#56260 Partial fix for a slide with a TextHeaderAtom but no other atoms related to it, which is followed by another TextHeaderAtom straight away

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@1577537 13f79535-47bb-0310-9956-ffa450edef68

3 files changed, 89 insertions, 7 deletions

diff --git a/src/scratchpad/src/org/apache/poi/hslf/model/Sheet.java b/src/scratchpad/src/org/apache/poi/hslf/model/Sheet.java
index 46670ff364..875b2dfafa 100644
--- a/src/scratchpad/src/org/apache/poi/hslf/model/Sheet.java
+++ b/src/scratchpad/src/org/apache/poi/hslf/model/Sheet.java
@@ -17,16 +17,34 @@
 
 package org.apache.poi.hslf.model;
 
-import org.apache.poi.ddf.*;
-import org.apache.poi.hslf.record.*;
-import org.apache.poi.hslf.usermodel.SlideShow;
-import org.apache.poi.util.POILogFactory;
-import org.apache.poi.util.POILogger;
-
+import java.awt.Graphics2D;
 import java.util.ArrayList;
 import java.util.Iterator;
 import java.util.List;
-import java.awt.*;
+
+import org.apache.poi.ddf.EscherContainerRecord;
+import org.apache.poi.ddf.EscherDgRecord;
+import org.apache.poi.ddf.EscherDggRecord;
+import org.apache.poi.ddf.EscherRecord;
+import org.apache.poi.hslf.record.CString;
+import org.apache.poi.hslf.record.ColorSchemeAtom;
+import org.apache.poi.hslf.record.EscherTextboxWrapper;
+import org.apache.poi.hslf.record.OEPlaceholderAtom;
+import org.apache.poi.hslf.record.PPDrawing;
+import org.apache.poi.hslf.record.Record;
+import org.apache.poi.hslf.record.RecordContainer;
+import org.apache.poi.hslf.record.RecordTypes;
+import org.apache.poi.hslf.record.RoundTripHFPlaceholder12;
+import org.apache.poi.hslf.record.SheetContainer;
+import org.apache.poi.hslf.record.StyleTextProp9Atom;
+import org.apache.poi.hslf.record.StyleTextPropAtom;
+import org.apache.poi.hslf.record.TextBytesAtom;
+import org.apache.poi.hslf.record.TextCharsAtom;
+import org.apache.poi.hslf.record.TextHeaderAtom;
+import org.apache.poi.hslf.record.TextRulerAtom;
+import org.apache.poi.hslf.usermodel.SlideShow;
+import org.apache.poi.util.POILogFactory;
+import org.apache.poi.util.POILogger;
 
 /**
  * This class defines the common format of "Sheets" in a powerpoint
@@ -182,6 +200,7 @@ public abstract class Sheet {
                 
                 // Is there a StyleTextPropAtom after the Text Atom?
                 // TODO Do we need to check for text ones two away as well?
+                // TODO Refactor this to happen later in this for loop
                 if (i < (records.length - 2)) {
                     next = records[i+2];
                     if (next instanceof StyleTextPropAtom) {
@@ -191,6 +210,16 @@ public abstract class Sheet {
 
                 // See what follows the TextHeaderAtom
                 next = records[i+1];
+                
+                // Is it one we ignore and check the one after that?
+                if (i < records.length - 2) {
+                    // TODO MasterTextPropAtom
+                    if (next instanceof TextRulerAtom) {
+                        next = records[i+2];
+                    }
+                }
+                
+                // Is it one we need to record?
                 if (next instanceof TextCharsAtom) {
                     TextCharsAtom tca = (TextCharsAtom)next;
                     trun = new TextRun(tha, tca, stpa);
@@ -199,6 +228,11 @@ public abstract class Sheet {
                     trun = new TextRun(tha, tba, stpa);
                 } else if (next instanceof StyleTextPropAtom) {
                     stpa = (StyleTextPropAtom)next;
+                } else if (next instanceof TextHeaderAtom) {
+                    // Seems to be a mostly, but not completely deleted block of
+                    //  text. Only the header remains, which isn't useful alone 
+                    // Skip on to the next TextHeaderAtom
+                    continue;
                 } else if (next.getRecordType() == (long)RecordTypes.TextSpecInfoAtom.typeID ||
                            next.getRecordType() == (long)RecordTypes.BaseTextPropAtom.typeID) { 
                     // Safe to ignore these ones
diff --git a/src/scratchpad/src/org/apache/poi/hslf/model/TextRun.java b/src/scratchpad/src/org/apache/poi/hslf/model/TextRun.java
index fdc30f6ea0..c5c876a827 100644
--- a/src/scratchpad/src/org/apache/poi/hslf/model/TextRun.java
+++ b/src/scratchpad/src/org/apache/poi/hslf/model/TextRun.java
@@ -22,8 +22,10 @@ import java.util.LinkedList;
 import java.util.List;
 
