Bug 880648 - [MMS] Accented characters are not correctly sent/received in a MMS (part 1, content blob should be encoded by utf-8). r=vicamo a=leo+

dom/mobilemessage/interfaces/nsIDOMMozMmsMessage.idl

@@ -11,7 +11,8 @@ dictionary MmsAttachment
 {
   DOMString? id;
   DOMString? location;
-  nsIDOMBlob content;
+  nsIDOMBlob content; // If the content blob is a text/plain type, the encoding
+                      // for text should always be "utf-8".
 };
 
 [scriptable, builtinclass, uuid(2e5e1c16-b7af-11e2-af04-8f4b1610a600)]

dom/mobilemessage/src/ril/WspPduHelper.jsm

@@ -2292,18 +2292,40 @@ this.PduHelper = {
         let octetArray = Octet.decodeMultiple(data, contentEnd);
         let content = null;
         if (octetArray) {
-          // If the content is a SMIL type, convert it to a string.
-          // We hope to save and expose the SMIL content as a string way.
-          if (headers["content-type"].media == "application/smil") {
-            let charset = headers["content-type"].params &&
-                          headers["content-type"].params.charset
-                        ? headers["content-type"].params.charset["charset"]
-                        : null;
-            content = this.decodeStringContent(octetArray, charset);
+          let charset = headers["content-type"].params &&
+                        headers["content-type"].params.charset
+                      ? headers["content-type"].params.charset.charset
+                      : null;
+
+          let mimeType = headers["content-type"].media;
+
+          if (mimeType) {
+            if (mimeType == "application/smil") {
+              // If the content is a SMIL type, convert it to a string.
+              // We hope to save and expose the SMIL content in a string way.
+              content = this.decodeStringContent(octetArray, charset);
+            } else if (mimeType.indexOf("text/") == 0 && charset != "utf-8") {
+              // If the content is a "text/plain" type, we have to make sure
+              // the encoding of the blob content should always be "utf-8".
+              let tmpStr = this.decodeStringContent(octetArray, charset);
+              let encoder = new TextEncoder("UTF-8");
+              content = new Blob([encoder.encode(tmpStr)], {type : mimeType});
+
+              // Make up the missing encoding info.
+              if (!headers["content-type"].params) {
+                headers["content-type"].params = {};
+              }
+              if (!headers["content-type"].params.charset) {
+                headers["content-type"].params.charset = {};
+              }
+              if (!headers["content-type"].params.charset.charset) {
+                headers["content-type"].params.charset.charset = "utf-8";
+              }
+            }
           }
+
           if (!content) {
-            content = new Blob([octetArray],
-                               {type : headers["content-type"].media});
+            content = new Blob([octetArray], {type : mimeType});
           }
         }