feature(excludeTrailingPunctuationFromURLs): excludes trailing punctuation from auto linked URLs

Closes #266, #308

README.md

@@ -210,7 +210,7 @@ var defaultOptions = showdown.getDefaultOptions();
     <h3>foo</h3>
     ```
 
- * **simplifiedAutoLink**: (boolean) [default false] Turning this on will enable GFM autolink style. This means that 
+ * **simplifiedAutoLink**: (boolean) [default false] Turning this option on will enable automatic linking to urls. This means that 
 
    ```md
    some text www.google.com
@@ -219,6 +219,17 @@ var defaultOptions = showdown.getDefaultOptions();
    ````
    <p>some text <a href="www.google.com">www.google.com</a>
    ```
+ 
+ * **excludeTrailingPunctuationFromURLs**: (boolean) [default false] This option excludes trailing punctuation from autolinking urls.
+   Punctuation excluded: `. !  ? ( )`. Only applies if **simplifiedAutoLink** option is set to `true`.
+   
+   ```md
+   check this link www.google.com!
+   ```
+   will be parsed as
+   ```html
+   <p>check this link <a href="www.google.com">www.google.com</a>!</p>
+   ```
    
  * **literalMidWordUnderscores**: (boolean) [default false] Turning this on will stop showdown from interpreting underscores in the middle of words as `<em>` and `<strong>` and instead treat them as literal underscores. 
 

dist/showdown.js

@@ -1,4 +1,4 @@
-;/*! showdown 30-11-2016 */
+;/*! showdown 01-12-2016 */
 (function(){
 /**
  * Created by Tivie on 13-07-2015.
@@ -38,6 +38,11 @@ function getDefaultOpts(simple) {
       describe: 'Turn on/off GFM autolink style',
       type: 'boolean'
     },
+    excludeTrailingPunctuationFromURLs: {
+      defaultValue: false,
+      describe: 'Excludes trailing punctuation from links generated with autoLinking',
+      type: 'boolean'
+    },
     literalMidWordUnderscores: {
       defaultValue: false,
       describe: 'Parse midword underscores as literal underscores',
@@ -115,6 +120,7 @@ var showdown = {},
         omitExtraWLInCodeBlocks:              true,
         prefixHeaderId:                       'user-content-',
         simplifiedAutoLink:                   true,
+        excludeTrailingPunctuationFromURLs:   true,
         literalMidWordUnderscores:            true,
         strikethrough:                        true,
         tables:                               true,
@@ -1185,9 +1191,10 @@ showdown.subParser('autoLinks', function (text, options, globals) {
 
   text = globals.converter._dispatch('autoLinks.before', text, options, globals);
 
-  var simpleURLRegex  = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+)(?=\s|$)(?!["<>])/gi,
+  var simpleURLRegex  = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+)()(?=\s|$)(?!["<>])/gi,
+      simpleURLRegex2 = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+?)([.!?()]?)(?=\s|$)(?!["<>])/gi,
       delimUrlRegex   = /<(((https?|ftp|dict):\/\/|www\.)[^'">\s]+)>/gi,
-      simpleMailRegex = /(?:^|\s)([A-Za-z0-9!#$%&'*+-/=?^_`\{|}~\.]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)(?:$|\s)/gi,
+      simpleMailRegex = /(?:^|\s)([A-Za-z0-9!#$%&'*+-/=?^_`{|}~.]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)(?:$|\s)/gi,
       delimMailRegex  = /<(?:mailto:)?([-.\w]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)>/gi;
 
   text = text.replace(delimUrlRegex, replaceLink);
@@ -1196,20 +1203,28 @@ showdown.subParser('autoLinks', function (text, options, globals) {
   // Email addresses: <address@domain.foo>
 
   if (options.simplifiedAutoLink) {
-    text = text.replace(simpleURLRegex, replaceLink);
+    if (options.excludeTrailingPunctuationFromURLs) {
+      text = text.replace(simpleURLRegex2, replaceLink);
+    } else {
+      text = text.replace(simpleURLRegex, replaceLink);
+    }
     text = text.replace(simpleMailRegex, replaceMail);
   }
 
-  function replaceLink(wm, link) {
-    var lnkTxt = link;
+  function replaceLink(wm, link, m2, m3, trailingPunctuation) {
+    var lnkTxt = link,
+        append = '';
     if (/^www\./i.test(link)) {
       link = link.replace(/^www\./i, 'http://www.');
     }
-    return '<a href="' + link + '">' + lnkTxt + '</a>';
+    if (options.excludeTrailingPunctuationFromURLs && trailingPunctuation) {
+      append = trailingPunctuation;
+    }
+    return '<a href="' + link + '">' + lnkTxt + '</a>' + append;
   }
 
