BUG 59830 -- incorrect reading of unicode stream name

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1765468 13f79535-47bb-0310-9956-ffa450edef68

diff --git a/src/java/org/apache/poi/poifs/macros/VBAMacroReader.java b/src/java/org/apache/poi/poifs/macros/VBAMacroReader.java
index 90727c24a81f45964d2f655ac671c51a4d7268cc..d0aa4660c2393b2bdf6323ea8467c6c8a74f958d 100644 (file)
--- a/src/java/org/apache/poi/poifs/macros/VBAMacroReader.java
+++ b/src/java/org/apache/poi/poifs/macros/VBAMacroReader.java
@@ -29,6 +29,7 @@ import java.io.IOException;
 import java.io.InputStream;
 import java.io.PushbackInputStream;
 import java.nio.charset.Charset;
+import java.nio.charset.StandardCharsets;
 import java.util.HashMap;
 import java.util.Map;
 import java.util.zip.ZipEntry;
@@ -268,6 +269,7 @@ public class VBAMacroReader implements Closeable {
     private static final int MODULE_NAME = 0x0019;
     private static final int MODULE_NAME_UNICODE = 0x0047;
     private static final int MODULE_DOC_STRING = 0x001c;
+    private static final int STREAMNAME_RESERVED = 0x0032;
 
     /**
      * Reads VBA Project modules from a VBA Project directory located at
@@ -287,6 +289,7 @@ public class VBAMacroReader implements Closeable {
                     // process DIR
                     RLEDecompressingInputStream in = new RLEDecompressingInputStream(dis);
                     String streamName = null;
+                    String streamNameUnicode = null;
                     int recordId = 0;
                     try {
                         while (true) {
@@ -306,6 +309,14 @@ public class VBAMacroReader implements Closeable {
                                 break;
                             case STREAMNAME:
                                 streamName = readString(in, recordLength, modules.charset);
+                                int reserved = in.readShort();
+                                if (reserved != STREAMNAME_RESERVED) {
+                                    throw new IOException("Expected x0032 after stream name before Unicode stream name, but found: "+
+                                            Integer.toHexString(reserved));
+                                }
+                                int unicodeNameRecordLength = in.readInt();
+                                streamNameUnicode = readUnicodeString(in, unicodeNameRecordLength);
+                                //do something with this at some point
                                 break;
                             case MODULEOFFSET:
                                 readModule(in, streamName, modules);
@@ -334,4 +345,10 @@ public class VBAMacroReader implements Closeable {
             }
         }
     }
+
+    private String readUnicodeString(RLEDecompressingInputStream in, int unicodeNameRecordLength) throws IOException {
+        byte[] buffer = new byte[unicodeNameRecordLength];
+        IOUtils.readFully(in, buffer);
+        return new String(buffer, Charset.forName("UTF-16LE"));
+    }
 }

diff --git a/src/testcases/org/apache/poi/poifs/macros/TestVBAMacroReader.java b/src/testcases/org/apache/poi/poifs/macros/TestVBAMacroReader.java
index 0e634b5c581a701085ef23ed1dd7f388eb6d1ff7..f622c548676103ce1537f896176c874fe9db950c 100644 (file)
--- a/src/testcases/org/apache/poi/poifs/macros/TestVBAMacroReader.java
+++ b/src/testcases/org/apache/poi/poifs/macros/TestVBAMacroReader.java
@@ -244,22 +244,15 @@ public class TestVBAMacroReader {
         assertContains(content, testMacroNoSub);
     }
     
-    @Ignore
+
     @Test
     public void bug59830() throws IOException {
-        // This file is intentionally omitted from the test-data directory
-        // unless we can extract the vbaProject.bin from this Word 97-2003 file
-        // so that it's less likely to be opened and executed on a Windows computer.
-        // The file is attached to bug 59830.
-        // The Macro Virus only affects Windows computers, as it makes a
-        // subprocess call to powershell.exe with an encoded payload
-        // The document contains macros that execute on workbook open if macros
-        // are enabled
-        File doc = POIDataSamples.getDocumentInstance().getFile("macro_virus.doc.do_not_open");
-        VBAMacroReader reader = new VBAMacroReader(doc);
-        Map<String, String> macros = reader.readMacros();
-        assertNotNull(macros);
-        reader.close();
+        //test file is "609751.xls" in govdocs1
+        File f = POIDataSamples.getSpreadSheetInstance().getFile("59830.xls");
+        VBAMacroReader r = new VBAMacroReader(f);
+        Map<String, String> macros = r.readMacros();
+        assertNotNull(macros.get("Module20"));
+        assertContains(macros.get("Module20"), "here start of superscripting");
     }
     
     // This test is written as expected-to-fail and should be rewritten

diff --git a/test-data/spreadsheet/59830.xls b/test-data/spreadsheet/59830.xls
new file mode 100644 (file)
index 0000000..592d9ba
Binary files /dev/null and b/test-data/spreadsheet/59830.xls differ