@@ -185,8 +186,16 @@
* @exception IOException if an I/O error occurs.
*/
public synchronized long skip(long n) throws IOException {
- checkOpen();
- return m_inputStream.skip(n);
+ checkOpen();
+ if ( m_bytesRead == m_contentLen )
+ {
+ return ( 0 );
+ }
+ else
+ {
+ return ( m_inputStream.skip( n ) );
+ }
+
}
/**
----------------------------------------------------------------------
In case of troubles, e-mail: [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
