mailparser - npm Package Compare versions

Comparing version 0.2.33 to 0.3.0

lib/mailparser.js

@@ -865,4 +865,3 @@
  * <p>Finds text parts and attachments from the tree. If there's several text/plain
- * or text/html parts, push the ones from the lower parts of the tree to the
- * alternatives array</p>
+ * or text/html parts, join these into one</p>
  *
@@ -875,3 +874,3 @@ * <p>Emits "end" when finished</p>
 
-    this.mailData = {html:[], text:[], alternatives:[], attachments:[]};
+    this.mailData = {html:[], text:[], attachments:[]};
 
@@ -886,22 +885,6 @@ if(!this.mimeTree.meta.mimeMultipart){
         for(i=0, len=this.mailData.html.length; i<len; i++){
-            if(!returnValue.html || this.mailData.html[i].level < htmlLevel){
-                if(returnValue.html){
-                    if(!returnValue.alternatives){
-                        returnValue.alternatives = [];
-                    }
-                    returnValue.alternatives.push({
-                        contentType: "text/html",
-                        content: returnValue.html
-                    });
-                }
-                htmlLevel = this.mailData.html[i].level;
+            if(!returnValue.html && this.mailData.html[i].content){
                 returnValue.html = this.mailData.html[i].content;
-            }else{
-                if(!returnValue.alternatives){
-                    returnValue.alternatives = [];
-                }
-                returnValue.alternatives.push({
-                    contentType: "text/html",
-                    content: this.mailData.html[i].content
-                });
+            }else if(this.mailData.html[i].content){
+                returnValue.html = this._concatHTML(returnValue.html, this.mailData.html[i].content);
             }
@@ -913,22 +896,6 @@ }
         for(i=0, len=this.mailData.text.length; i<len; i++){
-            if(!returnValue.text || this.mailData.text[i].level < textLevel){
-                if(returnValue.text){
-                    if(!returnValue.alternatives){
-                        returnValue.alternatives = [];
-                    }
-                    returnValue.alternatives.push({
-                        contentType: "text/plain",
-                        content: returnValue.text
-                    });
-                }
-                textLevel = this.mailData.text[i].level;
+            if(!returnValue.text && this.mailData.text[i].content){
                 returnValue.text = this.mailData.text[i].content;
-            }else{
-                if(!returnValue.alternatives){
-                    returnValue.alternatives = [];
-                }
-                returnValue.alternatives.push({
-                    contentType: "text/plain",
-                    content: this.mailData.text[i].content
-                });
+            }else if(this.mailData.text[i].content){
+                returnValue.text += this.mailData.text[i].content;
             }
@@ -1109,2 +1076,60 @@ }
 /**
+ * <p>Joins two HTML blocks by removing the header of the added element<p>
+ *
+ * @param {Sting} htmlNode Original HTML contents
+ * @param {String} newHTML HTML text to add to the original object node
+ * @return {String} Joined HTML
+ */
+MailParser.prototype._concatHTML = function(firstNode, secondNode){
+    var inserted = false,
+        headerNode = "",
+        htmlHeader = "";
+
+    firstNode = (firstNode || "").toString("utf-8");
+    secondNode = (secondNode || "").toString("utf-8");
+
+    if(!secondNode){
+        return firstNode;
+    }
+    if(!firstNode){
+        return secondNode;
+    }
+
+    if(firstNode.substr(0, 1024).replace(/\r?\n/g,"\u0000").match(/^[\s\u0000]*(<\!doctype\b[^>]*?>)?[\s\u0000]*<(html|head)\b[^>]*?>/i)){
+        headerNode = firstNode;
+    }else if(secondNode.substr(0, 1024).replace(/\r?\n/g,"\u0000").match(/^[\s\u0000]*(<\!doctype\b[^>]*?>)?[\s\u0000]*<(html|head)\b[^>]*?>/i)){
+        headerNode = secondNode;
+    }
+
+    if(headerNode){
+        headerNode.replace(/\r?\n/g, "\u0000").replace(/^[\s\u0000]*(<\!doctype\b[^>]*?>)?[\s\u0000]*<(html|head)\b[^>]*>.*?<\/(head)\b[^>]*>(.*?<body\b[^>]*>)?/i, function(h){
+            var doctype = h.match(/^[\s\u0000]*(<\!doctype\b[^>]*?>)/i),
+                html = h.match(/<html\b[^>]*?>/i),
+                head = h.match(/<head\b[^>]*?>/i),
+                body = h.match(/<body\b[^>]*?>/i);
+
+            doctype = doctype && doctype[1] && doctype[1] + "\n" || "";
+            html = html && html[0] || "<head>";
+            head = head && head[0] || "<head>";
+            body = body && body[0] || "<body>";
+            h = h.replace(/<[\!\/]?(doctype|html|head|body)\b[^>]*?>/ig, "\u0000").replace(/\u0000+/g, "\n").trim();
+
+            htmlHeader = doctype + html + "\n" + head + (h ? h + "\n" : "") + "</head>\n" + body + "\n";
+        });
+    }
+
+    firstNode = firstNode.replace(/\r?\n/g, "\u0000").
+        replace(/[\s\u0000]*<head\b[^>]*>.*?<\/(head|body)\b[^>]*>/gi, "").
+        replace(/[\s\u0000]*<[\!\/]?(doctype|html|body)\b[^>]*>[\s\u0000]*/gi, "").
+        replace(/\u0000/g, "\n");
+    
+    secondNode = secondNode.replace(/\r?\n/g, "\u0000").
+        replace(/[\s\u0000]*<head\b[^>]*>.*?<\/(head|body)\b[^>]*>/gi, "").
+        replace(/[\s\u0000]*<[\!\/]?(doctype|html|body)\b[^>]*>[\s\u0000]*/gi, "").
+        replace(/\u0000/g, "\n");
+
+    return htmlHeader + firstNode + secondNode + (htmlHeader? (firstNode || secondNode ? "\n" : "") + "</body>\n</html>" : "");
+};
+
+/**
  * <p>Converts a string from one charset to another</p>
@@ -1111,0 +1136,0 @@ *

