Use StreamConstraintsException exception in canonicalizer (#948)

FasterXML · Mar 14, 2023 · 25313f1 · 25313f1
1 parent 7fe79e9
commit 25313f1
Show file tree

Hide file tree

Showing 8 changed files with 107 additions and 62 deletions.
diff --git a/src/main/java/com/fasterxml/jackson/core/json/UTF8DataInputJsonParser.java b/src/main/java/com/fasterxml/jackson/core/json/UTF8DataInputJsonParser.java
@@ -5,6 +5,7 @@
 
 import com.fasterxml.jackson.core.*;
 import com.fasterxml.jackson.core.base.ParserBase;
+import com.fasterxml.jackson.core.exc.StreamConstraintsException;
 import com.fasterxml.jackson.core.io.CharTypes;
 import com.fasterxml.jackson.core.io.IOContext;
 import com.fasterxml.jackson.core.sym.ByteQuadsCanonicalizer;
@@ -38,8 +39,6 @@
 public class UTF8DataInputJsonParser
     extends ParserBase
 {
-    final static byte BYTE_LF = (byte) '\n';
-
     @SuppressWarnings("deprecation")
     private final static int FEAT_MASK_TRAILING_COMMA = Feature.ALLOW_TRAILING_COMMA.getMask();
     @SuppressWarnings("deprecation")
@@ -1768,7 +1767,8 @@ protected String _parseAposName() throws IOException
     /**********************************************************
      */
 
-    private final String findName(int q1, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q1 = pad(q1, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -1781,7 +1781,8 @@ private final String findName(int q1, int lastQuadBytes) throws JsonParseExcepti
         return addName(_quadBuffer, 1, lastQuadBytes);
     }
 
-    private final String findName(int q1, int q2, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int q2, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q2 = pad(q2, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -1795,7 +1796,8 @@ private final String findName(int q1, int q2, int lastQuadBytes) throws JsonPars
         return addName(_quadBuffer, 2, lastQuadBytes);
     }
 
-    private final String findName(int q1, int q2, int q3, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int q2, int q3, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q3 = pad(q3, lastQuadBytes);
         String name = _symbols.findName(q1, q2, q3);
@@ -1809,7 +1811,8 @@ private final String findName(int q1, int q2, int q3, int lastQuadBytes) throws
         return addName(quads, 3, lastQuadBytes);
     }
 
-    private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadBytes) throws JsonParseException
+    private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         if (qlen >= quads.length) {
             _quadBuffer = quads = _growArrayBy(quads, quads.length);
@@ -1828,7 +1831,8 @@ private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadB
      * multi-byte chars (if any), and then construct Name instance
      * and add it to the symbol table.
      */
-    private final String addName(int[] quads, int qlen, int lastQuadBytes) throws JsonParseException
+    private final String addName(int[] quads, int qlen, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         /* Ok: must decode UTF-8 chars. No other validation is
          * needed, since unescaping has been done earlier as necessary

diff --git a/src/main/java/com/fasterxml/jackson/core/json/UTF8StreamJsonParser.java b/src/main/java/com/fasterxml/jackson/core/json/UTF8StreamJsonParser.java
@@ -4,6 +4,7 @@
 
 import com.fasterxml.jackson.core.*;
 import com.fasterxml.jackson.core.base.ParserBase;
+import com.fasterxml.jackson.core.exc.StreamConstraintsException;
 import com.fasterxml.jackson.core.io.CharTypes;
 import com.fasterxml.jackson.core.io.IOContext;
 import com.fasterxml.jackson.core.sym.ByteQuadsCanonicalizer;
@@ -2309,7 +2310,8 @@ protected String _parseAposName() throws IOException
     /**********************************************************
      */
 
-    private final String findName(int q1, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q1 = _padLastQuad(q1, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -2322,7 +2324,8 @@ private final String findName(int q1, int lastQuadBytes) throws JsonParseExcepti
         return addName(_quadBuffer, 1, lastQuadBytes);
     }
 
-    private final String findName(int q1, int q2, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int q2, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q2 = _padLastQuad(q2, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -2336,7 +2339,8 @@ private final String findName(int q1, int q2, int lastQuadBytes) throws JsonPars
         return addName(_quadBuffer, 2, lastQuadBytes);
     }
 
-    private final String findName(int q1, int q2, int q3, int lastQuadBytes) throws JsonParseException
+    private final String findName(int q1, int q2, int q3, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q3 = _padLastQuad(q3, lastQuadBytes);
         String name = _symbols.findName(q1, q2, q3);
@@ -2350,7 +2354,8 @@ private final String findName(int q1, int q2, int q3, int lastQuadBytes) throws
         return addName(quads, 3, lastQuadBytes);
     }
 
