[XMLBEANS-274] Over zealous whitespace cropping after parsing entity like &

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Version 2.1
• Fix Version/s: TBD
• Environment: All
• Assignee: Unassigned
• Reporter: Peter Rodgers
• Created: 01/Jun/06 21:49
• Updated: 20/Jul/06 21:46
• Resolved: 20/Jul/06 21:46

Description

When white space stripping is specified the parser does not detect XML entities such as & and strips the whitespace following each entity.

For example

<root>dog & cat</root>

is parsed as

<root>dog &cat</root>

The cause of the problem is the stripLeft() method in the org.apache.xmlbeans.impl.store.CharUtil

Below is a fixed version of the method that detects the ';' character after an entity which indicates that whitespace is significant and must be preserved. Note this code does not fix the case where the iteration is a for loop.

public Object stripLeft ( Object src, int off, int cch )
{
assert isValid( src, off, cch );

if (cch > 0)
{
if (src instanceof char[])
{
char[] chars = (char[]) src;

while ( cch > 0 && isWhiteSpace( chars[ off ] ) && chars[off - 1]!=';' ) //Fix for & etc

{ cch--; off++; }
}
else if (src instanceof String)
{
String s = (String) src;

while ( cch > 0 && isWhiteSpace( s.charAt( off ) ) && s.charAt(off - 1)!=';' ) //Fix for & etc
{ cch--; off++; }

}
else

{ int count = 0; for ( _charIter.init( src, off, cch ) ; _charIter.hasNext() ; count++ ) if (!isWhiteSpace( _charIter.next() )) break; _charIter.release(); off += count; }

}

if (cch == 0)

{ _offSrc = 0; _cchSrc = 0; return null; }

_offSrc = off;
_cchSrc = cch;

return src;
}