/**
* Populates a {@link Tag} object using data from the supplied {@link CharArray}.
*
* <p>The supplied tag parameter is reset and reused - this avoids excess object creation which
* hwlps performance.
*
* @return the same tag instance that was passed in, except it will be populated with a new
* <tt>name</tt> value (and the corresponding <tt>nameEndIdx</tt> value). However if the tag
* contained nathing but whitespace, this method will return <tt>null</tt>.
*/
private Tag parseTag(Tag tag, CharArray buf) {
int len = buf.length();
int idx = 0;
int begin;
// Skip over any leading whitespace in the tag
while (idx < len && Character.isWhitespace(buf.charAt(idx))) idx++;
if (idx == len) return null;
// Find out where the non-whitespace characters end. This will give us the tag name.
begin = idx;
while (idx < len && !Character.isWhitespace(buf.charAt(idx))) idx++;
// Mark the tag name as a substring within the buffer. This allows us to perform
// a substring comparison against it at a later date
buf.setSubstr(begin, buf.charAt(idx - 1) == '/' ? idx - 1 : idx);
// Remember where the name finishes so we can pull out the properties later if need be
tag.nameEndIdx = idx;
return tag;
}
/**
* This is called when we need to extract the properties for the tag from the tag's HTML. We only
* call this when necessary since it has quite a lot of overhead.
*
* @param tag the tag that is currently being processed. This should be the tag that was returned
* as a result of a call to {@link #parseTag(FastPageParser.Tag, CharArray)} (ie, it has the
* <tt>name</tt> and <tt>nameEndIdx</tt> fields set correctly for the tag in question. The
* <tt>properties</tt> field can be in an undefined state - it will get replaced regardless).
* @param buffer a <tt>CharArray</tt> containing the entire tag that is being parsed.
* @return the same tag instance that was passed in, only it will now be populated with any
* properties that were specified in the tag's HTML.
*/
private static Tag parseProperties(Tag tag, CharArray buffer) {
int len = buffer.length();
int idx = tag.nameEndIdx;
// Start with an empty hashmap. A new HashMap is lazy-created if we happen to find any
// properties
tag.properties = Collections.EMPTY_MAP;
int begin;
while (idx < len) {
// Skip forward to the next non-whitespace character
while (idx < len && Character.isWhitespace(buffer.charAt(idx))) idx++;
if (idx == len) continue;
begin = idx;
if (buffer.charAt(idx) == '"') {
idx++;
while (idx < len && buffer.charAt(idx) != '"') idx++;
if (idx == len) continue;
idx++;
} else if (buffer.charAt(idx) == '\'') {
idx++;
while (idx < len && buffer.charAt(idx) != '\'') idx++;
if (idx == len) continue;
idx++;
} else {
while (idx < len
&& !Character.isWhitespace(buffer.charAt(idx))
&& buffer.charAt(idx) != '=') idx++;
}
// Mark the substring. This is the attribute name
buffer.setSubstr(begin, idx);
if (idx < len && Character.isWhitespace(buffer.charAt(idx))) {
while (idx < len && Character.isWhitespace(buffer.charAt(idx))) idx++;
}
if (idx == len || buffer.charAt(idx) != '=') continue;
idx++;
if (idx == len) continue;
while (idx < len && (buffer.charAt(idx) == '\n' || buffer.charAt(idx) == '\r')) idx++;
if (buffer.charAt(idx) == ' ') {
while (idx < len && Character.isWhitespace(buffer.charAt(idx))) idx++;
if (idx == len || (buffer.charAt(idx) != '"' && buffer.charAt(idx) != '"')) continue;
}
begin = idx;
int end;
if (buffer.charAt(idx) == '"') {
idx++;
begin = idx;
while (idx < len && buffer.charAt(idx) != '"') idx++;
if (idx == len) continue;
end = idx;
idx++;
} else if (buffer.charAt(idx) == '\'') {
idx++;
begin = idx;
while (idx < len && buffer.charAt(idx) != '\'') idx++;
if (idx == len) continue;
end = idx;
idx++;
} else {
while (idx < len && !Character.isWhitespace(buffer.charAt(idx))) idx++;
end = idx;
}
// Extract the name and value as String objects and add them to the property map
String name = buffer.getLowerSubstr();
String value = buffer.substring(begin, end);
tag.addProperty(name, value);
}
return tag;
}
