Remove unwanted spaces between HTML tags and support Word documents

app/assets/javascripts/discourse/components/d-editor.js.es6

@@ -662,8 +662,6 @@ export default Ember.Component.extend({
       if (table) {
         this.appEvents.trigger('composer:insert-text', table);
         handled = true;
-      } else if (html && html.includes("urn:schemas-microsoft-com:office:word")) {
-        html = ""; // use plain text data for microsoft word
       }
     }
 

app/assets/javascripts/discourse/lib/to-markdown.js.es6

@@ -38,15 +38,15 @@ class Tag {
   }
 
   static emphases() {
-    return  [ ["b", "**"], ["strong", "**"], ["i", "_"], ["em", "_"], ["s", "~~"], ["strike", "~~"] ];
+    return  [ ["b", "**"], ["strong", "**"], ["i", "*"], ["em", "*"], ["s", "~~"], ["strike", "~~"] ];
   }
 
   static slices() {
-    return ["dt", "dd", "tr", "thead", "tbody", "tfoot"];
+    return ["dt", "dd", "thead", "tbody", "tfoot"];
   }
 
   static trimmable() {
-    return [...Tag.blocks(), ...Tag.headings(), ...Tag.slices(), "li", "td", "th", "br", "hr", "blockquote", "table", "ol"];
+    return [...Tag.blocks(), ...Tag.headings(), ...Tag.slices(), "li", "td", "th", "br", "hr", "blockquote", "table", "ol", "tr"];
   }
 
   static block(name, prefix, suffix) {
@@ -73,14 +73,17 @@ class Tag {
       }
 
       decorate(text) {
-        text = text.trim();
-
         if (text.includes("\n")) {
           this.prefix = `<${this.name}>`;
           this.suffix = `</${this.name}>`;
         }
 
-        return super.decorate(text);
+        let space = text.match(/^\s/) || [""];
+        this.prefix = space[0] + this.prefix;
+        space = text.match(/\s$/) || [""];
+        this.suffix = this.suffix + space[0];
+
+        return super.decorate(text.trim());
       }
     };
   }
@@ -182,10 +185,6 @@ class Tag {
           throw "Unsupported format inside Markdown table cells";
         }
 
-        if (!this.element.next) {
-          this.suffix = "|";
-        }
-
         return this.decorate(text);
       }
     };
@@ -268,6 +267,17 @@ class Tag {
     };
   }
 
+  static tr() {
+    return class extends Tag.slice("tr", "|\n") {
+      decorate(text) {
+        if (!this.element.next) {
+          this.suffix = "|";
+        }
+        return `${text}${this.suffix}`;
+      }
+    };
+  }
+
 }
 
 const tags = [
@@ -278,7 +288,7 @@ const tags = [
   Tag.cell("td"), Tag.cell("th"),
   Tag.replace("br", "\n"), Tag.replace("hr", "\n---\n"), Tag.replace("head", ""),
   Tag.keep("ins"), Tag.keep("del"), Tag.keep("small"), Tag.keep("big"),
-  Tag.li(), Tag.link(), Tag.image(), Tag.code(), Tag.blockquote(), Tag.table(),, Tag.ol(),
+  Tag.li(), Tag.link(), Tag.image(), Tag.code(), Tag.blockquote(), Tag.table(), Tag.ol(), Tag.tr(),
 ];
 
 class Element {
@@ -375,6 +385,19 @@ class Element {
   }
 }
 
+function trimUnwantedSpaces(html) {
+  const body = html.match(/<body[^>]*>([\s\S]*?)<\/body>/);
+  html = body ? body[1] : html;
+  html = html.replace(/\r|\n|&nbsp;/g, " ");
+
+  let match;
+  while (match = html.match(/<[^\s>]+[^>]*>\s{2,}<[^\s>]+[^>]*>/)) {
+    html = html.replace(match[0], match[0].replace(/>\s{2,}</, "> <"));
+  }
+
+  return html;
+}
+
 function putPlaceholders(html) {
   const codeRegEx = /<code[^>]*>([\s\S]*?)<\/code>/gi;
   const origHtml = html;
@@ -390,7 +413,7 @@ function putPlaceholders(html) {
     match = codeRegEx.exec(origHtml);
   }
 
-  const elements = parseHTML(html);
+  const elements = parseHTML(trimUnwantedSpaces(html));
   return { elements, placeholders };
 }
 
@@ -406,7 +429,7 @@ export default function toMarkdown(html) {
     const { elements, placeholders } = putPlaceholders(html);
     let markdown = Element.parse(elements).trim();
     markdown = markdown.replace(/^<b>/, "").replace(/<\/b>$/, "").trim(); // fix for google doc copy paste
-    markdown = markdown.replace(/\r/g, "").replace(/\n \n/g, "\n\n").replace(/\n{3,}/g, "\n\n");
+    markdown = markdown.replace(/ +\n/g, "\n").replace(/\n \n/g, "\n\n").replace(/\n{3,}/g, "\n\n");
     return replacePlaceholders(markdown, placeholders);
   } catch(err) {
     return "";

test/javascripts/lib/to-markdown-test.js.es6

@@ -4,19 +4,21 @@ QUnit.module("lib:to-markdown");
 
