Merge branch '2.15'

release-notes/VERSION-2.x

@@ -37,6 +37,8 @@ JSON library.
   allow for deferred decoding in some cases
 #921: Add `JsonFactory.Feature` to disable charset detection
  (contributed by @yawkat) 
+#948: Use `StreamConstraintsException` in name canonicalizers
+ (contributed by @pjfanning)
 - Build uses package type "jar" but still produces valid OSGi bundle
  (changed needed to keep class timestamps with Reproducible Build)
 

src/main/java/tools/jackson/core/sym/ByteQuadsCanonicalizer.java

@@ -3,7 +3,7 @@
 import java.util.Arrays;
 import java.util.concurrent.atomic.AtomicReference;
 
-import tools.jackson.core.exc.StreamReadException;
+import tools.jackson.core.exc.StreamConstraintsException;
 import tools.jackson.core.json.JsonFactory;
 import tools.jackson.core.util.InternCache;
 
@@ -858,7 +858,13 @@ private boolean _verifyLongName2(int[] q, int qlen, int spillOffset)
     /**********************************************************
      */
 
-    public String addName(String name, int q1) {
+    /**
+     * @param name
+     * @param q1
+     * @return name (possibly interned)
+     * @throws StreamConstraintsException if the constraint exceptions
+     */
+    public String addName(String name, int q1) throws StreamConstraintsException {
         _verifySharing();
         if (_intern) {
             name = InternCache.instance.intern(name);
@@ -871,7 +877,14 @@ public String addName(String name, int q1) {
         return name;
     }
 
-    public String addName(String name, int q1, int q2) {
+    /**
+     * @param name
+     * @param q1
+     * @param q2
+     * @return name (possibly interned)
+     * @throws StreamConstraintsException if the constraint exceptions
+     */
+    public String addName(String name, int q1, int q2) throws StreamConstraintsException {
         _verifySharing();
         if (_intern) {
             name = InternCache.instance.intern(name);
@@ -891,7 +904,15 @@ public String addName(String name, int q1, int q2) {
         return name;
     }
 
-    public String addName(String name, int q1, int q2, int q3) {
+    /**
+     * @param name
+     * @param q1
+     * @param q2
+     * @param q3
+     * @return name (possibly interned)
+     * @throws StreamConstraintsException if the constraint exceptions
+     */
+    public String addName(String name, int q1, int q2, int q3) throws StreamConstraintsException {
         _verifySharing();
         if (_intern) {
             name = InternCache.instance.intern(name);
@@ -906,7 +927,14 @@ public String addName(String name, int q1, int q2, int q3) {
         return name;
     }
 
-    public String addName(String name, int[] q, int qlen)
+    /**
+     * @param name
+     * @param q
+     * @param qlen
+     * @return name (possibly interned)
+     * @throws StreamConstraintsException if the constraint exceptions
+     */
+    public String addName(String name, int[] q, int qlen) throws StreamConstraintsException
     {
         _verifySharing();
         if (_intern) {
@@ -915,7 +943,8 @@ public String addName(String name, int[] q, int qlen)
         int offset;
 
         switch (qlen) {
-        case 1: {
+        case 1:
+            {
                 offset = _findOffsetForAdd(calcHash(q[0]));
                 _hashArea[offset] = q[0];
                 _hashArea[offset+3] = 1;
@@ -961,9 +990,9 @@ private void _verifySharing()
             //   parent tables
             if (_parent == null) {
                 if (_count == 0) { // root
-                    throw new IllegalStateException("Cannot add names to Root symbol table");
+                    throw new IllegalStateException("Internal error: Cannot add names to Root symbol table");
                 }
-                throw new IllegalStateException("Cannot add names to Placeholder symbol table");
+                throw new IllegalStateException("Internal error: Cannot add names to Placeholder symbol table");
             }
 
             _hashArea = Arrays.copyOf(_hashArea, _hashArea.length);
@@ -975,7 +1004,7 @@ private void _verifySharing()
     /**
      * Method called to find the location within hash table to add a new symbol in.
      */
-    private int _findOffsetForAdd(int hash)
+    private int _findOffsetForAdd(int hash) throws StreamConstraintsException
     {
         // first, check the primary: if slot found, no need for resize
         int offset = _calcOffset(hash);
@@ -1030,7 +1059,7 @@ private int _findOffsetForAdd(int hash)
     }
 
