[Htmlparser-cvs] htmlparser/src/org/htmlparser/lexer/nodes Attribute.java,1.4,1.5 TagNode.java,1.6,1
Brought to you by:
derrickoswald
Menu
▾
▴
[Htmlparser-cvs] htmlparser/src/org/htmlparser/lexer/nodes Attribute.java,1.4,1.5 TagNode.java,1.6,1.7
|
From: <der...@us...> - 2003-09-07 21:29:02
|
Update of /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer/nodes
In directory sc8-pr-cvs1:/tmp/cvs-serv25784
Modified Files:
Attribute.java TagNode.java
Log Message:
Fix setAttribute and optimize getAttribute for speed.
Index: Attribute.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer/nodes/Attribute.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -C2 -d -r1.4 -r1.5
*** Attribute.java 24 Aug 2003 21:59:41 -0000 1.4
--- Attribute.java 7 Sep 2003 21:28:03 -0000 1.5
***************
*** 37,41 ****
/**
* An attribute within a tag.
! * <p>If Name is null, it's whitepace and Value has the text.
* <p>If Name is not null, and Value is null it's a standalone attribute.
* <p>If Name is not null, and Value is "", and Quote is zero it's an empty attribute.
--- 37,41 ----
/**
* An attribute within a tag.
! * <p>If Name is null, it is whitepace and Value has the text.
* <p>If Name is not null, and Value is null it's a standalone attribute.
* <p>If Name is not null, and Value is "", and Quote is zero it's an empty attribute.
***************
*** 156,159 ****
--- 156,160 ----
* Get the value of the attribute.
* The part after the equals sign, or the text if it's just a whitepace 'attribute'.
+ * <em>NOTE: This does not include any quotes that may have enclosed the value.</em>
* @return The value, or <code>null</code> if it's a stand-alone attribute,
* or the text if it's just a whitepace 'attribute'.
***************
*** 168,171 ****
--- 169,228 ----
/**
+ * Get the raw value of the attribute.
+ * The part after the equals sign, or the text if it's just a whitepace 'attribute'.
+ * @return The value, or <code>null</code> if it's a stand-alone attribute,
+ * or the text if it's just a whitepace 'attribute'.
+ */
+ public String getRawValue ()
+ {
+ char quote;
+ StringBuffer buffer;
+ String ret;
+
+ ret = getValue ();
+ if (null != ret && (0 != (quote = getQuote ())))
+ {
+ buffer = new StringBuffer (ret.length() + 2);
+ buffer.append (quote);
+ buffer.append (ret);
+ buffer.append (quote);
+ ret = buffer.toString ();
+ }
+
+ return (ret);
+ }
+
+ /**
+ * Get the raw value of the attribute.
+ * The part after the equals sign, or the text if it's just a whitepace 'attribute'.
+ * @return The value, or <code>null</code> if it's a stand-alone attribute,
+ * or the text if it's just a whitepace 'attribute'.
+ */
+ public void getRawValue (StringBuffer buffer)
+ {
+ char quote;
+
+ if (null == mValue)
+ {
+ if (0 <= mValueStart)
+ {
+ if (0 != (quote = getQuote ()))
+ buffer.append (quote);
+ mPage.getText (buffer, mValueStart, mValueEnd);
+ if (0 != quote)
+ buffer.append (quote);
+ }
+ }
+ else
+ {
+ if (0 != (quote = getQuote ()))
+ buffer.append (quote);
+ buffer.append (mValue);
+ if (0 != quote)
+ buffer.append (quote);
+ }
+ }
+
+ /**
* Get the quote, if any, surrounding the value of the attribute, if any.
