Merge pull request #105 from lonekorean/more-separator

Preserve more separator HTML comment

package-lock.json

@@ -17,7 +17,7 @@
 				"luxon": "^3.4.4",
 				"request": "^2.88.2",
 				"request-promise-native": "^1.0.8",
-				"turndown": "^7.0.0",
+				"turndown": "^7.1.2",
 				"turndown-plugin-gfm": "^1.0.2",
 				"xml2js": "^0.4.23"
 			},

package.json

@@ -28,7 +28,7 @@
 		"luxon": "^3.4.4",
 		"request": "^2.88.2",
 		"request-promise-native": "^1.0.8",
-		"turndown": "^7.0.0",
+		"turndown": "^7.1.2",
 		"turndown-plugin-gfm": "^1.0.2",
 		"xml2js": "^0.4.23"
 	},

src/translator.js

@@ -61,7 +61,7 @@ function getPostContent(post, turndownService, config) {
 
 	// insert an empty div element between double line breaks
 	// this nifty trick causes turndown to keep adjacent paragraphs separated
-	// without mucking up content inside of other elemnts (like <code> blocks)
+	// without mucking up content inside of other elements (like <code> blocks)
 	content = content.replace(/(\r?\n){2}/g, '\n<div></div>\n');
 
 	if (config.saveScrapedImages) {
@@ -75,6 +75,10 @@ function getPostContent(post, turndownService, config) {
 	// (using turndown's blankRule() and keep() solution did not work for me)
 	content = content.replace(/(<\/iframe>)/gi, '.$1');
 
+	// preserve "more" separator, max one per post, optionally with custom label
+	// by escaping angle brackets (will be unescaped during turndown conversion)
+	content = content.replace(/<(!--more( .*)?--)>/, '&lt;$1&gt;');
+
 	// use turndown to convert HTML to Markdown
 	content = turndownService.turndown(content);