#64693 - POI HwmfGraphics cannot read the embedded document title

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1881322 13f79535-47bb-0310-9956-ffa450edef68

6 files changed, 69 insertions, 13 deletions

diff --git a/src/ooxml/java/org/apache/poi/xslf/usermodel/XSLFTextRun.java b/src/ooxml/java/org/apache/poi/xslf/usermodel/XSLFTextRun.java
index 1c3890a159..b57e2da602 100644
--- a/src/ooxml/java/org/apache/poi/xslf/usermodel/XSLFTextRun.java
+++ b/src/ooxml/java/org/apache/poi/xslf/usermodel/XSLFTextRun.java
@@ -738,7 +738,9 @@ public class XSLFTextRun implements TextRun {
                 // SYMBOL is missing
 
                 if (font == null || !font.isSetTypeface() || "".equals(font.getTypeface())) {
-                    font = coll.getLatin();
+                    // don't fallback to latin but bubble up in the style hierarchy (slide -> layout -> master -> theme)
+                    return null;
+//                    font = coll.getLatin();
                 }
             }
 
diff --git a/src/ooxml/java/org/apache/poi/xslf/util/EMFHandler.java b/src/ooxml/java/org/apache/poi/xslf/util/EMFHandler.java
index 9d7e620bb4..761c819612 100644
--- a/src/ooxml/java/org/apache/poi/xslf/util/EMFHandler.java
+++ b/src/ooxml/java/org/apache/poi/xslf/util/EMFHandler.java
@@ -25,6 +25,7 @@ import java.awt.geom.Rectangle2D;
 import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.charset.Charset;
 import java.util.Collections;
 
 import org.apache.poi.common.usermodel.GenericRecord;
@@ -111,4 +112,9 @@ class EMFHandler extends MFProxy {
             ? ((EmbeddedExtractor) imgr).getEmbeddings()
             : Collections.emptyList();
     }
+
+    @Override
+    void setDefaultCharset(Charset charset) {
+        imgr.setDefaultCharset(charset);
+    }
 }
diff --git a/src/ooxml/java/org/apache/poi/xslf/util/MFProxy.java b/src/ooxml/java/org/apache/poi/xslf/util/MFProxy.java
index 435fdb3e78..5a02c63695 100644
--- a/src/ooxml/java/org/apache/poi/xslf/util/MFProxy.java
+++ b/src/ooxml/java/org/apache/poi/xslf/util/MFProxy.java
@@ -25,6 +25,7 @@ import java.io.Closeable;
 import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.charset.Charset;
 import java.util.Collections;
 import java.util.Set;
 
@@ -64,4 +65,6 @@ abstract class MFProxy implements Closeable {
     abstract GenericRecord getRoot();
 
     abstract Iterable<EmbeddedPart> getEmbeddings(int slideNo);
+
+    abstract void setDefaultCharset(Charset charset);
 }
diff --git a/src/ooxml/java/org/apache/poi/xslf/util/PPTHandler.java b/src/ooxml/java/org/apache/poi/xslf/util/PPTHandler.java
index debb34f005..79d676ddc0 100644
--- a/src/ooxml/java/org/apache/poi/xslf/util/PPTHandler.java
+++ b/src/ooxml/java/org/apache/poi/xslf/util/PPTHandler.java
@@ -27,6 +27,7 @@ import java.awt.geom.Dimension2D;
 import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.charset.Charset;
 import java.util.Set;
 import java.util.Spliterator;
 import java.util.Spliterators;
@@ -49,12 +50,14 @@ import org.apache.poi.sl.usermodel.SlideShow;
 import org.apache.poi.sl.usermodel.SlideShowFactory;
 import org.apache.poi.util.IOUtils;
 import org.apache.poi.util.Internal;
+import org.apache.poi.util.LocaleUtil;
 
 /** Handler for ppt and pptx files */
 @Internal
 class PPTHandler extends MFProxy {
     private SlideShow<?,?> ppt;
     private Slide<?,?> slide;
+    private Charset defaultCharset = LocaleUtil.CHARSET_1252;
 
     @Override
     public void parse(File file) throws IOException {
@@ -67,6 +70,9 @@ class PPTHandler extends MFProxy {
                 throw e;
             }
         }
+        if (ppt == null) {
+            throw new IOException("Unknown file format or missing poi-scratchpad.jar / poi-ooxml.jar");
+        }
         slide = ppt.getSlides().get(0);
     }
 
