epam · printsev · Dec 29, 2020 · Dec 28, 2020 · Dec 29, 2020 · Dec 29, 2020
diff --git a/src/main/java/com/epam/parso/impl/SasFileParser.java b/src/main/java/com/epam/parso/impl/SasFileParser.java
@@ -77,6 +77,13 @@ public final class SasFileParser {
      */
     private static final int MAX_PAGE_LENGTH = 10000000;
 
+    /**
+     * Byte buffer used for skip operations.
+     * Actually the data containing in this buffer is ignored,
+     * because it only used for dummy reads.
+     */
+    private static final byte[] SKIP_BYTE_BUFFER = new byte[4096];
+
     static {
         Map<Long, SubheaderIndexes> tmpMap = new HashMap<>();
         tmpMap.put((long) 0xF7F7F7F7, SubheaderIndexes.ROW_SIZE_SUBHEADER_INDEX);
@@ -359,16 +366,40 @@ private void processSasFileHeader(String builderEncoding) throws IOException {
         }
 
         if (sasFileStream != null) {
+            skipBytes(sasFileProperties.getHeaderLength() - currentFilePosition);
+            currentFilePosition = 0;
+        }
+    }
+
+    /**
+     * Skip specified number of bytes of data from the input stream,
+     * or fail if there are not enough left.
+     *
+     * @param numberOfBytesToSkip the number of bytes to skip
+     * @throws IOException if the number of bytes skipped was incorrect
+     */
+    private void skipBytes(long numberOfBytesToSkip) throws IOException {
+
+        long remainBytes = numberOfBytesToSkip;
+        long readBytes;
+        while (remainBytes > 0) {
             try {
-                int bytesLeft = sasFileProperties.getHeaderLength() - currentFilePosition;
-                long actuallySkipped = 0;
-                while (actuallySkipped < bytesLeft) {
-                    actuallySkipped += sasFileStream.skip(bytesLeft - actuallySkipped);
+                readBytes = sasFileStream.read(SKIP_BYTE_BUFFER, 0,
+                        (int) Math.min(remainBytes, SKIP_BYTE_BUFFER.length));
+                if (readBytes < 0) { // EOF
+                    break;
                 }
-                currentFilePosition = 0;
             } catch (IOException e) {
                 throw new IOException(EMPTY_INPUT_STREAM);
             }
+            remainBytes -= readBytes;
+        }
+
+        long actuallySkipped = numberOfBytesToSkip - remainBytes;
+
+        if (actuallySkipped != numberOfBytesToSkip) {
+            throw new IOException("Expected to skip " + numberOfBytesToSkip
+                    + " to the end of the header, but skipped " + actuallySkipped + " instead.");
         }
     }
 
@@ -893,14 +924,7 @@ private List<byte[]> getBytesFromFile(Long[] offset, Integer[] length) throws IO
         if (cachedPage == null) {
             for (int i = 0; i < offset.length; i++) {
                 byte[] temp = new byte[length[i]];
-                long actuallySkipped = 0;
-                while (actuallySkipped < offset[i] - currentFilePosition) {
-                    try {
-                        actuallySkipped += sasFileStream.skip(offset[i] - currentFilePosition - actuallySkipped);
-                    } catch (IOException e) {
-                        throw new IOException(EMPTY_INPUT_STREAM);
-                    }
-                }
+                skipBytes(offset[i] - currentFilePosition);
                 try {
                     sasFileStream.readFully(temp, 0, length[i]);
                 } catch (EOFException e) {

diff --git a/src/test/java/com/epam/parso/BugsTest.java b/src/test/java/com/epam/parso/BugsTest.java
@@ -20,7 +20,11 @@
 import com.epam.parso.impl.SasFileReaderImpl;
 import org.junit.Test;
 
