some zips can't be opened via ZipFile in JDK6, as the central directory

contains either non-latin entries or the compression type can't be handled
the workaround is to iterate over the stream and not the directory


git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1736933 13f79535-47bb-0310-9956-ffa450edef68

src/ooxml/java/org/apache/poi/openxml4j/opc/ZipPackage.java

@@ -18,6 +18,7 @@
 package org.apache.poi.openxml4j.opc;
 
 import java.io.File;
+import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
@@ -88,6 +89,7 @@ public final class ZipPackage extends Package {
      */
     ZipPackage(InputStream in, PackageAccess access) throws IOException {
         super(access);
+        @SuppressWarnings("resource")
         ThresholdInputStream zis = ZipHelper.openZipStream(in);
         this.zipArchive = new ZipInputStreamZipEntrySource(zis);
     }
@@ -101,18 +103,7 @@ public final class ZipPackage extends Package {
      *            The package access mode.
      */
     ZipPackage(String path, PackageAccess access) {
-        super(access);
-
-        final ZipFile zipFile;
-
-        try {
-            zipFile = ZipHelper.openZipFile(path);
-        } catch (IOException e) {
-            throw new InvalidOperationException(
-                    "Can't open the specified file: '" + path + "'", e);
-        }
-
-        this.zipArchive = new ZipFileZipEntrySource(zipFile);
+        this(new File(path), access);
     }
 
     /**
@@ -123,19 +114,33 @@ public final class ZipPackage extends Package {
      * @param access
      *            The package access mode.
      */
+    @SuppressWarnings("resource")
     ZipPackage(File file, PackageAccess access) {
         super(access);
 
-        final ZipFile zipFile;
-
+        ZipEntrySource ze;
         try {
-            zipFile = ZipHelper.openZipFile(file);
+            final ZipFile zipFile = ZipHelper.openZipFile(file);
+            ze = new ZipFileZipEntrySource(zipFile);
         } catch (IOException e) {
-            throw new InvalidOperationException(
-                    "Can't open the specified file: '" + file + "'", e);
+            // probably not happening with write access - not sure how to handle the default read-write access ...
+            if (access == PackageAccess.WRITE) {
+                throw new InvalidOperationException("Can't open the specified file: '" + file + "'", e);
+            }
+            logger.log(POILogger.ERROR, "Error in zip file "+file+" - falling back to stream processing (i.e. ignoring zip central directory)");
+            // some zips can't be opened via ZipFile in JDK6, as the central directory
+            // contains either non-latin entries or the compression type can't be handled
+            // the workaround is to iterate over the stream and not the directory
+            FileInputStream fis;
+            try {
+                fis = new FileInputStream(file);
+                ThresholdInputStream zis = ZipHelper.openZipStream(fis);
+                ze = new ZipInputStreamZipEntrySource(zis);
+            } catch (IOException e2) {
+                throw new InvalidOperationException("Can't open the specified file: '" + file + "'", e);
+            }
         }
-
-        this.zipArchive = new ZipFileZipEntrySource(zipFile);
+        this.zipArchive = ze;
     }
 
     /**

src/ooxml/java/org/apache/poi/openxml4j/opc/internal/ZipHelper.java

@@ -221,6 +221,7 @@ public final class ZipHelper {
      *            The stream to open.
      * @return The zip stream freshly open.
      */
+    @SuppressWarnings("resource")
     public static ThresholdInputStream openZipStream(InputStream stream) throws IOException {
         // Peek at the first few bytes to sanity check
         InputStream checkedStream = prepareToCheckHeader(stream);
@@ -228,8 +229,7 @@ public final class ZipHelper {
         
         // Open as a proper zip stream
         InputStream zis = new ZipInputStream(checkedStream);
-        ThresholdInputStream tis = ZipSecureFile.addThreshold(zis);
-        return tis;
+        return ZipSecureFile.addThreshold(zis);
     }
 
     /**
@@ -262,8 +262,6 @@ public final class ZipHelper {
      * @return The zip archive freshly open.
      */
     public static ZipFile openZipFile(String path) throws IOException {
-        File f = new File(path);
-        
-        return openZipFile(f);
+        return openZipFile(new File(path));
     }
 }

src/ooxml/testcases/org/apache/poi/extractor/TestExtractorFactory.java

@@ -32,6 +32,7 @@ import org.apache.poi.POIOLE2TextExtractor;
 import org.apache.poi.POITextExtractor;
 import org.apache.poi.POIXMLException;
 import org.apache.poi.POIXMLTextExtractor;
+import org.apache.poi.UnsupportedFileFormatException;
 import org.apache.poi.hdgf.extractor.VisioTextExtractor;
 import org.apache.poi.hpbf.extractor.PublisherTextExtractor;
 import org.apache.poi.hslf.extractor.PowerPointExtractor;
@@ -643,10 +644,7 @@ public class TestExtractorFactory {
     public void testPackage() throws Exception {
         // Excel
         POIXMLTextExtractor extractor = ExtractorFactory.createExtractor(OPCPackage.open(xlsx.toString(), PackageAccess.READ));
-        assertTrue(
-                extractor
-                instanceof XSSFExcelExtractor
-        );
+        assertTrue(extractor instanceof XSSFExcelExtractor);
         extractor.close();
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(xlsx.toString()));
         assertTrue(extractor.getText().length() > 200);
@@ -654,48 +652,33 @@ public class TestExtractorFactory {
 
