fix(clojure) Several issues with Clojure highlighting (#3397)

- fix(clojure) `$` in symbol breaks highlighting
- fix(clojure) Add complete regex for number detection
- enh(clojure) Add character mode for character literals
- fix(clojure) Inconsistent namespaced map highlighting
- enh(clojure) Add `regex` mode to regex literal
- fix(clojure) Remove inconsistent/broken highlighting for metadata
- enh(clojure) Add `punctuation` mode for commas.

Author: MrEbbinghaus

CHANGES.md

@@ -5,7 +5,15 @@ Grammars:
 - fix(python) Fix recognition of numeric literals followed by keywords without whitespace (#2985) [Richard Gibson][]
 - enh(swift) add SE-0290 unavailability condition (#3382) [Bradley Mackey][]
 - enh(java) add `sealed` and `non-sealed` keywords (#3386) [Bradley Mackey][]
-- fix(clojure) `comment` macro catches more than it should [Björn Ebbinghaus][]
+- fix(clojure) Several issues with Clojure highlighting (#3397) [Björn Ebbinghaus][]
+  - fix(clojure) `comment` macro catches more than it should (#3395)
+  - fix(clojure) `$` in symbol breaks highlighting
+  - fix(clojure) Add complete regex for number detection
+  - enh(clojure) Add character mode for character literals
+  - fix(clojure) Inconsistent namespaced map highlighting
+  - enh(clojure) Add `regex` mode to regex literal
+  - fix(clojure) Remove inconsistent/broken highlighting for metadata
+  - enh(clojure) Add `punctuation` mode for commas.
 
 [Richard Gibson]: https://github.com/gibson042
 [Bradley Mackey]: https://github.com/bradleymackey

src/languages/clojure.js

@@ -8,8 +8,8 @@ Category: lisp
 
 /** @type LanguageFn */
 export default function(hljs) {
-  const SYMBOLSTART = 'a-zA-Z_\\-!.?+*=<>&#\'';
-  const SYMBOL_RE = '[' + SYMBOLSTART + '][' + SYMBOLSTART + '0-9/;:]*';
+  const SYMBOLSTART = 'a-zA-Z_\\-!.?+*=<>&\'';
+  const SYMBOL_RE = '[#]?[' + SYMBOLSTART + '][' + SYMBOLSTART + '0-9/;:$#]*';
   const globals = 'def defonce defprotocol defstruct defmulti defmethod defn- defn defmacro deftype defrecord';
   const keywords = {
     $pattern: SYMBOL_RE,
@@ -45,20 +45,44 @@ export default function(hljs) {
       'lazy-seq spread list* str find-keyword keyword symbol gensym force rationalize'
   };
 
