Fixes #588, fixes #564, changes #568, relates #550 - Only filter out known problematic characters

danfickle · danfickle · commit 9ffd0e4ad4a1 · 2020-10-30T21:37:38.000+11:00
Can not be too aggressive as some fonts contain private area code points, etc and expect them to be output.
diff --git a/openhtmltopdf-core/src/main/java/com/openhtmltopdf/extend/TextRenderer.java b/openhtmltopdf-core/src/main/java/com/openhtmltopdf/extend/TextRenderer.java
@@ -35,15 +35,15 @@ public interface TextRenderer {
      *
      * @param input The string can be null
      * @return The cleaned string or <code>null</code> if the input is null
-     * @see com.openhtmltopdf.util.OpenUtil#isCodePointPrintable(int)
+     * @see com.openhtmltopdf.util.OpenUtil#isSafeFontCodePointToPrint(int)
      */
-    static String getEffectivePrintableString(String input) {
+    public static String getEffectivePrintableString(String input) {
         if (input == null || input.isEmpty() || areAllCharactersPrintable(input)) {
             return input;
         }
 
         StringBuilder effective = new StringBuilder(input.length());
-        input.codePoints().filter(OpenUtil::isCodePointPrintable).forEach(effective::appendCodePoint);
+        input.codePoints().filter(OpenUtil::isSafeFontCodePointToPrint).forEach(effective::appendCodePoint);
 
         return effective.toString();
     }
diff --git a/openhtmltopdf-core/src/main/java/com/openhtmltopdf/util/OpenUtil.java b/openhtmltopdf-core/src/main/java/com/openhtmltopdf/util/OpenUtil.java
@@ -10,6 +10,11 @@ private OpenUtil() {}
 	 * Checks if a code point is printable. If false, it can be safely discarded at the 
 	 * rendering stage, else it should be replaced with the replacement character,
 	 * if a suitable glyph can not be found.
+	 * 
+	 * NOTE: This should only be called after a character has been shown to be
+	 * NOT present in the font. It can not be called beforehand because some fonts
+	 * contain private area characters and so on. Issue#588.
+	 * 
 	 * @param codePoint
 	 * @return whether codePoint is printable
 	 */
@@ -26,14 +31,31 @@ public static boolean isCodePointPrintable(int codePoint) {
 				 category == Character.SURROGATE);
 	}
 
+    /**
+     * Whether the code point should be passed through to the font
+     * for rendering. It effectively filters out characters that
+     * have been shown to be problematic in some (broken) fonts such
+     * as visible soft-hyphens.
+     */
+    public static boolean isSafeFontCodePointToPrint(int codePoint) {
+        switch (codePoint) {
+        case 0xAD:        // Soft hyphen, PR#550, FALLTHRU
+        case 0xFFFC:      // Object replacement character, Issue#564.
+            return false;
+
+        default:
+            return true;
+        }
+    }
+
 	/**
 	 * Returns <code>true</code>, when all characters of the given string are printable.
 	 * @param str a non-null string to test
 	 * @return whether all characters are printable
 	 */
 	public static boolean areAllCharactersPrintable(String str) {
 		Objects.requireNonNull(str, "str");
-		return str.codePoints().allMatch(OpenUtil::isCodePointPrintable);
+		return str.codePoints().allMatch(OpenUtil::isSafeFontCodePointToPrint);
 	}
 
 	public static Integer parseIntegerOrNull(String possibleInteger) {
diff --git a/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxFastOutputDevice.java b/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxFastOutputDevice.java
@@ -43,7 +43,6 @@
 import com.openhtmltopdf.simple.extend.ReplacedElementScaleHelper;
 import com.openhtmltopdf.util.ArrayUtil;
 import com.openhtmltopdf.util.LogMessageId;
-import com.openhtmltopdf.util.OpenUtil;
 import com.openhtmltopdf.util.XRLog;
 import de.rototor.pdfbox.graphics2d.PdfBoxGraphics2D;
 import de.rototor.pdfbox.graphics2d.PdfBoxGraphics2DFontTextDrawer;
@@ -74,10 +73,6 @@
 import java.util.Map.Entry;
 import java.util.logging.Level;
 import java.util.regex.Pattern;
-import java.util.stream.IntStream;
-import java.util.stream.StreamSupport;
-
-import static com.openhtmltopdf.util.OpenUtil.areAllCharactersPrintable;
 
