fix escaping and unescaping of EXI4JSON keys

EXIficient · Aug 24, 2016 · 87b4dc1 · 87b4dc1
1 parent 7bb8a8c
commit 87b4dc1
Show file tree

Hide file tree

Showing 3 changed files with 117 additions and 15 deletions.
diff --git a/src/main/java/com/siemens/ct/exi/json/EXIforJSONGenerator.java b/src/main/java/com/siemens/ct/exi/json/EXIforJSONGenerator.java
@@ -387,6 +387,8 @@ protected String escapeKey(String key) {
 			// --> Conflict with existing EXI4JSON global schema element name
 			key = String.valueOf(EXI4JSONConstants.ESCAPE_START_CHARACTER)
 					+ String.valueOf(EXI4JSONConstants.ESCAPE_END_CHARACTER) + key;
+		} else {
+			key = escapeNCNamePlus(key);
 		}
 
 		// TODO represent '_' itself
@@ -421,24 +423,61 @@ public static String escapeNCNamePlus(String name) {
 				// first character (special)
 				if (isLetter(c)) {
 					// OK
+					if (sb != null) {
+						sb.append(c);
+					}
 				} else if (c == '_') {
-					// valid NCName, but needs to be escaped for EXI4JSON
-
+					// NOT OK: valid NCName, but needs to be escaped for EXI4JSON
+					if (sb == null) {
+						sb = new StringBuilder();
+					}
+					sb.append("_95.");
 				} else {
+					// NOT OK
 					if (sb == null) {
 						sb = new StringBuilder();
 					}
+					sb.append(EXI4JSONConstants.ESCAPE_START_CHARACTER);
+					// Is this a UTF-16 surrogate pair?
+					if (Character.isHighSurrogate(c)) {
+						// use code-point and increment loop count (2 char's)
+						sb.append((int)name.codePointAt(i++));
+					} else {
+						sb.append((int)c);
+					}
+					sb.append(EXI4JSONConstants.ESCAPE_END_CHARACTER);
 				}
 			} else {
 				// rest of the characters
-
 				if (isNCNameChar(c)) {
-					// OK
 					if(c == '_') {
-						// update
+						// NOT OK: valid NCName, but needs to be escaped for EXI4JSON
+						if (sb == null) {
+							sb = new StringBuilder();
+							sb.append(name, 0, i);
+						}
+						sb.append("_95.");
+					} else {
+						// OK
+						if (sb != null) {
+							sb.append(c);
+						}
 					}
 				} else {
 					// Not OK, fix
+					if (sb == null) {
+						sb = new StringBuilder();
+						sb.append(name, 0, i);
+					}
+					sb.append(EXI4JSONConstants.ESCAPE_START_CHARACTER);
+					// Is this a UTF-16 surrogate pair?
+					if (Character.isHighSurrogate(c)) {
+						// use code-point and increment loop count (2 char's)
+						sb.append((int)name.codePointAt(i++));
+					} else {
+						sb.append((int)c);
+					}
+					sb.append(EXI4JSONConstants.ESCAPE_END_CHARACTER);
 				}
 			}
 		}

diff --git a/src/main/java/com/siemens/ct/exi/json/EXIforJSONParser.java b/src/main/java/com/siemens/ct/exi/json/EXIforJSONParser.java
@@ -320,14 +320,44 @@ protected void parseV2(InputStream isEXI4JSON, OutputStream osJSON) throws EXIEx
 	}
 
 	protected String unescapeKey(String key) {
+		StringBuilder sb = null;
+
 		// conflicting names
 		if(key.length() > 2 && key.charAt(0) == EXI4JSONConstants.ESCAPE_START_CHARACTER && key.charAt(1) == EXI4JSONConstants.ESCAPE_END_CHARACTER) {
 			key = key.substring(2);
+		} else {
+			// check whether there is an escape character
+			int i = 0;
+			while(i < key.length()) {
+				char c = key.charAt(i);
+				if(c == EXI4JSONConstants.ESCAPE_START_CHARACTER) {
+					int endIndex = key.indexOf(EXI4JSONConstants.ESCAPE_END_CHARACTER, i);
+					if(endIndex <= 0) {
+						throw new RuntimeException("Unexpected Escape Key: " + key);
+					} else {
+						int cp = Integer.parseInt(key.substring(i+1, endIndex));
+						if(sb == null)  {
+							sb = new StringBuilder();
+							sb.append(key, 0, i);
+						}
+						sb.appendCodePoint(cp);
+						i += (endIndex-i);
+					}
+				} else {
+					// ok
+					if(sb != null) {
+						sb.append(c);
+					}
+				}
+				i++;
+			}
 		}
-		// TODO represent '_' itself
-		// TODO Conflict with NCName character(s)
 
