jkesselm 01/05/17 14:43:57
Modified: java/src/org/apache/xml/dtm/sax2dtm Tag: DTM_EXP
SAX2DTM.java
Log:
Revised text-coalescence code, improved handling of CDATA Sections.
Revision Changes Path
No revision
No revision
1.1.2.13 +47 -81
xml-xalan/java/src/org/apache/xml/dtm/sax2dtm/Attic/SAX2DTM.java
Index: SAX2DTM.java
===================================================================
RCS file:
/home/cvs/xml-xalan/java/src/org/apache/xml/dtm/sax2dtm/Attic/SAX2DTM.java,v
retrieving revision 1.1.2.12
retrieving revision 1.1.2.13
diff -u -r1.1.2.12 -r1.1.2.13
--- SAX2DTM.java 2001/05/16 23:29:00 1.1.2.12
+++ SAX2DTM.java 2001/05/17 21:43:55 1.1.2.13
@@ -146,9 +146,14 @@
/** Namespace support, only relevent at construction time. */
transient private IntStack m_contextIndexes = new IntStack();
- /** Type of text event. */
+ /** Type of next characters() event within text block in prgress. */
transient private int m_textType = DTM.TEXT_NODE;
+ /** Type of coalesced text block. See logic in the characters()
+ * method.
+ * */
+ transient private int m_coalescedTextType = DTM.TEXT_NODE;
+
/** The SAX Document locator */
transient private Locator m_locator = null;
@@ -217,10 +222,11 @@
private static final int ENTITY_FIELDS_PER = 4;
/**
- * Flag tells if text is pending that we need to
- * check for whitespace stripping.
+ * The starting offset within m_chars for the text or
+ * CDATA_SECTION node currently being acumulated,
+ * or -1 if there is no text node in progress
*/
- private boolean m_textPending = false;
+ private int m_textPendingStart = -1;
/**
* Construct a SAX2DTM object ready to be constructed from SAX2
@@ -799,9 +805,8 @@
m_info.setElementAt(expandedTypeID, startInfo + OFFSET_EXPANDEDNAMEID);
m_info.setElementAt(dataOrPrefix, startInfo + OFFSET_DATA_OR_QNAME);
- // Note that we don't want firstChild or nextSibling to be processed
until
- // charactersFlush() is called.
- if (!m_textPending)
+ // Text nodes no longer need special handling, because we
+ // don't add them until they're complete
{
if (DTM.NULL != parentIndex && type != DTM.ATTRIBUTE_NODE
&& type != DTM.NAMESPACE_NODE)
@@ -1278,63 +1283,37 @@
}
/**
- * Check the last text node to see if it should be stripped.
+ * Check whether accumulated text should be stripped; if not,
+ * append the appropriate flavor of text/cdata node.
*/
protected void charactersFlush()
{
-
- if (m_textPending)
+ if (m_textPendingStart >= 0 ) // -1 indicates no-text-in-progress
{
- m_textPending = false;
-
- boolean didStrip = false;
- int lastNodeIdentity = m_size;
+ int length=m_chars.size()-m_textPendingStart;
+ boolean doStrip = false;
if (getShouldStripWhitespace())
- {
- int dataIndex = getNodeInfoNoWait(lastNodeIdentity,
- OFFSET_DATA_OR_QNAME);
- int offset = m_data.elementAt(dataIndex);
- int length = m_data.elementAt(dataIndex + 1);
-
- if (m_chars.isWhitespace(offset, length))
- {
- m_chars.setLength(m_chars.size() - length);
- m_info.setSize(m_info.size() - getNodeInfoBlockSize());
-
- didStrip = true;
- }
- }
-
- if (!didStrip)
- {
- m_size++;
-
- int parentIndex = getNodeInfoNoWait(lastNodeIdentity, OFFSET_PARENT);
-
- if (DTM.NULL != parentIndex)
- {
- int startParentInfo = parentIndex * NODEINFOBLOCKSIZE;
-
- if (NOTPROCESSED
- == m_info.elementAt(startParentInfo + OFFSET_FIRSTCHILD))
- {
- m_info.setElementAt(lastNodeIdentity,
- startParentInfo + OFFSET_FIRSTCHILD);
- }
- }
-
- int prev = getNodeInfoNoWait(lastNodeIdentity, OFFSET_PREVSIBLING);
+ {
+ doStrip=m_chars.isWhitespace(m_textPendingStart, length);
+ }
- if (DTM.NULL != prev)
- {
- m_info.setElementAt(lastNodeIdentity,
- (prev * getNodeInfoBlockSize())
- + OFFSET_NEXTSIBLING);
- }
+ if(doStrip)
+ m_chars.setLength(m_textPendingStart); // Discard accumulated text
+ else
+ {
+ int exName = m_ent.getExpandedNameID(DTM.TEXT_NODE);
- m_previous = lastNodeIdentity;
- }
+ int nodeIndex= addNode(m_coalescedTextType,exName,m_level,
+ m_parents.peek(),m_previous,
+ m_textPendingStart,false);
+ // %REVIEW% I _think_ I've got this right...
+ m_data.setElementAt(nodeIndex,length);
+ }
+
+ // Reset for next text block
+ m_textPendingStart=-1;
+ m_textType=m_coalescedTextType=DTM.TEXT_NODE;
}
}
@@ -1814,32 +1793,19 @@
*/
public void characters(char ch[], int start, int length) throws
SAXException
{
-
- int exName = m_ent.getExpandedNameID(DTM.TEXT_NODE);
- int dataIndex = m_data.size();
-
- m_data.addElement(m_chars.length());
- m_data.addElement(length);
- m_chars.append(ch, start, length);
-
- if (m_textPending)
- {
- int lastNodeIdentity = m_size;
-
- dataIndex = getNodeInfoNoWait(lastNodeIdentity, OFFSET_DATA_OR_QNAME)
- + 1;
-
- m_data.setElementAt(m_data.elementAt(dataIndex) + length, dataIndex);
- }
- else
- {
- m_textPending = true;
-
- addNode(m_textType, exName, m_level, m_parents.peek(), m_previous,
- dataIndex, false);
+ if(m_textPendingStart==-1) // First one in this block
+ {
+ m_textPendingStart=m_chars.size();
+ m_coalescedTextType=m_textType;
+ }
+ m_chars.append(ch,start,length);
- m_size--; // doesn't really exist until charactersFlush.
- }
+ // Type logic: If all adjacent text is CDATASections, the
+ // concatentated text is treated as a single CDATASection (see
+ // initialization above). If any were ordinary Text, the whole
+ // thing is treated as Text. This may be worth %REVIEW%ing.
+ if(m_textType==DTM.TEXT_NODE)
+ m_coalescedTextType=DTM.TEXT_NODE;
}
/**
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