     // @since 2.10
-    private int _resizeAndFindOffsetForAdd(int hash)
+    private int _resizeAndFindOffsetForAdd(int hash) throws StreamConstraintsException
     {
         // First things first: we need to resize+rehash (or, if too big, nuke contents)
         rehash();
@@ -1158,10 +1187,16 @@ public int calcHash(int q1, int q2, int q3)
         return hash;
     }
 
+    /**
+     * @param q int array
+     * @param qlen length
+     * @return hash
+     * @throws IllegalArgumentException if <code>qlen</code> is less than 4
+     */
     public int calcHash(int[] q, int qlen)
     {
         if (qlen < 4) {
-            throw new IllegalArgumentException();
+            throw new IllegalArgumentException("qlen is too short, needs to be at least 4");
         }
         /* And then change handling again for "multi-quad" case; mostly
          * to make calculation of collisions less fun. For example,
@@ -1195,7 +1230,7 @@ public int calcHash(int[] q, int qlen)
     /**********************************************************
      */
 
-    private void rehash()
+    private void rehash() throws StreamConstraintsException
     {
         // Note: since we'll make copies, no need to unshare, can just mark as such:
         _hashShared = false;
@@ -1272,7 +1307,7 @@ private void rehash()
         // Sanity checks: since corruption difficult to detect, assert explicitly
         // with production code
         if (copyCount != oldCount) {
-            throw new IllegalStateException("Failed rehash(): old count="+oldCount+", copyCount="+copyCount);
+            throw new IllegalStateException("Internal error: Failed rehash(), old count="+oldCount+", copyCount="+copyCount);
         }
     }
 
@@ -1308,15 +1343,13 @@ private final int _spilloverStart() {
         return (offset << 3) - offset;
     }
 
-    protected void _reportTooManyCollisions()
+    protected void _reportTooManyCollisions() throws StreamConstraintsException
     {
         // First: do not fuzz about small symbol tables; may get balanced by doubling up
         if (_hashSize <= 1024) { // would have spill-over area of 128 entries
             return;
         }
-        // 20-Mar-2021, tatu: [core#686]: should use Jackson-specific exception
-        //    (to use new "processing limit" exception when available)
-        throw new StreamReadException(null,
+        throw new StreamConstraintsException(
 "Spill-over slots in symbol table with "+_count
 +" entries, hash area of "+_hashSize+" slots is now full (all "
 +(_hashSize >> 3)+" slots -- suspect a DoS attack based on hash collisions."

src/main/java/tools/jackson/core/sym/CharsToNameCanonicalizer.java

@@ -4,7 +4,7 @@
 import java.util.BitSet;
 import java.util.concurrent.atomic.AtomicReference;
 
-import tools.jackson.core.exc.StreamReadException;
+import tools.jackson.core.exc.StreamConstraintsException;
 import tools.jackson.core.json.JsonFactory;
 import tools.jackson.core.util.InternCache;
 
@@ -517,7 +517,7 @@ private String _addSymbol(char[] buffer, int start, int len, int h, int index)
      * Method called when an overflow bucket has hit the maximum expected length:
      * this may be a case of DoS attack. Deal with it based on settings by either
      * clearing up bucket (to avoid indefinite expansion) or throwing exception.
-     * Currently the first overflow for any single bucket DOES NOT throw an exception,
+     * Currently, the first overflow for any single bucket DOES NOT throw an exception,
      * only second time (per symbol table instance)
      */
     private void _handleSpillOverflow(int bucketIndex, Bucket newBucket, int mainIndex)
@@ -698,10 +698,8 @@ private void rehash() {
         }
     }
 
