Fix multiple zlib decoding bugs

Author: msimacek

graalpython/com.oracle.graal.python.test/src/tests/test_zlib.py

@@ -14,11 +14,12 @@ def _disable_native_zlib(self, flag):
             return None
     __graalpython__ = GP()
 
+import binascii
 import os
+import random
+import sys
 import unittest
 import zlib
-import binascii
-import sys
 
 pintNumber = 98765432109876543210
 longNumber = 9876543210
@@ -271,3 +272,35 @@ def test_GR65704():
     __graalpython__._disable_native_zlib(False)
 
     assert decompressed == contents
+
+def test_large_chunk():
+    contents = random.randbytes(5000)
+    wbits = 31
+
+    __graalpython__._disable_native_zlib(True)
+
+    compressed = zlib.compress(contents, wbits=wbits)
+    decompressor = zlib.decompressobj(wbits=wbits)
+
+    decompressed = decompressor.decompress(compressed)
+
+    __graalpython__._disable_native_zlib(False)
+
+    assert decompressed == contents
+
+def test_various_chunks():
+    contents = random.randbytes(5000)
+    wbits = 31
+
+    __graalpython__._disable_native_zlib(True)
+
+    compressed = zlib.compress(contents, wbits=wbits)
+    decompressor = zlib.decompressobj(wbits=wbits)
+
+    decompressed = decompressor.decompress(compressed[:10])
+    decompressed += decompressor.decompress(compressed[10:200])
+    decompressed += decompressor.decompress(compressed[200:])
+
+    __graalpython__._disable_native_zlib(False)
+
+    assert decompressed == contents

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/modules/zlib/JavaDecompress.java

@@ -47,6 +47,7 @@
 
 import java.io.ByteArrayInputStream;
 import java.io.ByteArrayOutputStream;
+import java.io.EOFException;
 import java.io.IOException;
 import java.io.InputStream;
 import java.util.zip.DataFormatException;
@@ -112,7 +113,7 @@ public Inflater getInflater() {
             return inf;
         }
 
-        public void setInput() throws IOException {
+        public void fillInput() throws IOException {
             fill();
         }
     }
@@ -162,7 +163,7 @@ private static boolean isGZIPStreamReady(DecompressStream stream, byte[] data, i
                 // GZIPInputStream will read the header during initialization
                 stream.stream = new GZIPDecompressStream(stream.in);
                 stream.inflater = stream.stream.getInflater();
-                stream.stream.setInput();
+                stream.stream.fillInput();
                 return true;
             }
         } catch (ZipException ze) {
@@ -178,7 +179,7 @@ private static boolean isGZIPStreamFinishing(DecompressStream stream, byte[] dat
             stream.in.append(data, 0, length);
             try {
                 if (stream.in.length() >= HEADER_TRAILER_SIZE) {
-                    stream.stream.setInput();
+                    stream.stream.fillInput();
                     // this should trigger reading trailer
                     stream.stream.read();
                     stream.stream = null;
@@ -246,19 +247,30 @@ private byte[] createByteArray(byte[] bytes, int length, int maxLength, int bufS
         int maxLen = maxLength <= 0 ? Integer.MAX_VALUE : maxLength;
         byte[] result = new byte[Math.min(maxLen, bufSize)];
 
-        int bytesWritten = result.length;
         ByteArrayOutputStream baos = new ByteArrayOutputStream();
         boolean zdictIsSet = false;
-        while (baos.size() < maxLen && bytesWritten == result.length) {
+        while (baos.size() < maxLen && !stream.inflater.finished()) {
+            if (stream.inflater.needsInput()) {
+                if (stream.stream == null) {
+                    break;
+                }
+                try {
+                    stream.stream.fillInput();
+                } catch (EOFException e) {
+                    break;
+                } catch (IOException e) {
+                    throw CompilerDirectives.shouldNotReachHere(e);
+                }
+            }
+            int bytesWritten;
             try {
                 int len = Math.min(maxLen - baos.size(), result.length);
                 bytesWritten = stream.inflater.inflate(result, 0, len);
                 if (bytesWritten == 0 && !zdictIsSet && stream.inflater.needsDictionary()) {
                     if (getZdict().length > 0) {
                         setDictionary();
                         zdictIsSet = true;
-                        // we inflate again with a dictionary
-                        bytesWritten = stream.inflater.inflate(result, 0, len);
+                        continue;
                     } else {
                         throw PRaiseNode.raiseStatic(nodeForRaise, ZLibError, WHILE_SETTING_ZDICT);
                     }
@@ -320,7 +332,7 @@ protected static byte[] decompress(byte[] bytes, int length, int wbits, int bufs
     private void saveUnconsumedInput(byte[] data, int length,
                     byte[] unusedDataBytes, int unconsumedTailLen, Node inliningTarget) {
         int unusedLen = getRemaining();
-        byte[] tail = PythonUtils.arrayCopyOfRange(data, length - unusedLen, length);
+        byte[] tail = PythonUtils.arrayCopyOfRange(data, Math.max(0, length - unusedLen), length);
         PythonLanguage language = PythonLanguage.get(inliningTarget);
         if (isEof()) {
             if (unconsumedTailLen > 0) {