private static String getMetaValue(Source source, String key) {
for (int pos = 0; pos < source.length(); ) {
StartTag startTag = source.getNextStartTag(pos, "name", key, false);
if (startTag == null) return null;
if (startTag.getName() == HTMLElementName.META)
return startTag.getAttributeValue("content"); // Attribute values are automatically decoded
pos = startTag.getEnd();
}
return null;
}
final List<Element> getChildElements(int depth) {
if (depth != -1) this.depth = depth;
if (childElements == null) {
if (!Config.IncludeServerTagsInElementHierarchy && end == startTag.end) {
childElements = Collections.emptyList();
} else {
final int childDepth = (depth == -1 ? -1 : depth + 1);
childElements = new ArrayList<Element>();
int pos = Config.IncludeServerTagsInElementHierarchy ? begin + 1 : startTag.end;
final int maxChildBegin =
(Config.IncludeServerTagsInElementHierarchy || endTag == null) ? end : endTag.begin;
while (true) {
final StartTag childStartTag = source.getNextStartTag(pos);
if (childStartTag == null || childStartTag.begin >= maxChildBegin) break;
if (Config.IncludeServerTagsInElementHierarchy) {
if (childStartTag.begin < startTag.end
&& !childStartTag.getTagType().isServerTag()
&& !startTag.getTagType().isServerTag()) {
// A start tag is found within another start tag, but neither is a server tag.
// This only legitimately happens in very rare cases like entity definitions in
// doctype.
// We don't want to include the child elements in the hierarchy.
pos = childStartTag.end;
continue;
}
} else if (childStartTag.getTagType().isServerTag()) {
pos = childStartTag.end;
continue;
}
final Element childElement = childStartTag.getElement();
if (childElement.end > end) {
if (source.logger.isInfoEnabled())
source.logger.info(
"Child "
+ childElement.getDebugInfo()
+ " extends beyond end of parent "
+ getDebugInfo());
if (!INCLUDE_INCORRECTLY_NESTED_CHILDREN_IN_HIERARCHY) {
pos = childElement.end;
continue;
}
}
childElement.getChildElements(childDepth);
if (childElement.parentElement
== Element.NOT_CACHED) { // make sure element was not added as a child of a descendent
// element (can happen with overlapping elements)
childElement.parentElement = this;
childElements.add(childElement);
}
pos = childElement.end;
}
}
}
return childElements;
}
public String getDebugInfo() {
if (this == NOT_CACHED) return "NOT_CACHED";
final StringBuilder sb = new StringBuilder();
sb.append("Element ");
startTag.appendDebugTag(sb);
if (!isEmpty()) sb.append('-');
if (endTag != null) sb.append(endTag);
sb.append(' ');
startTag.appendDebugTagType(sb);
sb.append(super.getDebugInfo());
return sb.toString();
}
/**
* Returns the parent of this element in the document element hierarchy.
*
* <p>The {@link Source#fullSequentialParse()} method must be called (either explicitly or
* implicitly) immediately after construction of the <code>Source</code> object if this method is
* to be used. An <code>IllegalStateException</code> is thrown if a full sequential parse has not
* been performed or if it was performed after this element was found.
*
* <p>This method returns <code>null</code> for a <a href="Source.html#TopLevelElement">top-level
* element</a>, as well as any element formed from a {@linkplain TagType#isServerTag() server
* tag}, regardless of whether it is nested inside a normal element.
*
* <p>See the {@link Source#getChildElements()} method for more details.
*
* @return the parent of this element in the document element hierarchy, or <code>null</code> if
* this element is a <a href="Source.html#TopLevelElement">top-level element</a>.
* @throws IllegalStateException if a {@linkplain Source#fullSequentialParse() full sequential
* parse} has not been performed or if it was performed after this element was found.
* @see #getChildElements()
*/
public Element getParentElement() {
if (parentElement == Element.NOT_CACHED) {
if (!source.wasFullSequentialParseCalled())
throw new IllegalStateException(
"This operation is only possible after a full sequential parse has been performed");
if (startTag.isOrphaned())
throw new IllegalStateException(
"This operation is only possible if a full sequential parse was performed immediately after construction of the Source object");
source.getChildElements();
if (parentElement == Element.NOT_CACHED) parentElement = null;
}
return parentElement;
}
/**
* Indicates whether this element is an <a target="_blank"
* href="http://www.w3.org/TR/REC-xml#dt-eetag">empty-element tag</a>.
*
* <p>This is equivalent to {@link #getStartTag()}<code>.</code>{@link
* StartTag#isEmptyElementTag() isEmptyElementTag()}.
*
* @return <code>true</code> if this element is an <a target="_blank"
* href="http://www.w3.org/TR/REC-xml#dt-eetag">empty-element tag</a>, otherwise <code>false
* </code>.
*/
public boolean isEmptyElementTag() {
return startTag.isEmptyElementTag();
}
/**
* Returns the {@linkplain StartTag#getName() name} of the {@linkplain #getStartTag() start tag}
* of this element, always in lower case.
*
* <p>This is equivalent to {@link #getStartTag()}<code>.</code>{@link StartTag#getName()
* getName()}.
*
* <p>See the {@link Tag#getName()} method for more information.
*
* @return the name of the {@linkplain #getStartTag() start tag} of this element, always in lower
* case.
*/
public String getName() {
return startTag.getName();
}