         // Word
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(docx.toString()));
-        assertTrue(
-                extractor
-                instanceof XWPFWordExtractor
-        );
+        assertTrue(extractor instanceof XWPFWordExtractor);
         extractor.close();
 
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(docx.toString()));
-        assertTrue(
-                extractor.getText().length() > 120
-        );
+        assertTrue(extractor.getText().length() > 120);
         extractor.close();
 
         // PowerPoint
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(pptx.toString()));
-        assertTrue(
-                extractor
-                instanceof XSLFPowerPointExtractor
-        );
+        assertTrue(extractor instanceof XSLFPowerPointExtractor);
         extractor.close();
 
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(pptx.toString()));
-        assertTrue(
-                extractor.getText().length() > 120
-        );
+        assertTrue(extractor.getText().length() > 120);
         extractor.close();
         
         // Visio
         extractor = ExtractorFactory.createExtractor(OPCPackage.open(vsdx.toString()));
-        assertTrue(
-                extractor
-                instanceof XDGFVisioExtractor
-        );
-        assertTrue(
-                extractor.getText().length() > 20
-        );
+        assertTrue(extractor instanceof XDGFVisioExtractor);
+        assertTrue(extractor.getText().length() > 20);
         extractor.close();
 
         // Text
         try {
             ExtractorFactory.createExtractor(OPCPackage.open(txt.toString()));
             fail();
-        } catch(InvalidOperationException e) {
+        } catch(UnsupportedFileFormatException e) {
             // Good
         }
     }

src/ooxml/testcases/org/apache/poi/openxml4j/opc/TestPackage.java

@@ -41,8 +41,6 @@ import java.util.List;
 import java.util.TreeMap;
 import java.util.regex.Pattern;
 import java.util.zip.ZipEntry;
-import java.util.zip.ZipError;
-import java.util.zip.ZipException;
 import java.util.zip.ZipFile;
 import java.util.zip.ZipOutputStream;
 
@@ -50,6 +48,7 @@ import org.apache.poi.EncryptedDocumentException;
 import org.apache.poi.POIDataSamples;
 import org.apache.poi.POITestCase;
 import org.apache.poi.POIXMLException;
+import org.apache.poi.UnsupportedFileFormatException;
 import org.apache.poi.openxml4j.OpenXML4JTestDataSamples;
 import org.apache.poi.openxml4j.exceptions.InvalidFormatException;
 import org.apache.poi.openxml4j.exceptions.InvalidOperationException;
@@ -744,7 +743,7 @@ public final class TestPackage {
         try {
             OPCPackage.open(files.getFile("SampleSS.txt"));
             fail("Shouldn't be able to open Plain Text");
-        } catch (InvalidOperationException e) {
+        } catch (UnsupportedFileFormatException e) {
             // Unhelpful low-level error, sorry
         }
     }

src/ooxml/testcases/org/apache/poi/openxml4j/opc/TestZipPackage.java

@@ -24,6 +24,7 @@ import static org.junit.Assert.assertTrue;
 import static org.junit.Assert.fail;
 
 import java.io.ByteArrayOutputStream;
+import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStreamWriter;
@@ -35,6 +36,8 @@ import org.apache.poi.POIXMLException;
 import org.apache.poi.extractor.ExtractorFactory;
 import org.apache.poi.hssf.HSSFTestDataSamples;
 import org.apache.poi.openxml4j.OpenXML4JTestDataSamples;
+import org.apache.poi.sl.usermodel.SlideShow;
+import org.apache.poi.sl.usermodel.SlideShowFactory;
 import org.apache.poi.ss.usermodel.Workbook;
 import org.apache.poi.ss.usermodel.WorkbookFactory;
 import org.apache.poi.xssf.XSSFTestDataSamples;
@@ -164,4 +167,11 @@ public class TestZipPackage {
             ExtractorFactory.setThreadPrefersEventExtractors(before);
         }
     }
+    
+    @Test
+    public void unparseableCentralDirectory() throws IOException {
+        File f = OpenXML4JTestDataSamples.getSampleFile("at.pzp.www_uploads_media_PP_Scheinecker-jdk6error.pptx");
+        SlideShow<?,?> ppt = SlideShowFactory.create(f);
+        ppt.close();
+    }
 }