enh(llvm) Improve grammar and align assembly like grammars

src/languages/llvm.js

@@ -7,7 +7,54 @@ Category: assembler
 */
 
 export default function(hljs) {
-  var identifier = '([-a-zA-Z$._][\\w\\-$.]*)';
+  const IDENT_RE = '([-a-zA-Z$._][\\w\\-$.]*)';
+  const TYPE = {
+    className: 'type',
+    begin: /\bi\d+(?=\s|\b)/
+  };
+  const OPERATOR = {
+    className: 'operator',
+    relevance: 0,
+    begin: /=/
+  };
+  const PUNCTUATION = {
+    className: 'punctuation',
+    relevance: 0,
+    begin: /,/
+  };
+  const NUMBER = {
+    className: 'number',
+    variants: [
+        { begin: '0[xX][a-fA-F0-9]+' },
+        { begin: '-?\\d+(?:[.]\\d+)?(?:[eE][-+]?\\d+(?:[.]\\d+)?)?' }
+    ],
+    relevance: 0
+  };
+  const LABEL = {
+    className: 'symbol',
+    variants: [
+        { begin: /^\s*[a-z]+:/ }, // labels
+    ],
+    relevance: 0
+  };
+  const VARIABLE = {
+    className: 'variable',
+    variants: [
+      { begin: '%' + IDENT_RE },
+      { begin: '%\\d+' },
+      { begin: '#\\d+' },
+    ]
+  };
+  const FUNCTION = {
+    className: 'title',
+    variants: [
+      { begin: '@' + IDENT_RE },
+      { begin: '@\\d+' },
+      { begin: '!' + IDENT_RE },
+      { begin: '!\\d+' + IDENT_RE }
+    ]
+  };
+
   return {
     name: 'LLVM IR',
     keywords:
@@ -49,48 +96,28 @@ export default function(hljs) {
       'extractvalue insertvalue atomicrmw cmpxchg fence ' +
       'argmemonly double',
     contains: [
-      {
-        className: 'keyword',
-        begin: 'i\\d+'
-      },
-      hljs.COMMENT(
-        ';', '\\n', {relevance: 0}
+      TYPE,
+      hljs.COMMENT(/;\s*$/,
+        // this matches "empty comments"...
+        // ...because it's far more likely this is a statement terminator in
+        // another language than an actual comment
+        { relevance: 0 }
       ),
-      // Double quote string
+      hljs.COMMENT(/;/, /$/),
       hljs.QUOTE_STRING_MODE,
       {
         className: 'string',
         variants: [
           // Double-quoted string
           { begin: '"', end: '[^\\\\]"' },
-        ],
-        relevance: 0
-      },
-      {
-        className: 'title',
-        variants: [
-          { begin: '@' + identifier },
-          { begin: '@\\d+' },
-          { begin: '!' + identifier },
-          { begin: '!\\d+' + identifier }
         ]
       },
-      {
-        className: 'symbol',
-        variants: [
-          { begin: '%' + identifier },
-          { begin: '%\\d+' },
-          { begin: '#\\d+' },
-        ]
-      },
-      {
-        className: 'number',
-        variants: [
-            { begin: '0[xX][a-fA-F0-9]+' },
-            { begin: '-?\\d+(?:[.]\\d+)?(?:[eE][-+]?\\d+(?:[.]\\d+)?)?' }
-        ],
-        relevance: 0
-      },
+      FUNCTION,
+      PUNCTUATION,
+      OPERATOR,
+      VARIABLE,
+      LABEL,
+      NUMBER
     ]
   };
 }

test/markup/llvm/simple.expect.txt

@@ -0,0 +1,8 @@
+<span class="hljs-comment">;; foooo</span>
+<span class="hljs-keyword">define</span> <span class="hljs-type">i32</span> <span class="hljs-title">@mul_add</span>(<span class="hljs-type">i32</span> <span class="hljs-variable">%x</span><span class="hljs-punctuation">,</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%y</span><span class="hljs-punctuation">,</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%z</span>) {
+<span class="hljs-symbol">  entry:</span>
+    <span class="hljs-variable">%tmp</span> <span class="hljs-operator">=</span> <span class="hljs-keyword">mul</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%x</span><span class="hljs-punctuation">,</span> <span class="hljs-variable">%y</span>
+    <span class="hljs-variable">%tmp2</span> <span class="hljs-operator">=</span> <span class="hljs-keyword">add</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%tmp</span><span class="hljs-punctuation">,</span> <span class="hljs-variable">%z</span>
+    <span class="hljs-variable">%tmp3</span> <span class="hljs-operator">=</span> <span class="hljs-keyword">add</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%tmp</span><span class="hljs-punctuation">,</span> <span class="hljs-number">0</span>
+    <span class="hljs-keyword">ret</span> <span class="hljs-type">i32</span> <span class="hljs-variable">%tmp3</span>
+}

test/markup/llvm/simple.txt

@@ -0,0 +1,8 @@
+;; foooo
+define i32 @mul_add(i32 %x, i32 %y, i32 %z) {
+  entry:
+    %tmp = mul i32 %x, %y
+    %tmp2 = add i32 %tmp, %z
+    %tmp3 = add i32 %tmp, 0
+    ret i32 %tmp3
+}