[GROOVY-4380] Round-tripping XML which contains a newline character reference in an attribute's value removes the character reference

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.6.4, 1.7.4
• Fix Version/s: 1.7.5, 1.8-beta-2
• Component/s: XML Processing
• Flags: Patch
• Assignee: Paul King
• Reporter: Ben Jansen
• Created: 25/Aug/10 19:43
• Updated: 05/Apr/15 14:43
• Resolved: 28/Aug/10 04:32

Description

Given an XML file which contains a newline character reference ( ) in the value of an attribute, the following groovy code incorrectly outputs the actual newline character instead of the original character reference. This is a significant problem because the XML standard dictates that newlines found in the value of an attribute must be treated as a space. Thus, if any XML parser reads the output of this groovy snippet, it is as if the newline is gone.

newline.xml:
<option name="LINE_SEPARATOR" value=" "/>

groovy:

def fileWriter = new FileWriter('./newline-roundtrip.xml')
def input = new XmlParser().parse('./newline.xml')
PrintWriter printWriter = new PrintWriter(fileWriter)
new XmlNodePrinter(printWriter).print(input)
printWriter.flush()

I've attached a junit test which demonstrates the problem.