-    // 20-Mar-2021, tatu: [core#686]: should use Jackson-specific exception
-    //    (to use new "processing limit" exception when available)
     protected void _reportTooManyCollisions(int maxLen) {
-        throw new StreamReadException(null,
+        throw new StreamConstraintsException(
 "Longest collision chain in symbol table (of size "+_size
 +") now exceeds maximum, "+maxLen+" -- suspect a DoS attack based on hash collisions."
 +" You can disable the check via `TokenStreamFactory.Feature.FAIL_ON_SYMBOL_HASH_OVERFLOW`");
@@ -727,8 +725,9 @@ protected void verifyInternalConsistency() {
             }
         }
         if (count != _size) {
-            throw new IllegalStateException(String.format("Internal error: expected internal size %d vs calculated count %d",
-                    _size, count));
+            throw new IllegalStateException(
+                    String.format("Internal error: expected internal size %d vs calculated count %d",
+                            _size, count));
         }
     }
 

src/test/java/tools/jackson/core/sym/TestByteBasedSymbols.java

@@ -2,6 +2,7 @@
 
 import java.io.*;
 import java.lang.reflect.Field;
+import java.nio.charset.StandardCharsets;
 import java.util.Random;
 
 import tools.jackson.core.*;
@@ -116,13 +117,13 @@ public void testIssue207() throws Exception
         }
         stringBuilder.append("\n}");
 
-        JsonParser p = jsonF.createParser(ObjectReadContext.empty(), stringBuilder.toString().getBytes("UTF-8"));
+        JsonParser p = jsonF.createParser(ObjectReadContext.empty(), stringBuilder.toString().getBytes(StandardCharsets.UTF_8));
         while (p.nextToken() != null) { }
         p.close();
     }
 
     // [core#548]
-    public void testQuadsIssue548()
+    public void testQuadsIssue548() throws Exception
     {
         Random r = new Random(42);
         ByteQuadsCanonicalizer root = ByteQuadsCanonicalizer.createRoot();
@@ -170,9 +171,9 @@ public void testQuadsIssue548()
     /**********************************************************
      */
 
-    protected JsonParser createParser(JsonFactory jf, String input) throws IOException
+    protected JsonParser createParser(JsonFactory jf, String input) throws Exception
     {
-        byte[] data = input.getBytes("UTF-8");
+        byte[] data = input.getBytes(StandardCharsets.UTF_8);
         InputStream is = new ByteArrayInputStream(data);
         return jf.createParser(ObjectReadContext.empty(), is);
     }

src/test/java/tools/jackson/core/sym/TestHashCollisionChars.java

@@ -153,18 +153,6 @@ void printString(String s) {
           System.out.println(s.length() + " " + s.hashCode() + " \"" + s + "\"");
       }
 
-      String asIntArray(String s) {
-        StringBuilder result = new StringBuilder().append("[");
-        for (int c = 0; c < s.length(); c++) {
-          if (c > 0) {
-            result.append(", ");
-          }
-          result.append((int) s.charAt(c));
-        }
-        result.append("]");
-        return result.toString();
-      }
-
     }
 
     public static void main(String[] args) {

src/test/java/tools/jackson/core/sym/TestSymbolTables.java

@@ -3,6 +3,7 @@
 import java.io.IOException;
 import java.lang.reflect.Field;
 import java.nio.charset.Charset;
+import java.nio.charset.StandardCharsets;
 
 import tools.jackson.core.*;
 import tools.jackson.core.json.JsonFactory;
@@ -15,7 +16,7 @@ public class TestSymbolTables extends tools.jackson.core.BaseTest
 {
     // Test for verifying stability of hashCode, wrt collisions, using
     // synthetic field name generation and character-based input
-    public void testSyntheticWithChars()
+    public void testSyntheticWithChars() throws IOException
     {
         // pass seed, to keep results consistent:
         CharsToNameCanonicalizer symbols = CharsToNameCanonicalizer.createRoot(1).makeChild(-1);
@@ -109,7 +110,7 @@ public void testThousandsOfSymbolsWithNew() throws IOException
         final int SEED = 33333;
 
         ByteQuadsCanonicalizer symbolsBRoot = ByteQuadsCanonicalizer.createRoot(SEED);
-        final Charset utf8 = Charset.forName("UTF-8");
+        final Charset utf8 = StandardCharsets.UTF_8;
         int exp = 0;
         ByteQuadsCanonicalizer symbolsB = null;