Merge pull request #6588 from yurydelendik/spacing

Better spacing in text layer.

src/core/evaluator.js

@@ -350,9 +350,6 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
 
         for (var i = 0, ii = glyphs.length; i < ii; i++) {
           var glyph = glyphs[i];
-          if (glyph === null) {
-            continue;
-          }
           buildPath(glyph.fontChar);
 
           // If the glyph has an accent we need to build a path for its
@@ -1012,10 +1009,6 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
         var defaultVMetrics = font.defaultVMetrics;
         for (var i = 0; i < glyphs.length; i++) {
           var glyph = glyphs[i];
-          if (!glyph) { // Previous glyph was a space.
-            width += textState.wordSpacing * textState.textHScale;
-            continue;
-          }
           var vMetricX = null;
           var vMetricY = null;
           var glyphWidth = null;
@@ -1051,11 +1044,14 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
           // var x = pt[0];
           // var y = pt[1];
 
-          var charSpacing = 0;
+          var charSpacing = textState.charSpacing;
-          if (textChunk.str.length > 0) {
+          if (glyph.isSpace) {
-            // Apply char spacing only when there are chars.
+            var wordSpacing = textState.wordSpacing;
-            // As a result there is only spacing between glyphs.
+            charSpacing += wordSpacing;
-            charSpacing = textState.charSpacing;
+            if (wordSpacing > 0) {
+              addFakeSpaces(wordSpacing * 1000 / textState.fontSize,
+                            textChunk.str);
+            }
           }
 
           var tx = 0;
@@ -1089,6 +1085,22 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
         return textChunk;
       }
 
+      function addFakeSpaces(width, strBuf) {
+        var spaceWidth = textState.font.spaceWidth;
+        if (spaceWidth <= 0) {
+          return;
+        }
+        var fakeSpaces = width / spaceWidth;
+        if (fakeSpaces > MULTI_SPACE_FACTOR) {
+          fakeSpaces = Math.round(fakeSpaces);
+          while (fakeSpaces--) {
+            strBuf.push(' ');
+          }
+        } else if (fakeSpaces > SPACE_FACTOR) {
+          strBuf.push(' ');
+        }
+      }
+
       var timeSlotManager = new TimeSlotManager();
 
       return new Promise(function next(resolve, reject) {
@@ -1167,29 +1179,26 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
                   // In the default coordinate system, a positive adjustment
                   // has the effect of moving the next glyph painted either to
                   // the left or down by the given amount.
-                  var val = items[j] * textState.fontSize / 1000;
+                  var advance = items[j];
+                  var val = advance * textState.fontSize / 1000;
                   if (textState.font.vertical) {
-                    offset = val * textState.textMatrix[3];
+                    offset = val *
-                    textState.translateTextMatrix(0, offset);
+                      (textState.textHScale * textState.textMatrix[2] +
+                       textState.textMatrix[3]);
+                    textState.translateTextMatrix(0, val);
                     // Value needs to be added to height to paint down.
                     textChunk.height += offset;
                   } else {
-                    offset = val * textState.textHScale *
+                    offset = val * (
-                                   textState.textMatrix[0];
+                      textState.textHScale * textState.textMatrix[0] +
-                    textState.translateTextMatrix(offset, 0);
+                      textState.textMatrix[1]);
+                    textState.translateTextMatrix(-val, 0);
                     // Value needs to be subtracted from width to paint left.
                     textChunk.width -= offset;
+                    advance = -advance;
                   }
-                  if (items[j] < 0 && textState.font.spaceWidth > 0) {
+                  if (advance > 0) {
-                    var fakeSpaces = -items[j] / textState.font.spaceWidth;
+                    addFakeSpaces(advance, textChunk.str);
-                    if (fakeSpaces > MULTI_SPACE_FACTOR) {
-                      fakeSpaces = Math.round(fakeSpaces);
-                      while (fakeSpaces--) {
-                        textChunk.str.push(' ');
-                      }
-                    } else if (fakeSpaces > SPACE_FACTOR) {
-                      textChunk.str.push(' ');
-                    }
                   }
                 }
               }

