cpp: Fix highlighting of unterminated raw strings

PR highlightjs#1897 switched C++ raw strings to use backreferences, however this
breaks souce files where raw strings are truncated. Like comments, it
would be preferable to highlight them.

Instead, go back to using separate begin and end regexps, but introduce
an endFilter feature to filter out false positive matches. This
internally works similarly to endSameAsBegin.

See also issue highlightjs#2259.

Author: davidben

docs/reference.rst

@@ -190,7 +190,7 @@ endSameAsBegin
 Acts as ``end`` matching exactly the same string that was found by the
 corresponding ``begin`` regexp.
 
-For example, in PostgreSQL string constants can uee "dollar quotes",
+For example, in PostgreSQL string constants can use "dollar quotes",
 consisting of a dollar sign, an optional tag of zero or more characters,
 and another dollar sign. String constant must be ended with the same
 construct using the same tag. It is possible to nest dollar-quoted string
@@ -208,6 +208,26 @@ In this case you can't simply specify the same regexp for ``begin`` and
 ``end`` (say, ``"\\$[a-z]\\$"``), but you can use ``begin: "\\$[a-z]\\$"``
 and ``endSameAsBegin: true``.
 
+.. _endFilter:
+
+endFilter
+^^^^^^^^^
+
+**type**: function
+
+Filters ``end`` matches to implement end rules that cannot be expressed as a
+standalone regular expression.
+
+This should be a function which takes two string parameters, the string that
+matched the ``begin`` regexp and the string that matched the ``end`` regexp. It
+should return true to end the mode and false otherwise.
+
+For example, C++11 raw string constants use syntax like ``R"tag(.....)tag"``,
+where ``tag`` is any zero to sixteen character string that must be repeated at
+the end. This could be matched with a single regexp containing backreferences,
+but truncated raw strings would not highlight. Instead, ``endFilter`` can be
+used to reject ``)tag"`` delimiters which do not match the starting value.
+
 .. _lexemes:
 
 lexemes

src/highlight.js

@@ -119,15 +119,19 @@ const HLJS = function(hljs) {
   function _highlight(languageName, code, ignore_illegals, continuation) {
     var codeToHighlight = code;
 
-    function endOfMode(mode, lexeme) {
-      if (regex.startsWith(mode.endRe, lexeme)) {
+    function endOfMode(mode, matchPlusRemainder, lexeme) {
+      var modeEnded = regex.startsWith(mode.endRe, matchPlusRemainder);
+      if (modeEnded && mode.endFilter) {
+        modeEnded = mode.endFilter(mode.beginValue, lexeme);
+      }
+      if (modeEnded) {
         while (mode.endsParent && mode.parent) {
           mode = mode.parent;
         }
         return mode;
       }
       if (mode.endsWithParent) {
-        return endOfMode(mode.parent, lexeme);
+        return endOfMode(mode.parent, matchPlusRemainder, lexeme);
       }
     }
 
@@ -206,7 +210,7 @@ const HLJS = function(hljs) {
       if (mode.className) {
         emitter.openNode(mode.className)
       }
-      top = Object.create(mode, {parent: {value: top}});
+      top = Object.create(mode, {parent: {value: top}, beginValue: {value: lexeme}});
     }
 
     function doIgnore(lexeme) {
@@ -255,7 +259,7 @@ const HLJS = function(hljs) {
     function doEndMatch(match) {
       var lexeme = match[0];
       var matchPlusRemainder = codeToHighlight.substr(match.index);
-      var end_mode = endOfMode(top, matchPlusRemainder);
+      var end_mode = endOfMode(top, matchPlusRemainder, lexeme);
       if (!end_mode) { return; }
 
       var origin = top;

src/languages/c-like.js

@@ -44,7 +44,16 @@ export default function(hljs) {
         begin: '(u8?|U|L)?\'(' + CHARACTER_ESCAPES + "|.)", end: '\'',
         illegal: '.'
       },
-      { begin: /(?:u8?|U|L)?R"([^()\\ ]{0,16})\((?:.|\n)*?\)\1"/ }
+      {
+        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;
+        },
+      }
     ]
   };
 

test/markup/cpp/truncated-block-comment.expect.txt

@@ -0,0 +1,3 @@
+<span class="hljs-comment">/*
+Truncated block comment
+</span>

test/markup/cpp/truncated-block-comment.txt

@@ -0,0 +1,2 @@
+/*
+Truncated block comment

test/markup/cpp/truncated-raw-string.expect.txt

@@ -0,0 +1,5 @@
+<span class="hljs-string">R"foo(
+Truncated raw string
+)nope"
+Still not completed.
+</span>

test/markup/cpp/truncated-raw-string.txt

@@ -0,0 +1,4 @@
+R"foo(
+Truncated raw string
+)nope"
+Still not completed.