package.json

 {
     "name": "mailparser",
     "description": "Asynchronous and non-blocking parser for mime encoded e-mail messages",
-    "version": "0.2.33",
+    "version": "0.3.0",
     "author" : "Andris Reinman",
@@ -6,0 +6,0 @@ "maintainers":[

README.md

 MailParser
 ==========
 
-**NB!** MailParser is now out of the box compatible with Windows - instead of compiled [Iconv](https://npmjs.org/package/iconv) module, [iconv-lite](https://npmjs.org/package/iconv-lite) is used instead when in Windows.
-
 [![Build Status](https://secure.travis-ci.org/andris9/mailparser.png)](http://travis-ci.org/andris9/mailparser)
@@ -80,3 +78,2 @@
   * **html** - html body
-  * **alternatives** - an array of alternative bodies in addition to the default `html` and `text` - `[{contentType:"text/plain", content: "..."}]`
   * **attachments** - an array of attachments
@@ -83,0 +80,0 @@

test/mailparser.js

 var MailParser = require("../lib/mailparser").MailParser,
     testCase = require('nodeunit').testCase,
     utillib = require("util"),
-    encodinglib = require("encoding");
+    encodinglib = require("encoding"),
+    fs = require("fs");
 
@@ -889,106 +890,2 @@ exports["General tests"] = {
         });
-    },
-    "Alternative content - Main TEXT first": function(test){
-        var encodedText = "Content-type: multipart/mixed; boundary=ABC\r\n"+
-                          "\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: text/plain; charset=utf-8\r\n"+
-                              "\r\n"+
-                              "ÕÄÖÜ1\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: multipart/related; boundary=DEF\r\n"+
-                              "\r\n"+
-                              "--DEF\r\n"+
-                                  "Content-type: text/plain; charset=utf-8\r\n"+
-                                   "\r\n"+
-                                   "ÕÄÖÜ2\r\n"+
-                              "--DEF--\r\n"+
-                           "--ABC--",
-            mail = new Buffer(encodedText, "utf-8");
-
-        var mailparser = new MailParser();
-        mailparser.end(mail);
-
-        mailparser.on("end", function(mail){
-            test.equal(mail.text, "ÕÄÖÜ1");
-            test.done();
-        });
-    },
-    "Alternative content - Main TEXT last": function(test){
-        var encodedText = "Content-type: multipart/mixed; boundary=ABC\r\n"+
-                          "\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: multipart/related; boundary=DEF\r\n"+
-                              "\r\n"+
-                              "--DEF\r\n"+
-                                  "Content-type: text/plain; charset=utf-8\r\n"+
-                                   "\r\n"+
-                                   "ÕÄÖÜ2\r\n"+
-                              "--DEF--\r\n"+
-                           "--ABC\r\n"+
-                              "Content-type: text/plain; charset=utf-8\r\n"+
-                              "\r\n"+
-                              "ÕÄÖÜ1\r\n"+
-                           "--ABC--",
-            mail = new Buffer(encodedText, "utf-8");
-
-        var mailparser = new MailParser();
-        mailparser.end(mail);
-
-        mailparser.on("end", function(mail){
-            test.equal(mail.text, "ÕÄÖÜ1");
-            test.done();
-        });
-    },
-    "Alternative content - Main HTML first": function(test){
-        var encodedHTML = "Content-type: multipart/mixed; boundary=ABC\r\n"+
-                          "\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: text/html; charset=utf-8\r\n"+
-                              "\r\n"+
-                              "ÕÄÖÜ1\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: multipart/related; boundary=DEF\r\n"+
-                              "\r\n"+
-                              "--DEF\r\n"+
-                                  "Content-type: text/html; charset=utf-8\r\n"+
-                                   "\r\n"+
-                                   "ÕÄÖÜ2\r\n"+
-                              "--DEF--\r\n"+
-                           "--ABC--",