 QUnit.test("converts styles between normal words", assert => {
   const html = `Line with <s>styles</s> <b><i>between</i></b> words.`;
-  const markdown = `Line with ~~styles~~ **_between_** words.`;
+  const markdown = `Line with ~~styles~~ ***between*** words.`;
   assert.equal(toMarkdown(html), markdown);
+
+  assert.equal(toMarkdown("A <b>bold </b>word"), "A **bold** word");
 });
 
 QUnit.test("converts inline nested styles", assert => {
   let html = `<em>Italicised line with <strong>some random</strong> <b>bold</b> words.</em>`;
-  let markdown = `_Italicised line with **some random** **bold** words._`;
+  let markdown = `*Italicised line with **some random** **bold** words.*`;
   assert.equal(toMarkdown(html), markdown);
 
   html = `<i class="fa">Italicised line
-   with <b title="strong">some
+   with <b title="strong">some<br>
    random</b> <s>bold</s> words.</i>`;
-  markdown = `<i>Italicised line\n with <b>some\n random</b> ~~bold~~ words.</i>`;
+  markdown = `<i>Italicised line with <b>some\nrandom</b> ~~bold~~ words.</i>`;
   assert.equal(toMarkdown(html), markdown);
 });
 
@@ -26,7 +28,7 @@ QUnit.test("converts a link", assert => {
   assert.equal(toMarkdown(html), markdown);
 
   html = `<a href="https://discourse.org">Disc\n\n\nour\n\nse</a>`;
-  markdown = `[Disc\nour\nse](https://discourse.org)`;
+  markdown = `[Disc our se](https://discourse.org)`;
   assert.equal(toMarkdown(html), markdown);
 });
 
@@ -82,7 +84,7 @@ QUnit.test("converts ul list tag", assert => {
     <li>Item 3</li>
   </ul>
   `;
-  const markdown = `* Item 1\n* Item 2\n\n  * Sub Item 1\n  * Sub Item 2\n\n    * Sub _Sub_ Item 1\n    * Sub **Sub** Item 2\n\n* Item 3`;
+  const markdown = `* Item 1\n* Item 2\n\n  * Sub Item 1\n  * Sub Item 2\n\n    * Sub *Sub* Item 1\n    * Sub **Sub** Item 2\n\n* Item 3`;
   assert.equal(toMarkdown(html), markdown);
 });
 
@@ -101,10 +103,12 @@ QUnit.test("converts table tags", assert => {
     <thead> <tr><th>Heading 1</th><th>Head 2</th></tr> </thead>
       <tbody>
         <tr><td>Lorem</td><td>ipsum</td></tr>
-        <tr><td><b>dolor</b></td> <td><i>sit amet</i></td></tr></tbody>
+        <tr><td><b>dolor</b></td> <td><i>sit amet</i></td> </tr>
+        
+        </tbody>
 </table>
   `;
-  const markdown = `Discourse Avenue\n\n**laboris**\n\n|Heading 1|Head 2|\n| --- | --- |\n|Lorem|ipsum|\n|**dolor**|_sit amet_|`;
+  const markdown = `Discourse Avenue\n\n**laboris**\n\n|Heading 1|Head 2|\n| --- | --- |\n|Lorem|ipsum|\n|**dolor**|*sit amet*|`;
   assert.equal(toMarkdown(html), markdown);
 });
 
@@ -164,11 +168,11 @@ QUnit.test("supporting html tags by keeping them", assert => {
   output = `[Lorem <del>ipsum dolor</del> sit](http://example.com).`;
   assert.equal(toMarkdown(html), output);
 
-  html = `Lorem <del>ipsum \n\n dolor</del> sit.`;
+  html = `Lorem <del>ipsum dolor</del> sit.`;
   assert.equal(toMarkdown(html), html);
 
   html = `Lorem <a href="http://example.com"><del>ipsum \n\n\n dolor</del> sit.</a>`;
-  output = `Lorem [<del>ipsum \n dolor</del> sit.](http://example.com)`;
+  output = `Lorem [<del>ipsum dolor</del> sit.](http://example.com)`;
   assert.equal(toMarkdown(html), output);
 });
 
@@ -223,6 +227,6 @@ QUnit.test("converts ol list tag", assert => {
     <li>Item 3</li>
   </ol>
   `;
-  const markdown = `Testing\n\n1. Item 1\n2. Item 2\n\n  100. Sub Item 1\n  101. Sub Item 2\n\n    * Sub _Sub_ Item 1\n    * Sub **Sub** Item 2\n\n3. Item 3`;
+  const markdown = `Testing\n\n1. Item 1\n2. Item 2\n\n  100. Sub Item 1\n  101. Sub Item 2\n\n    * Sub *Sub* Item 1\n    * Sub **Sub** Item 2\n\n3. Item 3`;
   assert.equal(toMarkdown(html), markdown);
 });