Markdown: Better language detection for code blocks (#2114)

This improves the language detection for code blocks by using only the first word instead of the whole expression and replacing some characters so languages like `c#` and `c++` are supported.

components/prism-markdown.js

@@ -280,7 +280,12 @@
 					typeof codeLang.content === 'string') {
 
 					// this might be a language that Prism does not support
-					var alias = 'language-' + codeLang.content.trim().split(/\s+/)[0].toLowerCase();
+
+					// do some replacements to support C++, C#, and F#
+					var lang = codeLang.content.replace(/\b#/g, 'sharp').replace(/\b\+\+/g, 'pp')
+					// only use the first word
+					lang = (/[a-z][\w-]*/i.exec(lang) || [''])[0].toLowerCase();
+					var alias = 'language-' + lang;
 
 					// add alias
 					if (!codeBlock.alias) {

tests/languages/markdown/code_block_language_detection_feature.js

@@ -0,0 +1,11 @@
+module.exports = {
+
+	'```js\nlet a = 0;\n```': '<span class="token code"><span class="token punctuation">```</span><span class="token code-language">js</span>\n<span class="token code-block language-js">let a = 0;</span>\n<span class="token punctuation">```</span></span>',
+
+	'``` c++\nint a = 0;\n```': '<span class="token code"><span class="token punctuation">```</span><span class="token code-language"> c++</span>\n<span class="token code-block language-cpp">int a = 0;</span>\n<span class="token punctuation">```</span></span>',
+
+	'``` c#\nvar a = 0;\n```': '<span class="token code"><span class="token punctuation">```</span><span class="token code-language"> c#</span>\n<span class="token code-block language-csharp">var a = 0;</span>\n<span class="token punctuation">```</span></span>',
+
+	'```{r pressure, echo=FALSE}\nplot(pressure)\n```': '<span class="token code"><span class="token punctuation">```</span><span class="token code-language">{r pressure, echo=FALSE}</span>\n<span class="token code-block language-r">plot(pressure)</span>\n<span class="token punctuation">```</span></span>'
+
+};