-            mail = new Buffer(encodedHTML, "utf-8");
-
-        var mailparser = new MailParser();
-        mailparser.end(mail);
-
-        mailparser.on("end", function(mail){
-            test.equal(mail.html, "ÕÄÖÜ1");
-            test.done();
-        });
-    },
-    "Alternative content - Main HTML last": function(test){
-        var encodedHTML = "Content-type: multipart/mixed; boundary=ABC\r\n"+
-                          "\r\n"+
-                          "--ABC\r\n"+
-                              "Content-type: multipart/related; boundary=DEF\r\n"+
-                              "\r\n"+
-                              "--DEF\r\n"+
-                                  "Content-type: text/html; charset=utf-8\r\n"+
-                                   "\r\n"+
-                                   "ÕÄÖÜ2\r\n"+
-                              "--DEF--\r\n"+
-                           "--ABC\r\n"+
-                              "Content-type: text/html; charset=utf-8\r\n"+
-                              "\r\n"+
-                              "ÕÄÖÜ1\r\n"+
-                           "--ABC--",
-            mail = new Buffer(encodedHTML, "utf-8");
-
-        var mailparser = new MailParser();
-        mailparser.end(mail);
-
-        mailparser.on("end", function(mail){
-            test.equal(mail.html, "ÕÄÖÜ1");
-            test.done();
-        });
     }
@@ -1220,3 +1117,53 @@ };
         });
+    },
+
+    "Inline attachments": function(test){
+        var encodedText = "Content-type: multipart/mixed; boundary=ABC\r\n"+
+                          "X-Test: =?UTF-8?Q?=C3=95=C3=84?= =?UTF-8?Q?=C3=96=C3=9C?=\r\n"+
+                          "Subject: ABCDEF\r\n"+
+                          "\r\n"+
+                          "--ABC\r\n"+
+                              "Content-Type: text/html\r\n"+
+                              "\r\n"+
+                              "<p>test 1</p>\r\n"+
+                          "--ABC\r\n"+
+                              "Content-Type: application/octet-stream\r\n"+
+                              "Content-Transfer-Encoding: base64\r\n"+
+                              "Content-Disposition: attachment; filename=\"test.pdf\"\r\n"+
+                              "\r\n"+
+                              "AAECAwQFBg==\r\n"+
+                          "--ABC\r\n"+
+                              "Content-Type: text/html\r\n"+
+                              "\r\n"+
+                              "<p>test 2</p>\r\n"+
+                          "--ABC--",
+            mail = new Buffer(encodedText, "utf-8");
+
+        test.expect(1);
+        var mailparser = new MailParser({showAttachmentLinks: true});
+
+        mailparser.end(mail);
+        mailparser.on("end", function(mail){
+            test.equal(mail.html, '<p>test 1</p><br/>\n\n<div class="mailparser-attachment"><a href="cid:test.pdf@node">&lt;test.pdf&gt;</a></div><br/>\n<p>test 2</p>');
+            test.done();
+        });
     }
 };
+
+exports["Advanced nested HTML"] = function(test){
+        var mail = fs.readFileSync(__dirname + "/nested.eml");
+
+        test.expect(2);
+        var mailparser = new MailParser();
+
+        for(var i=0, len = mail.length; i<len; i++){
+            mailparser.write(new Buffer([mail[i]]));
+        }
+
+        mailparser.end();
+        mailparser.on("end", function(mail){
+            test.equal(mail.text, "\nDear Sir,\n\nGood evening.\n\n\n \n\n\n\nThe footer\n");
+            test.equal(mail.html, "<p>Dear Sir</p>\n<p>Good evening.</p>\n<p></p><p>The footer</p>\n");
+            test.done();
+        });
+    }

New alerts

Filesystem access

Supply chain risk

Accesses the file system, and could potentially read sensitive data.

Found 1 instance in 1 package

Improved metrics

Number of package files

11

increased by10%

Worsened metrics

Total package byte prevSize

119523

decreased by-0.07%

Lines of code

2576

decreased by-1.42%

Number of lines in readme file

217

decreased by-1.36%

Number of low supply chain risk alerts

1

increased byInfinity%

No dependency changes