Rewrite the line wrapper to use an implementation of the Unicode Line Breaking Algorithm

Should solve a huge number of issues. The regular express based word matching from before was not good. It was overzealous and caused the most bugs of anything in the project. It also didn't work at all for languages like Chinese which don't have spaces between words.

The new wrapping algorithm supports all of this. It is a separate module at http://github.com/devongovett/linebreak.

lib/line_wrapper.coffee

@@ -1,6 +1,5 @@
-# This regular expression is used for splitting a string into wrappable words
-WORD_RE = /([^ ,\/!.?:;\-\n]*[ ,\/!.?:;\-]*)|\n/g
 {EventEmitter} = require 'events'
+LineBreaker = require 'linebreak'
 
 class LineWrapper extends EventEmitter
     constructor: (@document) ->
@@ -23,7 +22,7 @@ class LineWrapper extends EventEmitter
                 options.align = align
                 @lastLine = false
         
-    wrap: (paragraphs, options) ->
+    wrap: (text, options) ->
         width = @document.widthOfString.bind(@document)
         indent = options.indent or 0
         charSpacing = options.characterSpacing or 0
@@ -46,57 +45,65 @@ class LineWrapper extends EventEmitter
         
         # word width cache
         wordWidths = {}
-        
         @emit 'sectionStart', options, this
         
-        for text, i in paragraphs
-            @emit 'firstLine', options, this
+        breaker = new LineBreaker(text)
+        last = null
+        buffer = ''
+        textWidth = 0
+        wc = 0
+        
+        emitLine = =>
+            options.textWidth = textWidth + wordSpacing * (wc - 1)
+            options.wordCount = wc
+            options.lineWidth = @lineWidth
+            @emit 'line', buffer, options, this
+        
+        spaceLeft = @lineWidth - indent
+        
+        while bk = breaker.nextBreak()
+            if not last? or last.required
+                @emit 'firstLine', options, this
+                
+            word = text.slice(last?.position or 0, bk.position)
+            w = wordWidths[word] ?= width(word, options) + charSpacing + wordSpacing
             
-            # split the line into words
-            words = text.match(WORD_RE) or [text]
-                          
-            # space left on the line to fill with words
-            spaceLeft = @lineWidth - indent
-            options.lineWidth = spaceLeft
-            
-            len = words.length
-            buffer = ''
-            wc = 0
-            
-            for word, wi in words
-                w = wordWidths[word] ?= width(word, options) + charSpacing + wordSpacing
-
-                if w > spaceLeft or word is '\n'
-                    options.textWidth = width(buffer.trim(), options) + wordSpacing * (wc - 1)
-                    @emit 'line', buffer.trim(), options, this
-                                        
-                    # if we've reached the edge of the page, 
-                    # continue on a new page or column
-                    if @document.y > @maxY
-                        @nextSection()
+            if w <= spaceLeft
+                buffer += word
+                lineWidth += w
+                wc++
                             
+            if bk.required or w > spaceLeft
+                if bk.required
+                    @emit 'lastLine', options, this
+                
+                emitLine()
+                
+                # if we've reached the edge of the page, 
+                # continue on a new page or column
+                if @document.y > @maxY
+                    @nextSection()
+                
+                # reset the space left and buffer
+                if bk.required
+                    spaceLeft = @lineWidth - indent                    
+                    buffer = ''
+                    lineWidth = 0
+                    wc = 0
+                else
                     # reset the space left and buffer
                     spaceLeft = @lineWidth - w
-                    buffer = if word is '\n' then '' else word
+                    buffer = word
+                    lineWidth = w
                     wc = 1
-                            
-                else
-                    # add the word to the buffer
-                    spaceLeft -= w
-                    buffer += word
-                    wc++
-                        
-            # add the last line
-            @lastLine = true
-            @emit 'lastLine', options, this
-            options.textWidth = width(buffer.trim(), options) + wordSpacing * (wc - 1)
-            @emit 'line', buffer.trim(), options, this
+            else
+                spaceLeft -= w
+                
+            last = bk
             
-            # make sure that the first line of a paragraph is never by 
-            # itself at the bottom of a page (orphans)
-            nextY = @document.y + @document.currentLineHeight(true)
-            if i < paragraphs.length - 1 and nextY > @maxY
-                @nextSection()
+        if wc > 0
+            @emit 'lastLine', options, this
+            emitLine()
                 
         @emit 'sectionEnd', options, this
                     

lib/mixins/text.coffee

@@ -1,4 +1,3 @@
-WORD_RE = /([^ ,\/!.?:;\-\n]*[ ,\/!.?:;\-]*)|\n/g
 LineWrapper = require '../line_wrapper'
 
 module.exports = 
@@ -34,18 +33,16 @@ module.exports =
         # if the wordSpacing option is specified, remove multiple consecutive spaces
         if options.wordSpacing
             text = text.replace(/\s{2,}/g, ' ')
-            
-        paragraphs = text.split '\n'
-        
+
         # word wrapping
         if options.width
             wrapper = new LineWrapper(this)
             wrapper.on 'line', @_line.bind(this)
-            wrapper.wrap(paragraphs, options)
+            wrapper.wrap text, options
             
         # render paragraphs as single lines
         else
-            @_line line, options for line in paragraphs
+            @_line line, options for line in text.split '\n'
         
         return this
         
@@ -97,7 +94,7 @@ module.exports =
             @x -= pos
             wrapper.lineWidth += pos
                     
-        wrapper.wrap(items, options)
+        wrapper.wrap items.join('\n'), options
         
         @x -= indent
         return this

package.json

@@ -1,29 +1,39 @@
 {
   "name": "pdfkit",
   "description": "A PDF generation library for Node.js",
-  "keywords": ["pdf", "pdf writer", "pdf generator", "graphics", "document", "vector"],
+  "keywords": [
+    "pdf",
+    "pdf writer",
+    "pdf generator",
+    "graphics",
+    "document",
+    "vector"
+  ],
   "version": "0.2.8",
   "homepage": "http://pdfkit.org/",
   "author": {
-      "name": "Devon Govett",
-      "email": "devongovett@gmail.com",
-      "url": "http://badassjs.com/"
+    "name": "Devon Govett",
+    "email": "devongovett@gmail.com",
+    "url": "http://badassjs.com/"
   },
   "repository": {
-      "type": "git",
-      "url": "https://github.com/devongovett/pdfkit.git"
+    "type": "git",
+    "url": "https://github.com/devongovett/pdfkit.git"
   },
   "bugs": "http://github.com/devongovett/pdfkit/issues",
   "devDependencies": {
-      "coffee-script": ">=1.0.1"
+    "coffee-script": ">=1.0.1"
   },
   "dependencies": {
-      "png-js": ">=0.1.0"
+    "png-js": ">=0.1.0",
+    "linebreak": "~0.1.0"
   },
   "scripts": {
-      "prepublish": "coffee -o js -c lib/ && cp -r lib/font/data js/font/data",
-      "postpublish": "rm -rf ./js"
+    "prepublish": "coffee -o js -c lib/ && cp -r lib/font/data js/font/data",
+    "postpublish": "rm -rf ./js"
   },
   "main": "./js/document",
-  "engine": [ "node >= v0.6.0" ]
-}
+  "engine": [
+    "node >= v0.6.0"
+  ]
+}