-  const SIMPLE_NUMBER_RE = '[-+]?\\d+(\\.\\d+)?';
-
   const SYMBOL = {
     begin: SYMBOL_RE,
     relevance: 0
   };
   const NUMBER = {
-    className: 'number',
-    begin: SIMPLE_NUMBER_RE,
-    relevance: 0
+    scope: 'number',
+    relevance: 0,
+    variants: [
+      {match: /[-+]?0[xX][0-9a-fA-F]+N?/},                           // hexadecimal                 // 0x2a
+      {match: /[-+]?0[0-7]+N?/},                                     // octal                       // 052
+      {match: /[-+]?[1-9][0-9]?[rR][0-9a-zA-Z]+N?/},                 // variable radix from 2 to 36 // 2r101010, 8r52, 36r16
+      {match: /[-+]?[0-9]+\/[0-9]+N?/},                              // ratio                       // 1/2
+      {match: /[-+]?[0-9]+((\.[0-9]*([eE][+-]?[0-9]+)?M?)|([eE][+-]?[0-9]+M?|M))/}, // float        // 0.42 4.2E-1M 42E1 42M
+      {match: /[-+]?([1-9][0-9]*|0)N?/},                             // int (don't match leading 0) // 42 42N
+    ]
   };
+  const CHARACTER = {
+    scope: 'character',
+    variants: [
+      {match: /\\o[0-3]?[0-7]{1,2}/},                             // Unicode Octal 0 - 377
+      {match: /\\u[0-9a-fA-F]{4}/},                               // Unicode Hex 0000 - FFFF
+      {match: /\\(newline|space|tab|formfeed|backspace|return)/}, // special characters
+      {match: /\\\S/, relevance: 0}                               // any non-whitespace char
+    ]
+  }
+  const REGEX = {
+    scope: 'regex',
+    begin: /#"/,
+    end: /"/
+  }
   const STRING = hljs.inherit(hljs.QUOTE_STRING_MODE, {
     illegal: null
   });
+  const COMMA = {
+    scope: 'punctuation',
+    match: /,/,
+    relevance: 0
+  }
   const COMMENT = hljs.COMMENT(
     ';',
     '$',
@@ -71,15 +95,10 @@ export default function(hljs) {
     begin: /\b(true|false|nil)\b/
   };
   const COLLECTION = {
-    begin: '[\\[\\{]',
+    begin: "\\[|(#::?" + SYMBOL_RE + ")?\\{",
     end: '[\\]\\}]',
     relevance: 0
   };
-  const HINT = {
-    className: 'comment',
-    begin: '\\^' + SYMBOL_RE
-  };
-  const HINT_COL = hljs.COMMENT('\\^\\{', '\\}');
   const KEY = {
     className: 'symbol',
     begin: '[:]{1,2}' + SYMBOL_RE
@@ -100,10 +119,11 @@ export default function(hljs) {
     starts: BODY
   };
   const DEFAULT_CONTAINS = [
+    COMMA,
     LIST,
+    CHARACTER,
+    REGEX,
     STRING,
-    HINT,
-    HINT_COL,
     COMMENT,
     KEY,
     COLLECTION,
@@ -138,17 +158,17 @@ export default function(hljs) {
   ];
   BODY.contains = DEFAULT_CONTAINS;
   COLLECTION.contains = DEFAULT_CONTAINS;
-  HINT_COL.contains = [ COLLECTION ];
 
   return {
     name: 'Clojure',
     aliases: [ 'clj', 'edn' ],
     illegal: /\S/,
     contains: [
+      COMMA,
       LIST,
+      CHARACTER,
+      REGEX,
       STRING,
-      HINT,
-      HINT_COL,
       COMMENT,
       KEY,
       COLLECTION,

test/markup/clojure/character.expect.txt

@@ -0,0 +1,5 @@
+<span class="hljs-character">\A</span>
+<span class="hljs-character">\a</span>
+<span class="hljs-character">\formfeed</span>
+<span class="hljs-character">\u00DF</span>
+<span class="hljs-character">\o303</span>

test/markup/clojure/character.txt

@@ -0,0 +1,5 @@
+\A
+\a
+\formfeed
+\u00DF
+\o303

test/markup/clojure/deps_edn.expect.txt

@@ -1,14 +1,14 @@
-{<span class="hljs-symbol">:aliases</span> {<span class="hljs-symbol">:export</span> {<span class="hljs-symbol">:exec-fn</span> stelcodes.dev-blog.generator/export},
-           <span class="hljs-symbol">:repl</span> {<span class="hljs-symbol">:extra-deps</span> {cider/cider-nrepl {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;0.25.2&quot;</span>},
-                               nrepl/nrepl {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;0.8.3&quot;</span>}},
-                  <span class="hljs-symbol">:extra-paths</span> [<span class="hljs-string">&quot;dev&quot;</span>],
+{<span class="hljs-symbol">:aliases</span> {<span class="hljs-symbol">:export</span> {<span class="hljs-symbol">:exec-fn</span> stelcodes.dev-blog.generator/export}<span class="hljs-punctuation">,</span>
+           <span class="hljs-symbol">:repl</span> {<span class="hljs-symbol">:extra-deps</span> {cider/cider-nrepl {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;0.25.2&quot;</span>}<span class="hljs-punctuation">,</span>
+                               nrepl/nrepl {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;0.8.3&quot;</span>}}<span class="hljs-punctuation">,</span>
+                  <span class="hljs-symbol">:extra-paths</span> [<span class="hljs-string">&quot;dev&quot;</span>]<span class="hljs-punctuation">,</span>
                   <span class="hljs-symbol">:main-opts</span> [<span class="hljs-string">&quot;-m&quot;</span>
                               <span class="hljs-string">&quot;nrepl.cmdline&quot;</span>
                               <span class="hljs-string">&quot;--middleware&quot;</span>
                               <span class="hljs-string">&quot;[cider.nrepl/cider-middleware]&quot;</span>
-                              <span class="hljs-string">&quot;--interactive&quot;</span>]},
-           <span class="hljs-symbol">:webhook</span> {<span class="hljs-symbol">:exec-fn</span> stelcodes.dev-blog.webhook/listen}},
- <span class="hljs-symbol">:deps</span> {http-kit/http-kit {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;2.5.3&quot;</span>},
-        org.clojure/clojure {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;1.10.1&quot;</span>},
-        stasis/stasis {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;2.5.1&quot;</span>}},
+                              <span class="hljs-string">&quot;--interactive&quot;</span>]}<span class="hljs-punctuation">,</span>
+           <span class="hljs-symbol">:webhook</span> {<span class="hljs-symbol">:exec-fn</span> stelcodes.dev-blog.webhook/listen}}<span class="hljs-punctuation">,</span>
+ <span class="hljs-symbol">:deps</span> {http-kit/http-kit {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;2.5.3&quot;</span>}<span class="hljs-punctuation">,</span>
+        org.clojure/clojure {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;1.10.1&quot;</span>}<span class="hljs-punctuation">,</span>
+        stasis/stasis {<span class="hljs-symbol">:mvn/version</span> <span class="hljs-string">&quot;2.5.1&quot;</span>}}<span class="hljs-punctuation">,</span>
  <span class="hljs-symbol">:paths</span> [<span class="hljs-string">&quot;src&quot;</span> <span class="hljs-string">&quot;resources&quot;</span>]}