* @return Either ' or " if the attribute value was quoted, or zero
***************
*** 194,218 ****
public void toString (StringBuffer buffer)
{
- String value;
String name;
- value = getValue ();
name = getName ();
if (null == name)
! {
! if (value != null)
! buffer.append (value);
! }
else
{
buffer.append (name);
! if (null != value)
{
buffer.append ("=");
! if (0 != getQuote ())
! buffer.append (getQuote ());
! buffer.append (value);
! if (0 != getQuote ())
! buffer.append (getQuote ());
}
}
--- 251,266 ----
public void toString (StringBuffer buffer)
{
String name;
name = getName ();
if (null == name)
! getRawValue (buffer);
else
{
buffer.append (name);
! if (0 <= mValueStart)
{
buffer.append ("=");
! getRawValue (buffer);
}
}
Index: TagNode.java
===================================================================
RCS file: /cvsroot/htmlparser/htmlparser/src/org/htmlparser/lexer/nodes/TagNode.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -C2 -d -r1.6 -r1.7
*** TagNode.java 3 Sep 2003 23:36:19 -0000 1.6
--- TagNode.java 7 Sep 2003 21:28:03 -0000 1.7
***************
*** 38,41 ****
--- 38,42 ----
import org.htmlparser.util.NodeList;
import org.htmlparser.util.ParserException;
+ import org.htmlparser.util.Translate;
/**
***************
*** 121,152 ****
/**
! * In case the tag is parsed at the scan method this will return value of a
! * parameter not implemented yet
! * @param name of parameter
*/
public String getAttribute (String name)
{
! return ((String)getAttributes().get(name.toUpperCase()));
}
/**
* Set attribute with given key, value pair.
! * @param key
! * @param value
*/
! public void setAttribute(String key, String value)
{
! getAttributes ().put(key,value);
}
/**
! * In case the tag is parsed at the scan method this will return value of a
! * parameter not implemented yet
! * @param name of parameter
* @deprecated use getAttribute instead
*/
! public String getParameter(String name)
{
! return (String)getAttributes().get (name.toUpperCase());
}
--- 122,293 ----
/**
! * Create a tag with the location and attributes provided
! * @param page The page this tag was read from.
! * @param start The starting offset of this node within the page.
! * @param end The ending offset of this node within the page.
! * @param attributes The list of attributes that were parsed in this tag.
! * @see Attribute
! */
! public TagNode ()
! {
! super (null, -1, -1);
! mAttributes = new Vector ();
! }
!
! /**
! * Returns the value of an attribute.
! * @param name Name of attribute, case insensitive.
! * @return The value associated with the attribute or null if it does
! * not exist, or is a stand-alone or
*/
public String getAttribute (String name)
{
! Vector attributes;
! int size;
! Attribute attribute;
! String string;
! String ret;
!
! ret = null;
!
! attributes = getAttributesEx ();
! if (name.equalsIgnoreCase (TAGNAME))
! ret = ((Attribute)attributes.elementAt (0)).getName ();
! else
! {
! size = attributes.size ();
! for (int i = 1; i < size; i++)
! {
! attribute = (Attribute)attributes.elementAt (i);
! string = attribute.getName ();
! if ((null != string) && name.equalsIgnoreCase (string))
! {
! ret = attribute.getValue ();
! i = size; // exit fast
! }
! }
! }
!
! return (ret);
}
/**
* Set attribute with given key, value pair.
! * Figures out a quote character to use if necessary.
! * @param key The name of the attribute.
! * @param value The value of the attribute.
*/
! public void setAttribute (String key, String value)
{
! char ch;
! boolean needed;
! boolean singleq;
! boolean doubleq;
! String ref;
! StringBuffer buffer;
! char quote;
!
! // first determine if there's whitespace in the value
! // and while we'return at it find a suitable quote character
! needed = false;
! singleq = true;
! doubleq = true;
! for (int i = 0; i < value.length (); i++)
! {
! ch = value.charAt (i);
! if (Character.isWhitespace (ch))
! needed = true;
! else if ('\'' == ch)
! singleq = false;
! else if ('"' == ch)
! doubleq = false;
! }
!
! // now apply quoting
! if (needed)
! {
! if (doubleq)
! quote = '"';
! else if (singleq)
! quote = '\'';
! else
! {
! // uh-oh, we need to convert some quotes into character references
! // convert all double quotes into "
! quote = '"';
! ref = Translate.convertToString (quote);
! // JDK 1.4: value = value.replaceAll ("\"", ref);
! buffer = new StringBuffer (value.length() * 5);
! for (int i = 0; i < value.length (); i++)
! {
! ch = value.charAt (i);
! if ('"' == ch)
! buffer.append (ref);
! else
! buffer.append (ch);
! }
! value = buffer.toString ();
! }
! }
! else
! quote = 0;
! setAttribute (key, value, quote);
}
/**
! * Set attribute with given key, value pair where the value is quoted by quote.