-    private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadBytes) throws JsonParseException
+    private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         if (qlen >= quads.length) {
             _quadBuffer = quads = growArrayBy(quads, quads.length);
@@ -2368,7 +2373,8 @@ private final String findName(int[] quads, int qlen, int lastQuad, int lastQuadB
      * multi-byte chars (if any), and then construct Name instance
      * and add it to the symbol table.
      */
-    private final String addName(int[] quads, int qlen, int lastQuadBytes) throws JsonParseException
+    private final String addName(int[] quads, int qlen, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         /* Ok: must decode UTF-8 chars. No other validation is
          * needed, since unescaping has been done earlier as necessary

diff --git a/src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingJsonParserBase.java b/src/main/java/com/fasterxml/jackson/core/json/async/NonBlockingJsonParserBase.java
@@ -4,6 +4,7 @@
 
 import com.fasterxml.jackson.core.*;
 import com.fasterxml.jackson.core.base.ParserBase;
+import com.fasterxml.jackson.core.exc.StreamConstraintsException;
 import com.fasterxml.jackson.core.io.IOContext;
 import com.fasterxml.jackson.core.json.JsonReadContext;
 import com.fasterxml.jackson.core.sym.ByteQuadsCanonicalizer;
@@ -639,7 +640,8 @@ protected final JsonToken _closeObjectScope() throws IOException
     /**********************************************************
      */
 
-    protected final String _findName(int q1, int lastQuadBytes) throws JsonParseException
+    protected final String _findName(int q1, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q1 = _padLastQuad(q1, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -652,7 +654,8 @@ protected final String _findName(int q1, int lastQuadBytes) throws JsonParseExce
         return _addName(_quadBuffer, 1, lastQuadBytes);
     }
 
-    protected final String _findName(int q1, int q2, int lastQuadBytes) throws JsonParseException
+    protected final String _findName(int q1, int q2, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q2 = _padLastQuad(q2, lastQuadBytes);
         // Usually we'll find it from the canonical symbol table already
@@ -666,7 +669,8 @@ protected final String _findName(int q1, int q2, int lastQuadBytes) throws JsonP
         return _addName(_quadBuffer, 2, lastQuadBytes);
     }
 
-    protected final String _findName(int q1, int q2, int q3, int lastQuadBytes) throws JsonParseException
+    protected final String _findName(int q1, int q2, int q3, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         q3 = _padLastQuad(q3, lastQuadBytes);
         String name = _symbols.findName(q1, q2, q3);
@@ -684,10 +688,11 @@ protected final String _findName(int q1, int q2, int q3, int lastQuadBytes) thro
     // table miss. It needs to demultiplex individual bytes, decode
     // multi-byte chars (if any), and then construct Name instance
     // and add it to the symbol table.
-    protected final String _addName(int[] quads, int qlen, int lastQuadBytes) throws JsonParseException
+    protected final String _addName(int[] quads, int qlen, int lastQuadBytes)
+            throws JsonParseException, StreamConstraintsException
     {
         /* Ok: must decode UTF-8 chars. No other validation is
-         * needed, since unescaping has been done earlier as necessary
+         * needed, since unescaping has been done earlier, as necessary
          * (as well as error reporting for unescaped control chars)
          */
         // 4 bytes per quad, except last one maybe less

diff --git a/src/main/java/com/fasterxml/jackson/core/sym/ByteQuadsCanonicalizer.java b/src/main/java/com/fasterxml/jackson/core/sym/ByteQuadsCanonicalizer.java
@@ -4,6 +4,7 @@
 import java.util.concurrent.atomic.AtomicReference;
 
 import com.fasterxml.jackson.core.JsonFactory;
+import com.fasterxml.jackson.core.exc.StreamConstraintsException;
 import com.fasterxml.jackson.core.util.InternCache;
 
 /**
@@ -863,7 +864,13 @@ 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);
@@ -876,7 +883,14 @@ 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);
@@ -896,7 +910,15 @@ 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);
@@ -911,7 +933,14 @@ 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) {
@@ -921,7 +950,7 @@ public String addName(String name, int[] q, int qlen)
 
         switch (qlen) {
         case 1:
-        {
+            {
                 offset = _findOffsetForAdd(calcHash(q[0]));
                 _hashArea[offset] = q[0];
                 _hashArea[offset+3] = 1;
@@ -961,16 +990,16 @@ public String addName(String name, int[] q, int qlen)
         return name;
     }
 
-    private void _verifySharing()
+    private void _verifySharing() throws StreamConstraintsException
     {
         if (_hashShared) {
             // 12-Mar-2021, tatu: prevent modifying of "placeholder" and
             //   parent tables
             if (_parent == null) {
                 if (_count == 0) { // root
-                    throw new IllegalStateException("Cannot add names to Root symbol table");
+                    throw new StreamConstraintsException("Cannot add names to Root symbol table");
                 }
-                throw new IllegalStateException("Cannot add names to Placeholder symbol table");
+                throw new StreamConstraintsException("Cannot add names to Placeholder symbol table");
             }
 
             _hashArea = Arrays.copyOf(_hashArea, _hashArea.length);
@@ -982,7 +1011,7 @@ 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);
@@ -1037,7 +1066,7 @@ 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();
@@ -1165,10 +1194,16 @@ 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,
@@ -1202,7 +1237,7 @@ 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;
@@ -1279,7 +1314,7 @@ 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 StreamConstraintsException("Failed rehash(): old count="+oldCount+", copyCount="+copyCount);
         }
     }
 
@@ -1315,13 +1350,13 @@ 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;
         }
-        throw new IllegalStateException("Spill-over slots in symbol table with "+_count
+        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."
                 +" You can disable the check via `JsonFactory.Feature.FAIL_ON_SYMBOL_HASH_OVERFLOW`");