src/core/fonts.js

@@ -2170,23 +2170,26 @@ function getFontType(type, subtype) {
 }
 
 var Glyph = (function GlyphClosure() {
-  function Glyph(fontChar, unicode, accent, width, vmetric, operatorListId) {
+  function Glyph(fontChar, unicode, accent, width, vmetric, operatorListId,
+                 isSpace) {
     this.fontChar = fontChar;
     this.unicode = unicode;
     this.accent = accent;
     this.width = width;
     this.vmetric = vmetric;
     this.operatorListId = operatorListId;
+    this.isSpace = isSpace;
   }
 
-  Glyph.prototype.matchesForCache =
+  Glyph.prototype.matchesForCache = function(fontChar, unicode, accent, width,
-      function(fontChar, unicode, accent, width, vmetric, operatorListId) {
+                                             vmetric, operatorListId, isSpace) {
     return this.fontChar === fontChar &&
            this.unicode === unicode &&
            this.accent === accent &&
            this.width === width &&
            this.vmetric === vmetric &&
-           this.operatorListId === operatorListId;
+           this.operatorListId === operatorListId &&
+           this.isSpace === isSpace;
   };
 
   return Glyph;
@@ -4701,7 +4704,7 @@ var Font = (function FontClosure() {
       return width;
     },
 
-    charToGlyph: function Font_charToGlyph(charcode) {
+    charToGlyph: function Font_charToGlyph(charcode, isSpace) {
       var fontCharCode, width, operatorListId;
 
       var widthCode = charcode;
@@ -4744,9 +4747,9 @@ var Font = (function FontClosure() {
       var glyph = this.glyphCache[charcode];
       if (!glyph ||
           !glyph.matchesForCache(fontChar, unicode, accent, width, vmetric,
-                                 operatorListId)) {
+                                 operatorListId, isSpace)) {
         glyph = new Glyph(fontChar, unicode, accent, width, vmetric,
-                          operatorListId);
+                          operatorListId, isSpace);
         this.glyphCache[charcode] = glyph;
       }
       return glyph;
@@ -4782,22 +4785,16 @@ var Font = (function FontClosure() {
           charcode = c.charcode;
           var length = c.length;
           i += length;
-          glyph = this.charToGlyph(charcode);
+          // Space is char with code 0x20 and length 1 in multiple-byte codes.
+          var isSpace = length === 1 && chars.charCodeAt(i - 1) === 0x20;
+          glyph = this.charToGlyph(charcode, isSpace);
           glyphs.push(glyph);
-          // placing null after each word break charcode (ASCII SPACE)
-          // Ignore occurences of 0x20 in multiple-byte codes.
-          if (length === 1 && chars.charCodeAt(i - 1) === 0x20) {
-            glyphs.push(null);
-          }
         }
       } else {
         for (i = 0, ii = chars.length; i < ii; ++i) {
           charcode = chars.charCodeAt(i);
-          glyph = this.charToGlyph(charcode);
+          glyph = this.charToGlyph(charcode, charcode === 0x20);
           glyphs.push(glyph);
-          if (charcode === 0x20) {
-            glyphs.push(null);
-          }
         }
       }
 

src/display/canvas.js

@@ -1397,16 +1397,13 @@ var CanvasGraphics = (function CanvasGraphicsClosure() {
       var x = 0, i;
       for (i = 0; i < glyphsLength; ++i) {
         var glyph = glyphs[i];
-        if (glyph === null) {
+        if (isNum(glyph)) {
-          // word break
-          x += fontDirection * wordSpacing;
-          continue;
-        } else if (isNum(glyph)) {
           x += spacingDir * glyph * fontSize / 1000;
           continue;
         }
 
         var restoreNeeded = false;
+        var spacing = (glyph.isSpace ? wordSpacing : 0) + charSpacing;
         var character = glyph.fontChar;
         var accent = glyph.accent;
         var scaledX, scaledY, scaledAccentX, scaledAccentY;
@@ -1450,7 +1447,7 @@ var CanvasGraphics = (function CanvasGraphicsClosure() {
           }
         }
 
