Author: fanningpj
Date: Sun Jul 5 23:16:17 2026
New Revision: 1935919
Log:
replace invalid xml chars in element content in Saver. Thanks to aizu-m. This
closes #74
Modified:
xmlbeans/trunk/src/main/java/org/apache/xmlbeans/impl/store/Saver.java
xmlbeans/trunk/src/test/java/misc/checkin/SaveOptimizeForSpeedTest.java
Modified: xmlbeans/trunk/src/main/java/org/apache/xmlbeans/impl/store/Saver.java
==============================================================================
--- xmlbeans/trunk/src/main/java/org/apache/xmlbeans/impl/store/Saver.java
Sun Jul 5 20:10:03 2026 (r1935918)
+++ xmlbeans/trunk/src/main/java/org/apache/xmlbeans/impl/store/Saver.java
Sun Jul 5 23:16:17 2026 (r1935919)
@@ -2118,6 +2118,10 @@ abstract class Saver {
int index = 0;
for (int i = 0; i < bufLimit; i++) {
char c = _buf[i];
+ if (isBadChar(c)) {
+ _buf[i] = '?';
+ c = '?';
+ }
switch (c) {
case '<':
emit(_buf, index, i - index);
Modified:
xmlbeans/trunk/src/test/java/misc/checkin/SaveOptimizeForSpeedTest.java
==============================================================================
--- xmlbeans/trunk/src/test/java/misc/checkin/SaveOptimizeForSpeedTest.java
Sun Jul 5 20:10:03 2026 (r1935918)
+++ xmlbeans/trunk/src/test/java/misc/checkin/SaveOptimizeForSpeedTest.java
Sun Jul 5 23:16:17 2026 (r1935919)
@@ -99,6 +99,23 @@ public class SaveOptimizeForSpeedTest {
}
@Test
+ void testBadCharInText() throws Exception {
+ // an invalid XML 1.0 character (a C0 control) in element content. The
+ // default saver replaces it with '?'; the speed path emitted it raw,
+ // producing output that is not well-formed.
+ XmlObject o = XmlObject.Factory.parse("<root/>");
+ try (XmlCursor cur = o.newCursor()) {
+ cur.toFirstChild();
+ cur.toFirstContentToken();
+ cur.insertChars("a\u0001b");
+ }
+ String out = saveForSpeed(o);
+ assertFalse(out.indexOf('\u0001') >= 0);
+ // before the fix the raw control char makes this a fatal parse error
+ XmlObject.Factory.parse(out);
+ }
+
+ @Test
void testProcInstTerminatorAcrossChunkBoundary() throws Exception {
// a '?>' that straddles the 512-char chunk boundary: '?' is the last
char
// of the first chunk, '>' the first char of the second. The per-chunk
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