@@ -81,6 +87,9 @@ class PPTHandler extends MFProxy {
                 throw e;
             }
         }
+        if (ppt == null) {
+            throw new IOException("Unknown file format or missing poi-scratchpad.jar / poi-ooxml.jar");
+        }
         slide = ppt.getSlides().get(0);
     }
 
@@ -162,8 +171,8 @@ class PPTHandler extends MFProxy {
         ;
     }
 
-    private static EmbeddedPart fromObjectShape(Shape s) {
-        final ObjectShape os = (ObjectShape)s;
+    private static EmbeddedPart fromObjectShape(Shape<?,?> s) {
+        final ObjectShape<?,?> os = (ObjectShape<?,?>)s;
         final ObjectData od = os.getObjectData();
         EmbeddedPart embed = new EmbeddedPart();
         embed.setName(od.getFileName());
@@ -177,4 +186,8 @@ class PPTHandler extends MFProxy {
         });
         return embed;
     }
+
+    @Override
+    void setDefaultCharset(Charset charset) {
+    }
 }
diff --git a/src/ooxml/java/org/apache/poi/xslf/util/PPTX2PNG.java b/src/ooxml/java/org/apache/poi/xslf/util/PPTX2PNG.java
index e3bbf62001..384255c57a 100644
--- a/src/ooxml/java/org/apache/poi/xslf/util/PPTX2PNG.java
+++ b/src/ooxml/java/org/apache/poi/xslf/util/PPTX2PNG.java
@@ -27,15 +27,18 @@ import java.io.File;
 import java.io.FileOutputStream;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.charset.Charset;
 import java.util.Locale;
 import java.util.Set;
 import java.util.regex.Pattern;
 
 import org.apache.poi.common.usermodel.GenericRecord;
 import org.apache.poi.poifs.filesystem.FileMagic;
+import org.apache.poi.sl.draw.Drawable;
 import org.apache.poi.sl.draw.EmbeddedExtractor.EmbeddedPart;
 import org.apache.poi.util.Dimension2DDouble;
 import org.apache.poi.util.GenericRecordJsonWriter;
+import org.apache.poi.util.LocaleUtil;
 import org.apache.poi.xslf.util.OutputFormat.BitmapFormat;
 import org.apache.poi.xslf.util.OutputFormat.SVGFormat;
 
@@ -71,7 +74,8 @@ public final class PPTX2PNG {
             "    -inputType <type> default input file type (OLE2,WMF,EMF), default is OLE2 = Powerpoint\n" +
             "                      some files (usually wmf) don't have a header, i.e. an identifiable file magic\n" +
             "    -textAsShapes     text elements are saved as shapes in SVG, necessary for variable spacing\n" +
-            "                      often found in math formulas";
+            "                      often found in math formulas\n" +
+            "    -charset          sets the default charset to be used, defaults to Windows-1252";
 
         System.out.println(msg);
         // no System.exit here, as we also run in junit tests!
@@ -99,6 +103,7 @@ public final class PPTX2PNG {
     private boolean extractEmbedded = false;
     private FileMagic defaultFileType = FileMagic.OLE2;
     private boolean textAsShapes = false;
+    private Charset charset = LocaleUtil.CHARSET_1252;
 
     private PPTX2PNG() {
     }
@@ -176,6 +181,15 @@ public final class PPTX2PNG {
                 case "-extractEmbedded":
                     extractEmbedded = true;
                     break;
+                case "-charset":
+                    if (opt != null) {
+                        charset = Charset.forName(opt);
+                        i++;
+                    } else {
+                        charset = LocaleUtil.CHARSET_1252;
+                    }
+                    break;
+
                 default:
                     file = new File(args[i]);
                     break;
@@ -264,6 +278,7 @@ public final class PPTX2PNG {
                     graphics.setRenderingHint(RenderingHints.KEY_COLOR_RENDERING, RenderingHints.VALUE_COLOR_RENDER_SPEED);
                     graphics.setRenderingHint(RenderingHints.KEY_INTERPOLATION, RenderingHints.VALUE_INTERPOLATION_BICUBIC);
                     graphics.setRenderingHint(RenderingHints.KEY_FRACTIONALMETRICS, RenderingHints.VALUE_FRACTIONALMETRICS_ON);
+                    graphics.setRenderingHint(Drawable.DEFAULT_CHARSET, getDefaultCharset());
 
                     graphics.scale(scale / lenSide, scale / lenSide);
 