-        var charWidth = width * widthAdvanceScale + charSpacing * fontDirection;
+        var charWidth = width * widthAdvanceScale + spacing * fontDirection;
         x += charWidth;
 
         if (restoreNeeded) {
@@ -1495,18 +1492,14 @@ var CanvasGraphics = (function CanvasGraphicsClosure() {
 
       for (i = 0; i < glyphsLength; ++i) {
         glyph = glyphs[i];
-        if (glyph === null) {
+        if (isNum(glyph)) {
-          // word break
-          this.ctx.translate(wordSpacing, 0);
-          current.x += wordSpacing * textHScale;
-          continue;
-        } else if (isNum(glyph)) {
           spacingLength = spacingDir * glyph * fontSize / 1000;
           this.ctx.translate(spacingLength, 0);
           current.x += spacingLength * textHScale;
           continue;
         }
 
+        var spacing = (glyph.isSpace ? wordSpacing : 0) + charSpacing;
         var operatorList = font.charProcOperatorList[glyph.operatorListId];
         if (!operatorList) {
           warn('Type3 character \"' + glyph.operatorListId +
@@ -1521,7 +1514,7 @@ var CanvasGraphics = (function CanvasGraphicsClosure() {
         this.restore();
 
         var transformed = Util.applyTransform([glyph.width, 0], fontMatrix);
-        width = transformed[0] * fontSize + charSpacing;
+        width = transformed[0] * fontSize + spacing;
 
         ctx.translate(width, 0);
         current.x += width * textHScale;

test/pdfs/.gitignore

@@ -29,6 +29,7 @@
 !issue3205r.pdf
 !issue3879r.pdf
 !close-path-bug.pdf
+!issue6019.pdf
 !issue1293r.pdf
 !issue6541.pdf
 !issue2948.pdf

test/pdfs/issue6019.pdf

@@ -0,0 +1,80 @@
+%PDF-1.7
+%âãÏÓ
+1 0 obj 
+<<
+/Type /Encoding
+/BaseEncoding /WinAnsiEncoding
+>>
+endobj 
+2 0 obj 
+<<
+/Pages 3 0 R
+/Type /Catalog
+>>
+endobj 
+3 0 obj 
+<<
+/Kids [4 0 R]
+/Count 1
+/Type /Pages
+>>
+endobj 
+4 0 obj 
+<<
+/Rotate 90
+/Parent 3 0 R
+/MediaBox [0 0 50 400]
+/Resources 
+<<
+/Font 
+<<
+/F1 5 0 R
+>>
+>>
+/Contents 6 0 R
+/Type /Page
+>>
+endobj 
+5 0 obj 
+<<
+/BaseFont /Times-Italic
+/Subtype /Type1
+/Encoding 1 0 R
+/Type /Font
+>>
+endobj 
+6 0 obj 
+<<
+/Length 108
+>>
+stream
+BT
+/F1 1 Tf
+0 24 -24 0 30 30 Tm
+0 Tw
+(Ð)Tj
+0.94 0 TD
+-0.0002 Tc
+0.0002 Tw
+[(Medicine Bow)-3653(  936)]TJ
+ET
+
+endstream 
+endobj xref
+0 7
+0000000000 65535 f 
+0000000015 00000 n 
+0000000085 00000 n 
+0000000136 00000 n 
+0000000195 00000 n 
+0000000335 00000 n 
+0000000426 00000 n 
+trailer
+
+<<
+/Root 2 0 R
+/Size 7
+>>
+startxref
+587
+%%EOF

test/test_manifest.json

@@ -1288,6 +1288,13 @@
        "link": false,
        "type": "eq"
     },
+    {
+      "id": "issue6019-text",
+      "file": "pdfs/issue6019.pdf",
+      "md5": "7a2e5dda3b0fc5c2e9060e378a8cdc4e",
+      "rounds": 1,
+      "type": "text"
+    },
     {  "id": "bug893730",
        "file": "pdfs/bug893730.pdf",
        "md5": "2587379fb1b3bbff89c14f0863e78383",