-  function replaceMail(wholeMatch, m1) {
-    var unescapedStr = showdown.subParser('unescapeSpecialChars')(m1);
+  function replaceMail(wholeMatch, mail) {
+    var unescapedStr = showdown.subParser('unescapeSpecialChars')(mail);
     return showdown.subParser('encodeEmailAddress')(unescapedStr);
   }
 

src/options.js

@@ -36,6 +36,11 @@ function getDefaultOpts(simple) {
       describe: 'Turn on/off GFM autolink style',
       type: 'boolean'
     },
+    excludeTrailingPunctuationFromURLs: {
+      defaultValue: false,
+      describe: 'Excludes trailing punctuation from links generated with autoLinking',
+      type: 'boolean'
+    },
     literalMidWordUnderscores: {
       defaultValue: false,
       describe: 'Parse midword underscores as literal underscores',

src/showdown.js

@@ -12,6 +12,7 @@ var showdown = {},
         omitExtraWLInCodeBlocks:              true,
         prefixHeaderId:                       'user-content-',
         simplifiedAutoLink:                   true,
+        excludeTrailingPunctuationFromURLs:   true,
         literalMidWordUnderscores:            true,
         strikethrough:                        true,
         tables:                               true,

src/subParsers/autoLinks.js

@@ -3,9 +3,10 @@ showdown.subParser('autoLinks', function (text, options, globals) {
 
   text = globals.converter._dispatch('autoLinks.before', text, options, globals);
 
-  var simpleURLRegex  = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+)(?=\s|$)(?!["<>])/gi,
+  var simpleURLRegex  = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+)()(?=\s|$)(?!["<>])/gi,
+      simpleURLRegex2 = /\b(((https?|ftp|dict):\/\/|www\.)[^'">\s]+\.[^'">\s]+?)([.!?()]?)(?=\s|$)(?!["<>])/gi,
       delimUrlRegex   = /<(((https?|ftp|dict):\/\/|www\.)[^'">\s]+)>/gi,
-      simpleMailRegex = /(?:^|\s)([A-Za-z0-9!#$%&'*+-/=?^_`\{|}~\.]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)(?:$|\s)/gi,
+      simpleMailRegex = /(?:^|\s)([A-Za-z0-9!#$%&'*+-/=?^_`{|}~.]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)(?:$|\s)/gi,
       delimMailRegex  = /<(?:mailto:)?([-.\w]+@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)>/gi;
 
   text = text.replace(delimUrlRegex, replaceLink);
@@ -14,20 +15,28 @@ showdown.subParser('autoLinks', function (text, options, globals) {
   // Email addresses: <address@domain.foo>
 
   if (options.simplifiedAutoLink) {
-    text = text.replace(simpleURLRegex, replaceLink);
+    if (options.excludeTrailingPunctuationFromURLs) {
+      text = text.replace(simpleURLRegex2, replaceLink);
+    } else {
+      text = text.replace(simpleURLRegex, replaceLink);
+    }
     text = text.replace(simpleMailRegex, replaceMail);
   }
 
-  function replaceLink(wm, link) {
-    var lnkTxt = link;
+  function replaceLink(wm, link, m2, m3, trailingPunctuation) {
+    var lnkTxt = link,
+        append = '';
     if (/^www\./i.test(link)) {
       link = link.replace(/^www\./i, 'http://www.');
     }
-    return '<a href="' + link + '">' + lnkTxt + '</a>';
+    if (options.excludeTrailingPunctuationFromURLs && trailingPunctuation) {
+      append = trailingPunctuation;
+    }
+    return '<a href="' + link + '">' + lnkTxt + '</a>' + append;
   }
 
-  function replaceMail(wholeMatch, m1) {
-    var unescapedStr = showdown.subParser('unescapeSpecialChars')(m1);
+  function replaceMail(wholeMatch, mail) {
+    var unescapedStr = showdown.subParser('unescapeSpecialChars')(mail);
     return showdown.subParser('encodeEmailAddress')(unescapedStr);
   }
 

test/features/excludeTrailingPunctuationFromURLs-option.html

@@ -0,0 +1,4 @@
+<p>url <a href="http://www.google.com">http://www.google.com</a>.</p>
+<p>url <a href="http://www.google.com">http://www.google.com</a>!</p>
+<p>url <a href="http://www.google.com">http://www.google.com</a>? foo</p>
+<p>url (<a href="http://www.google.com">http://www.google.com</a>) bazinga</p>

test/features/excludeTrailingPunctuationFromURLs-option.md

@@ -0,0 +1,7 @@
+url http://www.google.com.
+
+url http://www.google.com!
+
+url http://www.google.com? foo
+
+url (http://www.google.com) bazinga

test/node/testsuite.features.js

@@ -37,6 +37,8 @@ describe('makeHtml() features testsuite', function () {
       converter = new showdown.Converter({disableForced4SpacesIndentedSublists: true});
     } else if (testsuite[i].name === '#206.treat-single-line-breaks-as-br') {
       converter = new showdown.Converter({simpleLineBreaks: true});
+    } else if (testsuite[i].name === 'excludeTrailingPunctuationFromURLs-option') {
+      converter = new showdown.Converter({simplifiedAutoLink: true, excludeTrailingPunctuationFromURLs: true});
     } else {
       converter = new showdown.Converter();
     }