Extended behavior of preserved linefeeds to codepoints representing mandatory breaks in Unicode UAX#14:

- NEXT LINE (U+0085)
- LINE SEPARATOR (U+2028)
- PARAGRAPH SEPARATOR (U+2029)


git-svn-id: https://svn.apache.org/repos/asf/xmlgraphics/fop/trunk@664347 13f79535-47bb-0310-9956-ffa450edef68

2 files changed, 46 insertions, 22 deletions

diff --git a/src/java/org/apache/fop/layoutmgr/inline/TextLayoutManager.java b/src/java/org/apache/fop/layoutmgr/inline/TextLayoutManager.java
index acdb75f06..58ab9505f 100644
--- a/src/java/org/apache/fop/layoutmgr/inline/TextLayoutManager.java
+++ b/src/java/org/apache/fop/layoutmgr/inline/TextLayoutManager.java
@@ -140,7 +140,7 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
     private final MinOptMax[] letterAdjustArray; //size = textArray.length + 1
 
     private static final char NEWLINE = '\n';
-
+    
     /** Font used for the space between words. */
     private Font spaceFont = null;
     /** Start index of next TextArea */
@@ -523,7 +523,7 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
             || CharUtilities.isNonBreakableSpace(ch)
             || CharUtilities.isFixedWidthSpace(ch);
     }
-
+    
     /** {@inheritDoc} */
     public LinkedList getNextKnuthElements(final LayoutContext context, final int alignment) {
         this.lineStartBAP = context.getLineStartBorderAndPaddingWidth();
@@ -561,8 +561,9 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
                     TextLayoutManager.LOG.error("Unexpected breakAction: " + breakAction);
             }
             if (inWord) {
-                if (breakOpportunity || TextLayoutManager.isSpace(ch)
-                        || ch == TextLayoutManager.NEWLINE) {
+                if (breakOpportunity 
+                        || TextLayoutManager.isSpace(ch)
+                        || CharUtilities.isExplicitBreak(ch)) {
                     // this.textArray[lastIndex] == CharUtilities.SOFT_HYPHEN
                     prevAi = this.processWord(alignment, sequence, prevAi, ch,
                             breakOpportunity, true);
@@ -601,12 +602,13 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
                         (short) 0, (short) 0, ipd, false, true,
                         breakOpportunity, font);
                 this.thisStart = (short) (this.nextStart + 1);
-            } else if (ch == TextLayoutManager.NEWLINE) {
-                // linefeed; this can happen when linefeed-treatment="preserve"
+            } else if (CharUtilities.isExplicitBreak(ch)) {
+                //mandatory break-character: only advance index
                 this.thisStart = (short) (this.nextStart + 1);
             }
+            
             inWord = !TextLayoutManager.isSpace(ch)
-                    && ch != TextLayoutManager.NEWLINE;
+                    && !CharUtilities.isExplicitBreak(ch);
             inWhitespace = ch == CharUtilities.SPACE
                     && this.foText.getWhitespaceTreatment() != Constants.EN_PRESERVE;
             this.nextStart++;
@@ -620,7 +622,7 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
         } else if (ai != null) {
             ai = this.processLeftoverAi(alignment, sequence, ai, ch,
                     ch == CharUtilities.ZERO_WIDTH_SPACE);
-        } else if (ch == TextLayoutManager.NEWLINE) {
+        } else if (CharUtilities.isExplicitBreak(ch)) {
             sequence = this.processLinebreak(returnList, sequence);
         }
 
@@ -628,6 +630,7 @@ public class TextLayoutManager extends LeafNodeLayoutManager {
             //Remove an empty sequence because of a trailing newline
             returnList.removeLast();
         }
+        
         this.setFinished(true);
         if (returnList.isEmpty()) {
             return null;
diff --git a/src/java/org/apache/fop/util/CharUtilities.java b/src/java/org/apache/fop/util/CharUtilities.java
index 4910a371c..6baa5c0fd 100644
--- a/src/java/org/apache/fop/util/CharUtilities.java
+++ b/src/java/org/apache/fop/util/CharUtilities.java
@@ -54,10 +54,16 @@ public class CharUtilities {
     public static final int XMLWHITESPACE = 4;
 
 
+    /** linefeed character */
+    public static final char LINEFEED_CHAR = '\n';
+    /** carriage return */
+    public static final char CARRIAGE_RETURN = '\r';
     /** normal space */
     public static final char SPACE = '\u0020';
     /** non-breaking space */
     public static final char NBSPACE = '\u00A0';
+    /** next line control character */
+    public static final char NEXT_LINE = '\u0085';
     /** zero-width space */
     public static final char ZERO_WIDTH_SPACE = '\u200B';
     /** word joiner */
@@ -68,6 +74,10 @@ public class CharUtilities {
     public static final char ZERO_WIDTH_NOBREAK_SPACE = '\uFEFF';
     /** soft hyphen */
     public static final char SOFT_HYPHEN = '\u00AD';
+    /** line-separator */
+    public static final char LINE_SEPARATOR = '\u2028';
+    /** paragraph-separator */
+    public static final char PARAGRAPH_SEPARATOR = '\u2029';
     /** missing ideograph */
     public static final char MISSING_IDEOGRAPH = '\u25A1';
     /** Unicode value indicating the the character is "not a character". */
@@ -174,8 +184,7 @@ public class CharUtilities {
      * @return True if the character represents any kind of space
      */
     public static boolean isAnySpace(char c) {
-        boolean ret = (isBreakableSpace(c) || isNonBreakableSpace(c));
-        return ret;
+        return (isBreakableSpace(c) || isNonBreakableSpace(c));
     }
     
     /**
@@ -188,19 +197,31 @@ public class CharUtilities {
         //Generated from: Other_Alphabetic + Lu + Ll + Lt + Lm + Lo + Nl
         int generalCategory = Character.getType(ch);
         switch (generalCategory) {
-        case Character.UPPERCASE_LETTER: //Lu
-        case Character.LOWERCASE_LETTER: //Ll
-        case Character.TITLECASE_LETTER: //Lt
-        case Character.MODIFIER_LETTER: //Lm
-        case Character.OTHER_LETTER: //Lo
-        case Character.LETTER_NUMBER: //Nl
-            return true;
-        default:
-            //TODO if (ch in Other_Alphabetic) return true; (Probably need ICU4J for that)
-            //Other_Alphabetic contains mostly more exotic characters
-            return false;
+            case Character.UPPERCASE_LETTER: //Lu
+            case Character.LOWERCASE_LETTER: //Ll
+            case Character.TITLECASE_LETTER: //Lt
+            case Character.MODIFIER_LETTER: //Lm
+            case Character.OTHER_LETTER: //Lo
+            case Character.LETTER_NUMBER: //Nl
+                return true;
+            default:
+                //TODO if (ch in Other_Alphabetic) return true; (Probably need ICU4J for that)
+                //Other_Alphabetic contains mostly more exotic characters
+                return false;
         }
     }
-    
+
+    /**
+     * Indicates whether the given character is an explicit break-character
+     * @param ch    the character to check
+     * @return  true if the character represents an explicit break
+     */
+    public static boolean isExplicitBreak(char ch) {
+        return (ch == LINEFEED_CHAR
+            || ch == CARRIAGE_RETURN
+            || ch == NEXT_LINE
+            || ch == LINE_SEPARATOR
+            || ch == PARAGRAPH_SEPARATOR);
+    }
 }