-		return key;
+		if(sb == null)  {
+			return key;
+		} else {
+			return sb.toString();
+		}
 	}
 
 //	public static void main(String[] args) throws EXIException, IOException {

diff --git a/src/test/java/com/siemens/ct/exi/json/JSONDataTestsV2.java b/src/test/java/com/siemens/ct/exi/json/JSONDataTestsV2.java
@@ -73,13 +73,14 @@ public void testEscapeKey1() throws EXIException, IOException, JSONException {
 		String key = EXI4JSONConstants.LOCALNAME_NUMBER; // "number"
 		EXIforJSONGenerator e4jGenerator = new EXIforJSONGenerator();
 		String ekey = e4jGenerator.escapeKey(key);
-		
+
 		assertFalse(key.equals(ekey));
-		assertTrue((String.valueOf(EXI4JSONConstants.ESCAPE_START_CHARACTER)+String.valueOf(EXI4JSONConstants.ESCAPE_END_CHARACTER)+EXI4JSONConstants.LOCALNAME_NUMBER).equals(ekey));
-
+		assertTrue((String.valueOf(EXI4JSONConstants.ESCAPE_START_CHARACTER)
+				+ String.valueOf(EXI4JSONConstants.ESCAPE_END_CHARACTER) + EXI4JSONConstants.LOCALNAME_NUMBER)
+						.equals(ekey));
+
 		EXIforJSONParser e4jParser = new EXIforJSONParser();
 		String ukey = e4jParser.unescapeKey(ekey);
-
 		assertTrue(ukey.equals(key));
 	}
 
@@ -91,8 +92,10 @@ public void testEscapeKey2() throws EXIException, IOException, JSONException {
 
 		assertFalse(key.equals(ekey));
 		assertTrue("a_32.number".equals(ekey));
-
-		// TODO 
+
+		EXIforJSONParser e4jParser = new EXIforJSONParser();
+		String ukey = e4jParser.unescapeKey(ekey);
+		assertTrue(ukey.equals(key));
 	}
 
 	@Test
@@ -103,8 +106,12 @@ public void testEscapeKey3() throws EXIException, IOException, JSONException {
 
 		assertFalse(key.equals(ekey));
 		assertTrue("_95.foo".equals(ekey));
+
+		EXIforJSONParser e4jParser = new EXIforJSONParser();
+		String ukey = e4jParser.unescapeKey(ekey);
+		assertTrue(ukey.equals(key));
 	}
-	
+
 	@Test
 	public void testEscapeKey4() throws EXIException, IOException, JSONException {
 		String key = "foo_.A";
@@ -113,6 +120,32 @@ public void testEscapeKey4() throws EXIException, IOException, JSONException {
 
 		assertFalse(key.equals(ekey));
 		assertTrue("foo_95..A".equals(ekey));
+
+		EXIforJSONParser e4jParser = new EXIforJSONParser();
+		String ukey = e4jParser.unescapeKey(ekey);
+		assertTrue(ukey.equals(key));
+	}
+
+	@Test
+	public void testEscapeKey5() throws EXIException, IOException, JSONException {
+		// High surrogate pair
+		byte[] data = { 0, 0x41, // A
+				(byte) 0xD8, 1, // High surrogate
+				(byte) 0xDC, 2, // Low surrogate
+				0, 0x42, // B
+		};
+
+		String key = new String(data, "UTF-16");
+
+		EXIforJSONGenerator e4jGenerator = new EXIforJSONGenerator();
+		String ekey = e4jGenerator.escapeKey(key);
+
+		assertFalse(key.equals(ekey));
+		assertTrue("A_66562.B".equals(ekey));
+
+		EXIforJSONParser e4jParser = new EXIforJSONParser();
+		String ukey = e4jParser.unescapeKey(ekey);
+		assertTrue(ukey.equals(key));
 	}
 
 }