From 3a5d6ea6dbc0f1a6d401b60dac1ae191fd6ee2da Mon Sep 17 00:00:00 2001
From: Josh Goebel <me@joshgoebel.com>
Date: Mon, 11 Nov 2019 21:28:05 -0500
Subject: [PATCH] more generic implementation (though not perfect yet)

---
 src/highlight.js     | 62 +++++++++++++++++++++++++++++++++++---------
 src/languages/cpp.js | 12 +++------
 2 files changed, 54 insertions(+), 20 deletions(-)

diff --git a/src/highlight.js b/src/highlight.js
index bac0a5fbee..0dc394a1ca 100644
--- a/src/highlight.js
+++ b/src/highlight.js
@@ -108,7 +108,7 @@ https://highlightjs.org/
     }
   }
 
-  function inherit(parent) {  // inherit(parent, override_obj, override_obj, ...)
+  var inherit = function(parent) {  // inherit(parent, override_obj, override_obj, ...)
     var key;
     var result = {};
     var objects = Array.prototype.slice.call(arguments, 1);
@@ -122,6 +122,15 @@ https://highlightjs.org/
     return result;
   }
 
+  // ES2015 version (we would hope it's faster?)
+  if (Object.assign) {
+    inherit = function() {
+      Array.prototype.unshift.call(arguments, {})
+      return Object.assign.apply(null, arguments)
+    }
+  }
+
+
   /* Stream merging */
 
   function nodeStream(node) {
@@ -415,18 +424,32 @@ https://highlightjs.org/
         var match = matcherRe.exec(s);
         if (!match) { return null; }
 
-        for(var i = 0; i<match.length; i++) {
-          if (match[i] != undefined && matchIndexes["" +i] != undefined ) {
-            rule = matchIndexes[""+i];
+        var idx;
+        for(idx = 0; idx<match.length; idx++) {
+          if (match[idx] != undefined && matchIndexes["" +idx] != undefined ) {
+            rule = matchIndexes[""+idx];
+            // match.splice(0, i-1) // trim off the extra matches
+            match.splice(0, idx) // trim off the extra matches
             break;
           }
         }
 
+        // raw (but cooked slight, to fixed the indexes) match is passed to
+        // callbacks as `raw` and avoids exposing too much of our internal
+        // workings to callbacks
+
+        // we do this selectively with begin [performance], but for `end` because
+        // we could potentially be matching a parent end rule we always
+        // include a copy of the raw match
+
         // illegal or end match
         if (typeof rule === "string") {
+          match.raw = inherit(match)
           match.type = rule;
           match.extra = [mode.illegal, mode.terminator_end];
         } else {
+          if (hasBeginCallbacks(rule))
+            match.raw = inherit(match)
           match.type = "begin";
           match.rule = rule;
         }
@@ -436,6 +459,14 @@ https://highlightjs.org/
       return matcher;
     }
 
+    function hasBeginCallbacks(rule) {
+      return !!rule.onBegin;
+    }
+
+    function hasCallbacks(rule) {
+      return rule.onBegin || rule.endWhen;
+    }
+
     function compileMode(mode, parent) {
       if (mode.compiled)
         return;
@@ -501,11 +532,8 @@ https://highlightjs.org/
       return new RegExp(value.replace(/[-\/\\^$*+?.()|[\]{}]/g, '\\$&'), 'm');
     }
 
-    function endOfMode(mode, matchPlusRemainder, lexeme) {
+    function endOfMode(mode, matchPlusRemainder) {
       var modeEnded = testRe(mode.endRe, matchPlusRemainder);
-      if (modeEnded && mode.endFilter) {
-        modeEnded = mode.endFilter(mode.beginValue, lexeme);
-      }
       if (modeEnded) {
         while (mode.endsParent && mode.parent) {
           mode = mode.parent;
@@ -513,7 +541,7 @@ https://highlightjs.org/
         return mode;
       }
       if (mode.endsWithParent) {
-        return endOfMode(mode.parent, matchPlusRemainder, lexeme);
+        return endOfMode(mode.parent, matchPlusRemainder);
       }
     }
 
@@ -591,13 +619,15 @@ https://highlightjs.org/
 
     function startNewMode(mode, lexeme) {
       result += mode.className? buildSpan(mode.className, '', true): '';
-      top = Object.create(mode, {parent: {value: top}, beginValue: {value: lexeme}});
+      top = Object.create(mode, {parent: {value: top}, userState: {value: {}}})
+      return top;
     }
 
 
     function doBeginMatch(match) {
       var lexeme = match[0];
       var new_mode = match.rule;
+      var mode;
 
       if (new_mode && new_mode.endSameAsBegin) {
         new_mode.endRe = escapeRe( lexeme );
@@ -614,16 +644,24 @@ https://highlightjs.org/
           mode_buffer = lexeme;
         }
       }
-      startNewMode(new_mode, lexeme);
+      mode = startNewMode(new_mode, lexeme);
+      if (mode.onBegin) {
+        mode.onBegin(match.raw, mode.userState)
+      }
       return new_mode.returnBegin ? 0 : lexeme.length;
     }
 
     function doEndMatch(match) {
       var lexeme = match[0];
       var matchPlusRemainder = value.substr(match.index);
-      var end_mode = endOfMode(top, matchPlusRemainder, lexeme);
+      var end_mode = endOfMode(top, matchPlusRemainder);
       if (!end_mode) { return; }
 
+      if (end_mode.endWhen) {
+        var doEnd = end_mode.endWhen(match.raw, end_mode.userState)
+        if (doEnd === false) { return; }
+      }
+
       var origin = top;
       if (origin.skip) {
         mode_buffer += lexeme;
diff --git a/src/languages/cpp.js b/src/languages/cpp.js
index 26646d3854..8cf95f1537 100644
--- a/src/languages/cpp.js
+++ b/src/languages/cpp.js
@@ -28,14 +28,10 @@ function(hljs) {
         illegal: '.'
       },
       {
-        begin: /(?:u8?|U|L)?R"[^()\\ ]{0,16}\(/,
-        end: /\)[^()\\ ]{0,16}"/,
-        endFilter: function(begin, end) {
-          var quote = begin.indexOf('"');
-          var beginDelimiter = begin.substring(quote + 1, begin.length - 1);
-          var endDelimiter = end.substring(1, end.length - 1);
-          return beginDelimiter == endDelimiter;
-        },
+        begin: /(?:u8?|U|L)?R"([^()\\ ]{0,16})\(/,
+        end: /\)([^()\\ ]{0,16})"/,
+        onBegin: function(m, state) { state.heredoc = m[1] },
+        endWhen: function(m, state) { return state.heredoc === m[1] },
       }
     ]
   };