 import org.apache.poi.hslf.model.textproperties.TextPropCollection;
+import org.apache.poi.hslf.record.PPDrawing;
 import org.apache.poi.hslf.record.Record;
 import org.apache.poi.hslf.record.RecordContainer;
+import org.apache.poi.hslf.record.SlideListWithText;
 import org.apache.poi.hslf.record.StyleTextProp9Atom;
 import org.apache.poi.hslf.record.StyleTextPropAtom;
 import org.apache.poi.hslf.record.TextBytesAtom;
@@ -656,6 +658,14 @@ public final class TextRun
     protected void setIndex(int id){
         slwtIndex = id;
     }
+    
+    /**
+     * Is this Text Run one from a {@link PPDrawing}, or is it
+     *  one from the {@link SlideListWithText}?
+     */
+    public boolean isDrawingBased() {
+        return (slwtIndex == -1);
+    }
 
     /**
      * Returns the array of all hyperlinks in this text run
diff --git a/src/scratchpad/testcases/org/apache/poi/hslf/usermodel/TestBugs.java b/src/scratchpad/testcases/org/apache/poi/hslf/usermodel/TestBugs.java
index 50b21a007d..efc196d3f7 100644
--- a/src/scratchpad/testcases/org/apache/poi/hslf/usermodel/TestBugs.java
+++ b/src/scratchpad/testcases/org/apache/poi/hslf/usermodel/TestBugs.java
@@ -51,6 +51,11 @@ import org.apache.poi.hslf.model.TextBox;
 import org.apache.poi.hslf.model.TextRun;
 import org.apache.poi.hslf.model.TextShape;
 import org.apache.poi.hslf.model.TitleMaster;
+import org.apache.poi.hslf.record.Document;
+import org.apache.poi.hslf.record.Record;
+import org.apache.poi.hslf.record.SlideListWithText;
+import org.apache.poi.hslf.record.SlideListWithText.SlideAtomsSet;
+import org.apache.poi.hslf.record.TextHeaderAtom;
 import org.junit.Test;
 
 /**
@@ -502,4 +507,37 @@ public final class TestBugs {
         assertTrue("No Exceptions while reading headers", true);
     }
     
+    @Test
+    public void bug56260() throws Exception {
+        File file = _slTests.getFile("56260.ppt");
+        
+        HSLFSlideShow ss = new HSLFSlideShow(file.getAbsolutePath());
+        SlideShow _show = new SlideShow(ss);
+        Slide[] _slides = _show.getSlides();
+        assertEquals(13, _slides.length);
+        
+        // Check the number of TextHeaderAtoms on Slide 1
+        Document dr = _show.getDocumentRecord();
+        SlideListWithText slidesSLWT = dr.getSlideSlideListWithText();
+        SlideAtomsSet s1 = slidesSLWT.getSlideAtomsSets()[0];
+
+        int tha = 0;
+        for (Record r : s1.getSlideRecords()) {
+            if (r instanceof TextHeaderAtom) tha++;
+        }
+        assertEquals(2, tha);
+        
+        // Check to see that we have a pair next to each other
+        assertEquals(TextHeaderAtom.class, s1.getSlideRecords()[0].getClass());
+        assertEquals(TextHeaderAtom.class, s1.getSlideRecords()[1].getClass());
+        
+        
+        // Check the number of text runs based on the slide (not textbox)
+        // Will have skipped the empty one
+        int str = 0;
+        for (TextRun tr : _slides[0].getTextRuns()) {
+            if (! tr.isDrawingBased()) str++;
+        }
+        assertEquals(1, str);
+    }
 }