Improve check to cover all multi-byte encoded characters

FasterXML · Jan 9, 2024 · a1f62b1 · a1f62b1
1 parent 7711289
commit a1f62b1
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 11 deletions.
diff --git a/avro/src/main/java/com/fasterxml/jackson/dataformat/avro/deser/JacksonAvroParserImpl.java b/avro/src/main/java/com/fasterxml/jackson/dataformat/avro/deser/JacksonAvroParserImpl.java
@@ -614,9 +614,8 @@ private final String _finishShortText(int len) throws IOException
         _inputPtr += len;
         final byte[] inputBuf = _inputBuffer;
 
-        // Let's actually do a tight loop for ASCII first:
         final int end = inPtr + len;
-
+        // Let's actually do a tight loop for ASCII first:
         int i;
         while ((i = inputBuf[inPtr]) >= 0) {
             outBuf[outPtr++] = (char) i;
@@ -626,16 +625,20 @@ private final String _finishShortText(int len) throws IOException
         }
 
         final int[] codes = sUtf8UnitLengths;
-
         do {
             i = inputBuf[inPtr++] & 0xFF;
-            if (inPtr >= end) {
-                throw _constructError("Malformed UTF-8 character at end of short (non-chunked) text segment");
-            }
-            switch (codes[i]) {
-            case 0:
+            final int code = codes[i];
+            if (code == 0) { // still optimized for ASCII
                 outBuf[outPtr++] = (char) i;
-                break;
+                continue;
+            }
+            if ((inPtr + code) > end) {
+                if (code < 4) {
+                    throw _constructError(String.format(
+                            "Malformed %d-byte UTF-8 character at the end of Unicode text block", code));
+                }
+            }
+            switch (code) {
             case 1:
                 i = ((i & 0x1F) << 6) | (inputBuf[inPtr++] & 0x3F);
                 break;
@@ -655,7 +658,7 @@ private final String _finishShortText(int len) throws IOException
                 i = 0xDC00 | (i & 0x3FF);
                 break;
             default: // invalid
-                _reportError("Invalid byte "+Integer.toHexString(i)+" in Unicode text block");
+                _reportError(String.format("Invalid byte 0x2X in Unicode text block", i));
             }
             outBuf[outPtr++] = (char) i;
         } while (inPtr < end);

diff --git a/...src/test/java/com/fasterxml/jackson/dataformat/avro/fuzz/AvroFuzz449_65618_IOOBETest.java b/...src/test/java/com/fasterxml/jackson/dataformat/avro/fuzz/AvroFuzz449_65618_IOOBETest.java
@@ -39,7 +39,7 @@ public void testFuzz65618IOOBE() throws Exception {
             p.nextToken();
             fail("Should not pass (invalid content)");
         } catch (StreamReadException e) {
-            assertTrue(e.getMessage().contains("Malformed UTF-8 character"));
+            verifyException(e, "Malformed 2-byte UTF-8 character at the end of");
         }
     }
 }