test/markup/clojure/globals_definition.expect.txt

@@ -5,21 +5,26 @@
 <span class="hljs-comment">; function</span>
 (<span class="hljs-keyword">defn</span> <span class="hljs-title">clojure-function</span> [args]
   (<span class="hljs-name"><span class="hljs-built_in">let</span></span> [string   <span class="hljs-string">&quot;multiline\nstring&quot;</span>
-        regexp   #<span class="hljs-string">&quot;regexp&quot;</span>
-        number   <span class="hljs-number">100</span>,<span class="hljs-number">000</span>
+        regexp   <span class="hljs-regex">#&quot;regexp&quot;</span>
+        number   <span class="hljs-number">100000</span>
         booleans [<span class="hljs-literal">false</span> <span class="hljs-literal">true</span>]
         keyword  <span class="hljs-symbol">::the-keyword</span>]
     <span class="hljs-comment">;; this is comment</span>
     (<span class="hljs-name"><span class="hljs-built_in">if</span></span> <span class="hljs-literal">true</span>
       (<span class="hljs-name"><span class="hljs-built_in">-&gt;&gt;</span></span>
         (<span class="hljs-name"><span class="hljs-built_in">list</span></span> [vector] {<span class="hljs-symbol">:map</span> map} #{&#x27;set})))))
 
+#:person{<span class="hljs-symbol">:first</span> <span class="hljs-string">&quot;Han&quot;</span>
+         <span class="hljs-symbol">:last</span> <span class="hljs-string">&quot;Solo&quot;</span>
+         <span class="hljs-symbol">:ship</span> #:ship{<span class="hljs-symbol">:name</span> <span class="hljs-string">&quot;Millenium Falcon&quot;</span>}}
+#::{<span class="hljs-symbol">:a</span> <span class="hljs-number">1</span><span class="hljs-punctuation">,</span> <span class="hljs-symbol">:b</span> <span class="hljs-number">2</span>}
+
 <span class="hljs-comment">; global</span>
 (<span class="hljs-keyword">def</span> <span class="hljs-title">some-var</span>)
 <span class="hljs-comment">; another one</span>
 (<span class="hljs-keyword">def</span> <span class="hljs-title">alternative-var</span> <span class="hljs-string">&quot;132&quot;</span>)
 <span class="hljs-comment">; defonce</span>
-(<span class="hljs-keyword">defonce</span> ^<span class="hljs-symbol">:private</span> <span class="hljs-title">another-var</span> #<span class="hljs-string">&quot;foo&quot;</span>)
+(<span class="hljs-keyword">defonce</span> ^<span class="hljs-symbol">:private</span> <span class="hljs-title">another-var</span> <span class="hljs-regex">#&quot;foo&quot;</span>)
 
 <span class="hljs-comment">; private function</span>
 (<span class="hljs-keyword">defn-</span> <span class="hljs-title">add</span> [x y] (<span class="hljs-name"><span class="hljs-built_in">+</span></span> x y))
@@ -58,4 +63,4 @@
 