! * @param key The name of the attribute.
! * @param value The value of the attribute.
! * @param quote The quote character to be used around value.
! * If zero, it is an unquoted value.
! */
! public void setAttribute (String key, String value, char quote)
! {
! setAttribute (new Attribute (key, value, quote));
! }
!
! /**
! * Set an attribute.
! * This replaces an attribute of the same name.
! * To set the zeroth attribute (the tag name), use setTagName().
! * @param attribute The attribute to set.
! */
! public void setAttribute (Attribute attribute)
! {
! boolean replaced;
! Vector attributes;
! String name;
! Attribute test;
! String test_name;
!
! replaced = false;
! attributes = getAttributesEx ();
! if (0 < attributes.size ())
! {
! name = attribute.getName ();
! for (int i = 1; i < attributes.size (); i++)
! {
! test = (Attribute)attributes.elementAt (i);
! test_name = test.getName ();
! if (null != test_name)
! if (test_name.equalsIgnoreCase (name))
! {
! attributes.setElementAt (attribute, i);
! replaced = true;
! }
! }
! }
! if (!replaced)
! attributes.addElement (attribute);
! }
!
! /**
! * Eqivalent to <code>getAttribute (name)</code>.
! * @param name Name of attribute.
* @deprecated use getAttribute instead
*/
! public String getParameter (String name)
{
! return (getAttribute (name));
}
***************
*** 158,162 ****
* @return Returns a special hashtable of attributes in two element String arrays.
*/
! public Vector getAttributesEx()
{
return mAttributes;
--- 299,303 ----
* @return Returns a special hashtable of attributes in two element String arrays.
*/
! public Vector getAttributesEx ()
{
return mAttributes;
***************
*** 165,171 ****
/**
* Gets the attributes in the tag.
! * @return Returns a Hashtable of attributes
*/
! public Hashtable getAttributes()
{
Vector attributes;
--- 306,312 ----
/**
* Gets the attributes in the tag.
! * @return Returns a Hashtable of attributes.
*/
! public Hashtable getAttributes ()
{
Vector attributes;
***************
*** 188,213 ****
if (null != attribute.getName ())
{
! value = attribute.getValue ();
! if ('\'' == attribute.getQuote ())
! {
! _value = new StringBuffer (value.length () + 2);
! _value.append ("'");
! _value.append (value);
! _value.append ("'");
! value = _value.toString ();
! }
! else if ('"' == attribute.getQuote ())
{
! _value = new StringBuffer (value.length () + 2);
! _value.append ("\"");
! _value.append (value);
! _value.append ("\"");
! value = _value.toString ();
}
- else if ((null != value) && value.equals (""))
- value = NOTHING;
if (null == value)
value = NULLVALUE;
! ret.put (attribute.getName (), value);
}
}
--- 329,343 ----
if (null != attribute.getName ())
{
! if (0 != attribute.getQuote ())
! value = attribute.getRawValue ();
! else
{
! value = attribute.getValue ();
! if ((null != value) && value.equals (""))
! value = NOTHING;
}
if (null == value)
value = NULLVALUE;
! ret.put (attribute.getName ().toUpperCase (), value);
}
}
***************
*** 219,230 ****
}
! public String getTagName(){
! return getParameter(TAGNAME);
}
/**
! * Return the text contained in this tag
*/
! public String getText()
{
return (mPage.getText (elementBegin () + 1, elementEnd () - 1));
--- 349,402 ----
}
! /**
! * Return the name of this tag.
! * <p>
! * <em>
! * Note: This value is converted to uppercase.
! * To get at the original case version of the tag name use:
! * <pre>
! * getAttribute (TagNode.TAGNAME);
! * </pre>
! * </em>
! * @return The tag name.
! */
! public String getTagName ()
! {
! return (getAttribute (TAGNAME).toUpperCase ());
}
/**
! * Set the name of this tag.
! * This creates or replaces the first attribute of the tag (the
! * zeroth element of the attribute vector).
! * @param name The tag name.
*/
! public void setTagName (String name)
! {
! Attribute attribute;
! Vector attributes;
! Attribute zeroth;
!