@@ -315,7 +330,7 @@ public final class PPTX2PNG {
     }
 
     private void dumpRecords(MFProxy proxy) throws IOException {
-        if (dumpfile == null) {
+        if (dumpfile == null || "null".equals(dumpfile.getPath())) {
             return;
         }
         GenericRecord gr = proxy.getRoot();
@@ -387,6 +402,7 @@ public final class PPTX2PNG {
             proxy.setQuite(quiet);
             proxy.parse(file);
         }
+        proxy.setDefaultCharset(charset);
 
         return proxy;
     }
@@ -400,6 +416,10 @@ public final class PPTX2PNG {
         return INPUT_PATTERN.matcher(inname).replaceAll(outpat);
     }
 
+    private Charset getDefaultCharset() {
+        return charset;
+    }
+
     static class NoScratchpadException extends IOException {
         NoScratchpadException() {
         }
diff --git a/src/ooxml/testcases/org/apache/poi/xslf/usermodel/TestPPTX2PNG.java b/src/ooxml/testcases/org/apache/poi/xslf/usermodel/TestPPTX2PNG.java
index 96d30601e5..df23f14f28 100644
--- a/src/ooxml/testcases/org/apache/poi/xslf/usermodel/TestPPTX2PNG.java
+++ b/src/ooxml/testcases/org/apache/poi/xslf/usermodel/TestPPTX2PNG.java
@@ -19,10 +19,13 @@
 
 package org.apache.poi.xslf.usermodel;
 
+import static java.util.Arrays.asList;
 import static org.junit.Assume.assumeFalse;
 
 import java.io.File;
+import java.util.ArrayList;
 import java.util.Collection;
+import java.util.List;
 import java.util.function.Function;
 import java.util.stream.Collectors;
 import java.util.stream.Stream;
@@ -39,13 +42,14 @@ import org.junit.runners.Parameterized.Parameters;
 /**
  * Test class for testing PPTX2PNG utility which renders .ppt and .pptx slideshows
  */
+@SuppressWarnings("ConstantConditions")
 @RunWith(Parameterized.class)
 public class TestPPTX2PNG {
     private static boolean xslfOnly;
     private static final POIDataSamples samples = POIDataSamples.getSlideShowInstance();
     private static final File basedir = null;
     private static final String files =
-        "53446.ppt, alterman_security.ppt, alterman_security.pptx, KEY02.pptx, themes.pptx, " +
+        "bug64693.pptx, 53446.ppt, alterman_security.ppt, alterman_security.pptx, KEY02.pptx, themes.pptx, " +
         "backgrounds.pptx, layouts.pptx, sample.pptx, shapes.pptx, 54880_chinese.ppt, keyframes.pptx," +
         "customGeo.pptx, customGeo.ppt, wrench.emf, santa.wmf, missing-moveto.ppt";
 
@@ -62,7 +66,6 @@ public class TestPPTX2PNG {
     @Parameter
     public String pptFile;
 
-    @SuppressWarnings("ConstantConditions")
     @Parameters(name="{0}")
     public static Collection<String> data() {
         Function<String, Stream<String>> fun = (basedir == null) ? Stream::of :
@@ -75,7 +78,9 @@ public class TestPPTX2PNG {
     public void render() throws Exception {
         assumeFalse("ignore HSLF (.ppt) / HEMF (.emf) / HWMF (.wmf) files in no-scratchpad run", xslfOnly && pptFile.matches(".*\\.(ppt|emf|wmf)$"));
 
-        String[] args = {
+        // bug64693.pptx
+
+        final List<String> args = new ArrayList<>(asList(
             "-format", "null", // png,gif,jpg,svg or null for test
             "-slide", "-1", // -1 for all
             "-outdir", new File("build/tmp/").getCanonicalPath(),
@@ -84,10 +89,17 @@ public class TestPPTX2PNG {
             "-dump", "null",
             "-quiet",
             "-fixside", "long",
-            "-scale", "800",
-            // "-scale", "1.333333333",
-            (basedir == null ? samples.getFile(pptFile) : new File(basedir, pptFile)).getAbsolutePath()
-        };
-        PPTX2PNG.main(args);
+            "-scale", "800"
+        ));
+
+        if ("bug64693.pptx".equals(pptFile)) {
+            args.addAll(asList(
+                "-charset", "GBK"
+            ));
+        }
+
+        args.add((basedir == null ? samples.getFile(pptFile) : new File(basedir, pptFile)).getAbsolutePath());
+
+        PPTX2PNG.main(args.toArray(new String[0]));
     }
 }