enh(fortran) support intrinsic data types & better 0 width match error detection

CHANGES.md

@@ -14,6 +14,7 @@ Core Changes:
 
 Language Improvements:
 
+- (fortran) enh(fortran) support intrinsic data types (#2379) [Josh Goebel][]
 - enh(java) annotations can include numbers (#2377) [Josh Goebel][]
 - enh(java) annotations can take params (#2377) [Josh Goebel][]
 - enh(java) allow annotations inside function call params (#2377) [Josh Goebel][]

src/highlight.js

@@ -694,6 +694,12 @@ https://highlightjs.org/
       if (lastMatch.type=="begin" && match.type=="end" && lastMatch.index == match.index && lexeme === "") {
         // spit the "skipped" character that our regex choked on back into the output sequence
         mode_buffer += codeToHighlight.slice(match.index, match.index + 1);
+        if (!SAFE_MODE) {
+          var err = new Error('0 width match regex');
+          err.languageName = languageName;
+          err.badRule = lastMatch.rule;
+          throw(err);
+        }
         return 1;
       }
       lastMatch = match;
@@ -1018,6 +1024,7 @@ https://highlightjs.org/
   hljs.inherit = inherit;
   hljs.addPlugin = addPlugin;
   hljs.debugMode = function() { SAFE_MODE = false; }
+  hljs.safeMode = function() { SAFE_MODE = true; }
 
   // Common regexps
   hljs.IDENT_RE = '[a-zA-Z]\\w*';

src/languages/fortran.js

@@ -69,7 +69,8 @@ function(hljs) {
       hljs.COMMENT('!', '$', {relevance: 0}),
       {
         className: 'number',
-        begin: '(?=\\b|\\+|\\-|\\.)(?=\\.\\d|\\d)(?:\\d+)?(?:\\.?\\d*)(?:[de][+-]?\\d+)?\\b\\.?',
+        // regex in both fortran and irpf90 should match
+        begin: '(?=\\b|\\+|\\-|\\.)(?:\\.|\\d+\\.?)\\d*([de][+-]?\\d+)?(_[a-z_\\d]+)?',
         relevance: 0
       }
     ]

src/languages/irpf90.js

@@ -75,7 +75,8 @@ function(hljs) {
       hljs.COMMENT('begin_doc', 'end_doc', {relevance: 10}),
       {
         className: 'number',
-        begin: '(?=\\b|\\+|\\-|\\.)(?=\\.\\d|\\d)(?:\\d+)?(?:\\.?\\d*)(?:[de][+-]?\\d+)?\\b\\.?',
+        // regex in both fortran and irpf90 should match
+        begin: '(?=\\b|\\+|\\-|\\.)(?:\\.|\\d+\\.?)\\d*([de][+-]?\\d+)?(_[a-z_\\d]+)?',
         relevance: 0
       }
     ]

test/detect/index.js

@@ -5,6 +5,7 @@ delete require.cache[require.resolve('../../build/lib/highlight')]
 
 const fs       = require('fs').promises;
 const hljs     = require('../../build');
+hljs.debugMode(); // tests run in debug mode so errors are raised
 const path     = require('path');
 const utility  = require('../utility');
 

test/index.js

@@ -1,5 +1,8 @@
 'use strict';
 
+const hljs     = require('../build');
+hljs.debugMode(); // tests run in debug mode so errors are raised
+
 // Tests specific to the API exposed inside the hljs object.
 // Right now, that only includes tests for several common regular expressions.
 require('./api');

test/markup/fortran/numbers.expect.txt

@@ -14,6 +14,6 @@ var1
 va1r
 mo_tot_8 = <span class="hljs-number">1.</span>/(<span class="hljs-number">0.4</span>*<span class="hljs-built_in">log</span>(<span class="hljs-built_in">float</span>(elec_num_tot_8+<span class="hljs-number">0.4</span>)))
 
-<span class="hljs-number">6</span>_ikind
-<span class="hljs-number">1</span>_c_short
-<span class="hljs-number">6.</span><span class="hljs-number">6</span>66666666666666_DBL
+<span class="hljs-number">6_ikind</span>
+<span class="hljs-number">1_c_short</span>
+<span class="hljs-number">6.666666666666666_DBL</span>

test/parser/should-not-destroyData.js

@@ -5,6 +5,7 @@ describe("bugs", function () {
   // CONTEXT: https://github.com/highlightjs/highlight.js/pull/2219
   describe("a grammar with a mode that makes a 0 width match", () => {
     it("should instead count it as a 1 character match", () => {
+      hljs.safeMode();
       hljs.registerLanguage('test-language', (hljs) => {
 
         // broken regex from old Fortran ruleset
@@ -28,6 +29,12 @@ describe("bugs", function () {
         // Incorrect prior output:
         // 'The number is <span class="hljs-number"></span>23_longint yes.'
       )
+      hljs.debugMode();
+      should(() => {
+        hljs.highlight('test-language', 'The number is 123_longint yes.').value
+       }).throw(Error, {
+         message: "0 width match regex",
+         languageName: "test-language"})
     })
   })
 })