! attribute = new Attribute (name, null, (char)0);
! attributes = getAttributesEx ();
! if (0 == attributes.size ())
! // nothing added yet
! attributes.addElement (attribute);
! else
! {
! zeroth = (Attribute)attributes.elementAt (0);
! // check forn attribute that looks like a name
! if ((null == zeroth.getValue ()) && (0 == zeroth.getQuote ()))
! attributes.setElementAt (attribute, 0);
! else
! attributes.insertElementAt (attribute, 0);
! }
! }
!
! /**
! * Return the text contained in this tag.
! * @return The complete contents of the tag (within the angle brackets).
! */
! public String getText ()
{
return (mPage.getText (elementBegin () + 1, elementEnd () - 1));
***************
*** 282,287 ****
* @param tagBegin The nodeBegin to set
*/
! public void setTagBegin(int tagBegin) {
! this.nodeBegin = tagBegin;
}
--- 454,460 ----
* @param tagBegin The nodeBegin to set
*/
! public void setTagBegin (int tagBegin)
! {
! nodeBegin = tagBegin;
}
***************
*** 290,294 ****
* @return The nodeBegin value.
*/
! public int getTagBegin() {
return (nodeBegin);
}
--- 463,468 ----
* @return The nodeBegin value.
*/
! public int getTagBegin ()
! {
return (nodeBegin);
}
***************
*** 298,303 ****
* @param tagEnd The nodeEnd to set
*/
! public void setTagEnd(int tagEnd) {
! this.nodeEnd = tagEnd;
}
--- 472,478 ----
* @param tagEnd The nodeEnd to set
*/
! public void setTagEnd (int tagEnd)
! {
! nodeEnd = tagEnd;
}
***************
*** 306,310 ****
* @return The nodeEnd value.
*/
! public int getTagEnd() {
return (nodeEnd);
}
--- 481,486 ----
* @return The nodeEnd value.
*/
! public int getTagEnd ()
! {
return (nodeEnd);
}
***************
*** 323,328 ****
}
! public String toPlainTextString() {
! return EMPTY_STRING;
}
--- 499,505 ----
}
! public String toPlainTextString ()
! {
! return (EMPTY_STRING);
}
***************
*** 331,335 ****
* @see org.htmlparser.Node#toHtml()
*/
! public String toHtml()
{
StringBuffer ret;
--- 508,512 ----
* @see org.htmlparser.Node#toHtml()
*/
! public String toHtml ()
{
StringBuffer ret;
***************
*** 362,366 ****
* Print the contents of the tag
*/
! public String toString()
{
String tag;
--- 539,543 ----
* Print the contents of the tag
*/
! public String toString ()
{
String tag;
***************
*** 395,399 ****
* @see org.htmlparser.Node#collectInto(NodeList, String)
*/
! public void collectInto(NodeList collectionList, String filter)
{
}
--- 572,576 ----
* @see org.htmlparser.Node#collectInto(NodeList, String)
*/
! public void collectInto (NodeList collectionList, String filter)
{
}
***************
*** 404,408 ****
* @deprecated This method is deprecated. Use getAttributes() instead.
*/
! public Hashtable getParsed() {
return getAttributes ();
}
--- 581,586 ----
* @deprecated This method is deprecated. Use getAttributes() instead.
*/
! public Hashtable getParsed ()
! {
return getAttributes ();
}
***************
*** 417,421 ****
* @return Hashtable
*/
! public Hashtable redoParseAttributes()
{
mAttributes = null;
--- 595,599 ----
* @return Hashtable
*/
! public Hashtable redoParseAttributes ()
{
mAttributes = null;
***************
*** 424,431 ****
}
! public void accept(Object visitor) {
}
! public String getType() {
return TYPE;
}
--- 602,611 ----
}
! public void accept (Object visitor)
! {
}
! public String getType ()
! {
return TYPE;
}
***************
*** 436,444 ****
* @return boolean
*/
! public boolean isEmptyXmlTag() {
return emptyXmlTag;
}
! public void setEmptyXmlTag(boolean emptyXmlTag) {
this.emptyXmlTag = emptyXmlTag;
}
--- 616,626 ----
* @return boolean
*/
! public boolean isEmptyXmlTag ()
! {
return emptyXmlTag;
}
! public void setEmptyXmlTag (boolean emptyXmlTag)
! {
this.emptyXmlTag = emptyXmlTag;
}
|
