Bug 747820 - Style editor breaks with non-latin encoding. r=msucan f=cedricv

browser/devtools/styleeditor/StyleEditor.jsm

@@ -758,6 +758,87 @@ StyleEditor.prototype = {
     }
   },
 
+  /**
+   * Decode a CSS source string to unicode according to the character set rules
+   * defined in <http://www.w3.org/TR/CSS2/syndata.html#charset>.
+   *
+   * @param string aString
+   *        Source of a CSS stylesheet, loaded from file or cache.
+   * @param string aChannelCharset
+   *        Charset of the source string if set by the HTTP channel.
+   * @return string
+   *         The CSS string, in unicode.
+   */
+  _decodeCSSCharset: function SE__decodeCSSCharset(aString, aChannelCharset)
+  {
+    // StyleSheet's charset can be specified from multiple sources
+
+    if (aChannelCharset.length > 0) {
+      // step 1 of syndata.html: charset given in HTTP header.
+      return this._convertToUnicode(aString, aChannelCharset);
+    }
+
+    let sheet = this.styleSheet;
+    if (sheet) {
+      // Do we have a @charset rule in the stylesheet?
+      // step 2 of syndata.html (without the BOM check).
+      if (sheet.cssRules) {
+        let rules = sheet.cssRules;
+        if (rules.length
+            && rules.item(0).type == Ci.nsIDOMCSSRule.CHARSET_RULE) {
+          return this._convertToUnicode(aString, rules.item(0).encoding);
+        }
+      }
+
+      if (sheet.ownerNode) {
+        // step 3: see <link charset="…">
+        let linkCharset = sheet.ownerNode.getAttribute("charset");
+        if (linkCharset != null) {
+          return this._convertToUnicode(aString, linkCharset);
+        }
+      }
+
+      // step 4 (1 of 2): charset of referring stylesheet.
+      let parentSheet = sheet.parentStyleSheet;
+      if (parentSheet && parentSheet.cssRules &&
+          parentSheet.cssRules[0].type == Ci.nsIDOMCSSRule.CHARSET_RULE) {
+        return this._convertToUnicode(aString,
+            parentSheet.cssRules[0].encoding);
+      }
+
+      // step 4 (2 of 2): charset of referring document.
+      if (sheet.ownerNode && sheet.ownerNode.ownerDocument.characterSet) {
+        return this._convertToUnicode(aString,
+            sheet.ownerNode.ownerDocument.characterSet);
+      }
+    }
+
+    // step 5: default to utf-8.
+    return this._convertToUnicode(aString, "UTF-8");
+  },
+
+  /**
+   * Convert a given string, encoded in a given character set, to unicode.
+   * @param string aString
+   *        A string.
+   * @param string aCharset
+   *        A character set.
+   * @return string
+   *         A unicode string.
+   */
+  _convertToUnicode: function SE__convertToUnicode(aString, aCharset) {
+    // Decoding primitives.
+    let converter = Cc["@mozilla.org/intl/scriptableunicodeconverter"]
+        .createInstance(Ci.nsIScriptableUnicodeConverter);
+
+    try {
+      converter.charset = aCharset;
+      return converter.ConvertToUnicode(aString);
+    } catch(e) {
+      return aString;
+    }
+  },
+
   /**
    * Load source from a file or file-like resource.
    *
@@ -790,6 +871,7 @@ StyleEditor.prototype = {
   {
     let channel = Services.io.newChannel(aHref, null, null);
     let chunks = [];
+    let channelCharset = "";
     let streamListener = { // nsIStreamListener inherits nsIRequestObserver
       onStartRequest: function (aRequest, aContext, aStatusCode) {
         if (!Components.isSuccessCode(aStatusCode)) {
@@ -797,6 +879,10 @@ StyleEditor.prototype = {
         }
       }.bind(this),
       onDataAvailable: function (aRequest, aContext, aStream, aOffset, aCount) {
+        let channel = aRequest.QueryInterface(Ci.nsIChannel);
+        if (!channelCharset) {
+          channelCharset = channel.contentCharset;
+        }
         chunks.push(NetUtil.readInputStreamToString(aStream, aCount));
       },
       onStopRequest: function (aRequest, aContext, aStatusCode) {
@@ -804,7 +890,7 @@ StyleEditor.prototype = {
           return this._signalError(LOAD_ERROR);
         }
 
-        this._onSourceLoad(chunks.join(""));
+        this._onSourceLoad(chunks.join(""), channelCharset);
       }.bind(this)
     };
 
@@ -816,9 +902,14 @@ StyleEditor.prototype = {
    * Called when source has been loaded.
    *
    * @param string aSourceText
+   * @param string aCharset
+   *        Optional. The character set to use. The default is to detect the
+   *        character set following the standard (see
+   *        <http://www.w3.org/TR/CSS2/syndata.html#charset>).
    */
-  _onSourceLoad: function SE__onSourceLoad(aSourceText)
+  _onSourceLoad: function SE__onSourceLoad(aSourceText, aCharset)
   {
+    aSourceText = this._decodeCSSCharset(aSourceText, aCharset || "");
     this._restoreExpando();
     this._state.text = prettifyCSS(aSourceText);
     this._loaded = true;

browser/devtools/styleeditor/test/browser_styleeditor_init.js

@@ -71,6 +71,10 @@ function testEditorAdded(aChrome, aEditor)
 
 function testFirstStyleSheetEditor(aChrome, aEditor)
 {
+  // Note: the html <link> contains charset="UTF-8".
+  ok(aEditor._state.text.indexOf("\u263a") >= 0,
+     "stylesheet is unicode-aware.");
+
   //testing TESTCASE's simple.css stylesheet
   is(aEditor.styleSheetIndex, 0,
      "first stylesheet is at index 0");

browser/devtools/styleeditor/test/simple.css

@@ -1,6 +1,7 @@
 /* vim: set ts=2 et sw=2 tw=80: */
 /* Any copyright is dedicated to the Public Domain.
    http://creativecommons.org/publicdomain/zero/1.0/ */
+/* ☺ */
 
 body {
   margin: 0;

browser/devtools/styleeditor/test/simple.html

@@ -2,7 +2,7 @@
 <html>
 <head>
   <title>simple testcase</title>
-  <link rel="stylesheet" type="text/css" media="screen" href="simple.css"/>
+  <link rel="stylesheet" charset="UTF-8" type="text/css" media="screen" href="simple.css"/>
   <style type="text/css">
   body {
     background: white;