[GROOVY-7173] jsonslurper does not unescape Strings (sometimes) on parsing with JsonParsingUsingCharacterSource

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.3.6, 2.4.0-rc-1
• Fix Version/s: 2.3.10, 2.4.1
• Component/s: JSON
• Assignee: Pascal Schumacher
• Reporter: Tobias Conrad
• Created: 18/Nov/14 09:33
• Updated: 23/Feb/15 01:12
• Resolved: 14/Jan/15 14:50

Description

Depending on the position of the escape-character JsonParsingUsingCharacterSource.class does not recognize that the string has an escape. To raise this issue the ReaderCharacterSource.findNextChar() must be invoke recursive and the escape-character is not found in last invocation.

public void findNextCharWithRecursionHadEscape() {
		char[] control = '"\\'.toCharArray()
		
		// create ReaderCharacterSource with a very small buffer, 
		// findNextChar must be invoked recursive
		// use a string where escape-character is in last invokation
		ReaderCharacterSource cs = new ReaderCharacterSource(new StringReader('"value\\u0026"'),6)
		cs.findNextChar((int)control[0],(int)control[1]) 
		// this works
		assertTrue(cs.hadEscape()) 
		
		// use a string where escape-character is in first invokation
		cs = new ReaderCharacterSource(new StringReader('"\\u0026value"'),6)
		cs.findNextChar((int)control[0],(int)control[1]) 
		// this fail
		assertTrue("ReaderCharacterSource should have detect escape in sequence",  cs.hadEscape())
	}

ReaderCharacterSource must keep its property foundEscape if it was true on recursive findNextChar() calls.

groovy.json.internal.ReaderCharacterSource.java-line 231

 + boolean hasAlreadyFoundEscape = foundEscape;
 char results2[] = findNextChar(match, esc);

 + if(hasAlreadyFoundEscape){
 +      	foundEscape = true; //restore foundEscapeState
 + }