PDFBOX-5920: try to get space width by encoding first; add test by Miroslav Holubec

git-svn-id: https://svn.apache.org/repos/asf/pdfbox/trunk@1922425 13f79535-47bb-0310-9956-ffa450edef68

Author: THausherr

pdfbox/src/main/java/org/apache/pdfbox/pdmodel/font/PDFont.java

@@ -552,8 +552,8 @@ public Matrix getFontMatrix()
     }
 
     /**
-     * Determines the width of the space character.
-     * 
+     * Determines the width of the space character. This is very important for text extraction.
+     *
      * @return the width of the space character
      */
     public float getSpaceWidth()
@@ -572,7 +572,21 @@ public float getSpaceWidth()
                 }
                 else
                 {
-                    fontWidthOfSpace = getWidth(32);
+                    try
+                    {
+                        // PDFBOX-5920: try with encoding, which gets the correct code
+                        fontWidthOfSpace = getStringWidth(" ");
+                    }
+                    catch (IllegalArgumentException | UnsupportedOperationException ex)
+                    {
+                        // Happens if space is not available in the font
+                        // or if encoding isn't implemented
+                        LOG.debug(ex.getMessage(), ex);
+                    }
+                    if (fontWidthOfSpace <= 0)
+                    {
+                        fontWidthOfSpace = getWidth(32);
+                    }
                 }
 
                 // try to get it from the font itself
@@ -588,9 +602,10 @@ public float getSpaceWidth()
             }
             catch (Exception e)
             {
-                LOG.error("Can't determine the width of the space character, assuming 250", e);
+                LOG.error("Can't determine the width of the space character for font {}, assuming 250", getName(),e);
                 fontWidthOfSpace = 250f;
             }
+            LOG.debug("Space width for font {} is {}", getName(), fontWidthOfSpace);
         }
         return fontWidthOfSpace;
     }

pdfbox/src/test/java/org/apache/pdfbox/pdmodel/font/PDFontTest.java

@@ -462,4 +462,41 @@ void testPDFBox5484() throws IOException
             assertTrue(new Area(path1).equals(new Area(path2))); // assertEquals does not test equals()
         }
     }
+
+    /**
+     * Check space width.
+     *
+     * @throws IOException 
+     */
+    @Test
+    void PDFBOX5920Type0() throws IOException
+    {
+        try (InputStream is = 
+                PDFontTest.class.getResourceAsStream("/org/apache/pdfbox/resources/ttf/LiberationSans-Regular.ttf");
+                PDDocument document = new PDDocument())
+        {
+            PDFont font = PDType0Font.load(document, is, false);
+            assertEquals(20064.0f, font.getStringWidth("The quick brown fox jumps over the lazy dog."));
+            assertEquals(278.0f, font.getSpaceWidth());
+        }
+    }
+
+    /**
+     * Check space width.
+     *
+     * @throws IOException 
+     */
+    @Test
+    void PDFBOX5920TrueType() throws IOException
+    {
+        try (InputStream is = 
+                PDFontTest.class.getResourceAsStream("/org/apache/pdfbox/resources/ttf/LiberationSans-Regular.ttf");
+                PDDocument document = new PDDocument())
+        {
+            PDFont font = PDTrueTypeFont.load(document, is, WinAnsiEncoding.INSTANCE);
+            assertEquals(20064.0f, 
+                    font.getStringWidth("The quick brown fox jumps over the lazy dog."));
+            assertEquals(278.0f, font.getSpaceWidth());
+        }
+    }
 }