 <span class="hljs-comment">;; create a couple shapes and get their area</span>
 (<span class="hljs-keyword">def</span> <span class="hljs-title">myCircle</span> (<span class="hljs-name">Circle.</span> <span class="hljs-number">10</span>))
-(<span class="hljs-keyword">def</span> <span class="hljs-title">mySquare</span> (<span class="hljs-name">Square.</span> <span class="hljs-number">5</span> <span class="hljs-number">11</span>))
+(<span class="hljs-keyword">def</span> <span class="hljs-title">mySquare</span> (<span class="hljs-name">Square.</span> <span class="hljs-number">5</span> <span class="hljs-number">11</span>))

test/markup/clojure/globals_definition.txt

@@ -6,14 +6,19 @@
 (defn clojure-function [args]
   (let [string   "multiline\nstring"
         regexp   #"regexp"
-        number   100,000
+        number   100000
         booleans [false true]
         keyword  ::the-keyword]
     ;; this is comment
     (if true
       (->>
         (list [vector] {:map map} #{'set})))))
 
+#:person{:first "Han"
+         :last "Solo"
+         :ship #:ship{:name "Millenium Falcon"}}
+#::{:a 1, :b 2}
+
 ; global
 (def some-var)
 ; another one

test/markup/clojure/hint_col.expect.txt

@@ -0,0 +1,69 @@
+<span class="hljs-comment">; integer</span>
+<span class="hljs-number">00</span>
+<span class="hljs-number">42</span>
+<span class="hljs-number">+42</span>
+<span class="hljs-number">-42</span>
+
+<span class="hljs-comment">; BigInt</span>
+<span class="hljs-number">42N</span>
+<span class="hljs-number">0N</span>
+<span class="hljs-number">+42N</span>
+<span class="hljs-number">-42N</span>
+
+<span class="hljs-comment">; octal</span>
+<span class="hljs-number">052</span>
+<span class="hljs-number">00N</span>
+<span class="hljs-number">+052</span>
+<span class="hljs-number">-00N</span>
+
+<span class="hljs-comment">; hex</span>
+<span class="hljs-number">0x2a</span>
+<span class="hljs-number">0x0N</span>
+<span class="hljs-number">+0x2a</span>
+<span class="hljs-number">-0x0N</span>
+
+<span class="hljs-comment">; radix</span>
+<span class="hljs-number">2r101010</span>
+<span class="hljs-number">8r52</span>
+<span class="hljs-number">16r2a</span>
+<span class="hljs-number">36r16</span>
+<span class="hljs-number">-2r101010</span>
+<span class="hljs-number">+36r16</span>
+
+<span class="hljs-comment">; radix BigInt</span>
+<span class="hljs-number">2r101010N</span>
+<span class="hljs-number">8r52N</span>
+<span class="hljs-number">16r2aN</span>
+<span class="hljs-number">36r16N</span>
+<span class="hljs-number">+8r52N</span>
+<span class="hljs-number">-16r2aN</span>
+
+<span class="hljs-comment">;; ratios</span>
+<span class="hljs-number">1/2</span>
+<span class="hljs-number">-1/2</span>
+<span class="hljs-number">+123/224</span>
+
+<span class="hljs-comment">;; floats</span>
+<span class="hljs-number">42.0</span>
+<span class="hljs-number">-42.0</span>
+<span class="hljs-number">+42.0</span>
+<span class="hljs-number">42.</span>
+<span class="hljs-number">+42.</span>
+<span class="hljs-number">-42.</span>
+
+<span class="hljs-comment">; BigDecimal</span>
+<span class="hljs-number">42.0M</span>
+<span class="hljs-number">-42M</span>
+<span class="hljs-number">42.M</span>
+<span class="hljs-number">42M</span>
+
+<span class="hljs-comment">; with Exponent</span>
+<span class="hljs-number">42.0E2</span>
+<span class="hljs-number">-42.0E+9</span>
+<span class="hljs-number">42E-0</span>
+<span class="hljs-number">+42E-0</span>
+
+<span class="hljs-number">42.0E2M</span>
+<span class="hljs-number">42E+9M</span>
+<span class="hljs-number">-42E+9M</span>
+<span class="hljs-number">+42.0E2M</span>