 public class PdfBoxFastOutputDevice extends AbstractOutputDevice implements OutputDevice, PdfBoxOutputDevice {
     //
diff --git a/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxTextRenderer.java b/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxTextRenderer.java
@@ -23,7 +23,6 @@
 import java.util.ArrayList;
 import java.util.List;
 import java.util.logging.Level;
-import java.util.stream.IntStream;
 
 import com.openhtmltopdf.bidi.BidiReorderer;
 import com.openhtmltopdf.extend.FontContext;
@@ -35,11 +34,10 @@
 import com.openhtmltopdf.render.FSFontMetrics;
 import com.openhtmltopdf.render.JustificationInfo;
 import com.openhtmltopdf.util.LogMessageId;
+import com.openhtmltopdf.util.OpenUtil;
 import com.openhtmltopdf.util.ThreadCtx;
 import com.openhtmltopdf.util.XRLog;
 
-import static com.openhtmltopdf.util.OpenUtil.isCodePointPrintable;
-
 public class PdfBoxTextRenderer implements TextRenderer {
     private static float TEXT_MEASURING_DELTA = 0.01f;
 
@@ -194,7 +192,9 @@ public static List<FontRun> divideIntoFontRuns(FSFont font, String str, BidiReor
             i += Character.charCount(unicode);
             String ch = String.valueOf(Character.toChars(unicode));
 
-            if (!isCodePointPrintable(unicode)) {
+            if (!OpenUtil.isSafeFontCodePointToPrint(unicode)) {
+                // Filter out characters that should never be visible (such
+                // as soft-hyphen) but are in some fonts.
                 continue;
             }
 
@@ -264,8 +264,14 @@ else if (des != current.des) {
                     }
                 }
             }
-            
+
             if (!gotChar) {
+                if (!OpenUtil.isCodePointPrintable(unicode)) {
+                    // Filter out control, etc characters when they
+                    // are not present in any font.
+                    continue;
+                }
+
                 // We still don't have the character after all that. So use replacement character.
                 if (current.des == null) {
                     // First character of run.
@@ -366,6 +372,8 @@ public int getSmoothingLevel() {
         return 0;
     }
 
+    @Deprecated
+    @Override
     public void setSmoothingLevel(int level) {
     }
 

Original file line number	Diff line number	Diff line change
`@@ -35,15 +35,15 @@ public interface TextRenderer {`
`35`	`35`	`*`
`36`	`36`	`* @param input The string can be null`
`37`	`37`	`* @return The cleaned string or <code>null</code> if the input is null`
`38`		`- * @see com.openhtmltopdf.util.OpenUtil#isCodePointPrintable(int)`
	`38`	`+ * @see com.openhtmltopdf.util.OpenUtil#isSafeFontCodePointToPrint(int)`
`39`	`39`	`*/`
`40`		`- static String getEffectivePrintableString(String input) {`
	`40`	`+ public static String getEffectivePrintableString(String input) {`
`41`	`41`	`if (input == null \|\| input.isEmpty() \|\| areAllCharactersPrintable(input)) {`
`42`	`42`	`return input;`
`43`	`43`	`}`
`44`	`44`
`45`	`45`	`StringBuilder effective = new StringBuilder(input.length());`
`46`		`- input.codePoints().filter(OpenUtil::isCodePointPrintable).forEach(effective::appendCodePoint);`
	`46`	`+ input.codePoints().filter(OpenUtil::isSafeFontCodePointToPrint).forEach(effective::appendCodePoint);`
`47`	`47`
`48`	`48`	`return effective.toString();`
`49`	`49`	`}`