hwpf: append any character data before paragraphs to first paragraph

(workaround for bug#48075)


git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@982238 13f79535-47bb-0310-9956-ffa450edef68

src/scratchpad/src/org/apache/poi/hwpf/usermodel/Paragraph.java

@@ -101,6 +101,14 @@ public class Paragraph extends Range implements Cloneable {
     _istd = papx.getIstd();
   }
 
+  protected Paragraph(PAPX papx, Range parent, int start)
+  {
+    super(Math.max(parent._start, start), Math.min(parent._end, papx.getEnd()), parent);
+    _props = papx.getParagraphProperties(_doc.getStyleSheet());
+    _papx = papx.getSprmBuf();
+    _istd = papx.getIstd();
+  }
+
   public short getStyleIndex()
   {
     return _istd;

src/scratchpad/src/org/apache/poi/hwpf/usermodel/Range.java

@@ -829,8 +829,12 @@ public class Range { // TODO -instantiable superclass
 		Paragraph pap = null;
 		if (props.getIlfo() > 0) {
 			pap = new ListEntry(papx, this, _doc.getListTables());
+		} else {
+            if (((index + _parStart)==0) && papx.getStart()>0) {
+                pap = new Paragraph(papx, this, 0);
             } else {
     			pap = new Paragraph(papx, this);
+            }
 		}
 
 		return pap;

src/scratchpad/testcases/org/apache/poi/hwpf/extractor/TestWordExtractor.java

@@ -298,4 +298,14 @@ public final class TestWordExtractor extends TestCase {
         assertTrue(text.contains("\u0425\u0425\u0425\u0425\u0425"));
         assertTrue(text.contains("\u0423\u0423\u0423\u0423\u0423"));
     }
+
+    public void testFirstParagraphFix() throws Exception {
+        extractor = new WordExtractor(
+                POIDataSamples.getDocumentInstance().openResourceAsStream("MBD001D0B89.doc")
+        );
+
+        String text = extractor.getText();
+
+        assertTrue(text.startsWith("\u041f\u0440\u0438\u043b\u043e\u0436\u0435\u043d\u0438\u0435"));
+    }
 }