+import java.io.ByteArrayInputStream;
+import java.io.FileInputStream;
 import java.io.InputStream;
+import java.nio.file.Files;
+import java.nio.file.Paths;
 import java.util.List;
 
 import static org.assertj.core.api.Assertions.assertThat;
@@ -139,4 +143,77 @@ public void testCompressionMethodsIssue55() throws Exception {
             assertThat(sasFileReader.getSasFileProperties().getCompressionMethod()).isNull();
         }
     }
+
+    @Test(timeout = 1000)
+    public void testInfinityLoopOnEmptyStreamIssue56() {
+        ByteArrayInputStream emptyStream = new ByteArrayInputStream(new byte[]{});
+        SasFileReaderImpl sasFileReader = new SasFileReaderImpl(emptyStream);
+        assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+    }
+
+    @Test(timeout = 1000)
+    public void testInfinityLoopOnIncompleteStreamIssue56() {
+        ByteArrayInputStream emptyStream = new ByteArrayInputStream(new byte[]{1, 2, 3, 4, 5, 6, 7, 8, 9});
+        SasFileReaderImpl sasFileReader = new SasFileReaderImpl(emptyStream);
+        assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+    }
+
+    /**
+     * This has invalid page and header lengths:
+     * - page length = 31612 (7C 7B 00 00)
+     * - header length = 32126 (7E 7D 00 00)
+     * See this row in a hex editor:
+     * 000000C0:  00 00 00 00 7E 7D 00 00 │ 7C 7B 00 00 B8 00 00 00
+     * <p>
+     * File size is 17408 - it is smaller than specified lengths, so both lengths refer
+     * to the  out of file bounds offsets.
+     */
+    private static final String INVALID_FILE_NAME = "invalid_lengths.sas7bdat";
+
+    @Test(timeout = 1000)
+    public void testBrokenFileOnFileChannelIssue58() throws Exception {
+        // These skips were performed performed for the FileChannel by the parser before the fix:
+        // [32, 2, 1, 32, 21, 16, 8, 17120, 0, 0, 0, ... infinity loop of zeros]
+        try (InputStream is = Files.newInputStream(
+                Paths.get(this.getClass().getResource("/bugs/" + INVALID_FILE_NAME).toURI()))) {
+            SasFileReader sasFileReader = new SasFileReaderImpl(is);
+            assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+        }
+    }
+
+    @Test(timeout = 1000)
+    public void testBrokenFileOnBufferedFileInputStreamIssue58() throws Exception {
+        // These skips were performed performed for the buffered stream by the parser before the fix:
+        // [32, 2, 1, 32, 21, 16, 8, 7936, 23902]
+        try (InputStream is = this.getClass().getResourceAsStream("/bugs/" + INVALID_FILE_NAME)) {
+            SasFileReader sasFileReader = new SasFileReaderImpl(is);
+            assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+        }
+    }
+
+    @Test(timeout = 1000)
+    public void testBrokenFileOnUnbufferedFileInputStreamIssue58() throws Exception {
+        // These skips were performed for the buffered stream by the parser before the fix:
+        // [32, 2, 1, 32, 21, 16, 8, 31838]
+        try (InputStream is = new FileInputStream("target/test-classes/bugs/" + INVALID_FILE_NAME)) {
+            SasFileReader sasFileReader = new SasFileReaderImpl(is);
+            assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+        }
+    }
+
+    @Test(timeout = 1000)
+    public void testInfinityLoopBufferedIssue58() throws Exception {
+        try (InputStream is = this.getClass().getResourceAsStream("/bugs/sas_infinite_loop.sas7bdat")) {
+            SasFileReader sasFileReader = new SasFileReaderImpl(is);
+            assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+        }
+    }
+
+    @Test(timeout = 1000)
+    public void testInfinityLoopUnbufferedIssue58() throws Exception {
+        try (InputStream is = new FileInputStream("target/test-classes/bugs/sas_infinite_loop.sas7bdat")) {
+            SasFileReader sasFileReader = new SasFileReaderImpl(is);
+            assertThat(sasFileReader.getSasFileProperties().getRowCount()).isEqualTo(0);
+        }
+    }
 }
diff --git a/src/test/resources/bugs/invalid_lengths.sas7bdat b/src/test/resources/bugs/invalid_lengths.sas7bdat
diff --git a/src/test/resources/bugs/sas_infinite_loop.sas7bdat b/src/test/resources/bugs/sas_infinite_loop.sas7bdat