Update MD034/no-bare-urls to re-scan documents with potential violations using proper reference definition handling to avoid false positives (fixes #787).

demo/markdownlint-browser.js

@@ -1374,10 +1374,12 @@ var _require = __webpack_require__(/*! markdownlint-micromark */ "markdownlint-m
  *
  * @param {string} markdown Markdown document.
  * @param {Object} [options] Options for micromark.
+ * @param {boolean} [refsDefined] Whether to treat references as defined.
  * @returns {Object[]} Micromark events.
  */
 function getMicromarkEvents(markdown) {
   var options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
+  var refsDefined = arguments.length > 2 && arguments[2] !== undefined ? arguments[2] : true;
   // Customize options object to add useful extensions
   options.extensions = options.extensions || [];
   options.extensions.push(gfmAutolinkLiteral, gfmFootnote(), gfmTable);
@@ -1386,10 +1388,12 @@ function getMicromarkEvents(markdown) {
   var encoding = undefined;
   var eol = true;
   var parseContext = parse(options);
-  // Customize ParseContext to treat all references as defined
-  parseContext.defined.includes = function (searchElement) {
-    return searchElement.length > 0;
-  };
+  if (refsDefined) {
+    // Customize ParseContext to treat all references as defined
+    parseContext.defined.includes = function (searchElement) {
+      return searchElement.length > 0;
+    };
+  }
   var chunks = preprocess()(markdown, encoding, eol);
   var events = postprocess(parseContext.document().write(chunks));
   return events;
@@ -1400,12 +1404,14 @@ function getMicromarkEvents(markdown) {
  *
  * @param {string} markdown Markdown document.
  * @param {Object} [options] Options for micromark.
+ * @param {boolean} [refsDefined] Whether to treat references as defined.
  * @returns {Token[]} Micromark tokens (frozen).
  */
 function micromarkParse(markdown) {
   var options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
+  var refsDefined = arguments.length > 2 && arguments[2] !== undefined ? arguments[2] : true;
   // Use micromark to parse document into Events
-  var events = getMicromarkEvents(markdown, options);
+  var events = getMicromarkEvents(markdown, options, refsDefined);
 
   // Create Token objects
   var document = [];
@@ -5008,60 +5014,68 @@ var _require = __webpack_require__(/*! ../helpers */ "../helpers/helpers.js"),
   addErrorContext = _require.addErrorContext;
 var _require2 = __webpack_require__(/*! ../helpers/micromark.cjs */ "../helpers/micromark.cjs"),
   filterByPredicate = _require2.filterByPredicate,
-  getHtmlTagInfo = _require2.getHtmlTagInfo;
+  getHtmlTagInfo = _require2.getHtmlTagInfo,
+  parse = _require2.parse;
 module.exports = {
   "names": ["MD034", "no-bare-urls"],
   "description": "Bare URL used",
   "tags": ["links", "url"],
   "function": function MD034(params, onError) {
-    var literalAutolinks = filterByPredicate(params.parsers.micromark.tokens, function (token) {
-      return token.type === "literalAutolink";
-    }, function (token) {
-      var children = token.children;
-      var result = [];
-      for (var i = 0; i < children.length; i++) {
-        var openToken = children[i];
-        var openTagInfo = getHtmlTagInfo(openToken);
-        if (openTagInfo && !openTagInfo.close) {
-          var count = 1;
-          for (var j = i + 1; j < children.length; j++) {
-            var closeToken = children[j];
-            var closeTagInfo = getHtmlTagInfo(closeToken);
-            if (closeTagInfo && openTagInfo.name === closeTagInfo.name) {
-              if (closeTagInfo.close) {
-                count--;
-                if (count === 0) {
-                  i = j;
-                  break;
+    var literalAutolinks = function literalAutolinks(tokens) {
+      return filterByPredicate(tokens, function (token) {
+        return token.type === "literalAutolink";
+      }, function (token) {
+        var children = token.children;
+        var result = [];
+        for (var i = 0; i < children.length; i++) {
+          var openToken = children[i];
+          var openTagInfo = getHtmlTagInfo(openToken);
+          if (openTagInfo && !openTagInfo.close) {
+            var count = 1;
+            for (var j = i + 1; j < children.length; j++) {
+              var closeToken = children[j];
+              var closeTagInfo = getHtmlTagInfo(closeToken);
+              if (closeTagInfo && openTagInfo.name === closeTagInfo.name) {
+                if (closeTagInfo.close) {
+                  count--;
+                  if (count === 0) {
+                    i = j;
+                    break;
+                  }
+                } else {
+                  count++;
                 }
-              } else {
-                count++;
               }
             }
+          } else {
+            result.push(openToken);
           }
-        } else {
-          result.push(openToken);
         }
+        return result;
+      });
+    };
+    if (literalAutolinks(params.parsers.micromark.tokens).length > 0) {
+      // Re-parse with correct link/image reference definition handling
+      var document = params.lines.join("\n");
+      var tokens = parse(document, undefined, false);
+      var _iterator = _createForOfIteratorHelper(literalAutolinks(tokens)),
+        _step;
+      try {
+        for (_iterator.s(); !(_step = _iterator.n()).done;) {
+          var token = _step.value;
+          var range = [token.startColumn, token.endColumn - token.startColumn];
+          var fixInfo = {
+            "editColumn": range[0],
+            "deleteCount": range[1],
+            "insertText": "<".concat(token.text, ">")
+          };
+          addErrorContext(onError, token.startLine, token.text, null, null, range, fixInfo);
+        }
+      } catch (err) {
+        _iterator.e(err);
+      } finally {
+        _iterator.f();
       }
-      return result;
-    });
-    var _iterator = _createForOfIteratorHelper(literalAutolinks),
-      _step;
-    try {
-      for (_iterator.s(); !(_step = _iterator.n()).done;) {
-        var token = _step.value;
-        var range = [token.startColumn, token.endColumn - token.startColumn];
-        var fixInfo = {
-          "editColumn": range[0],
-          "deleteCount": range[1],
-          "insertText": "<".concat(token.text, ">")
-        };
-        addErrorContext(onError, token.startLine, token.text, null, null, range, fixInfo);
-      }
-    } catch (err) {
-      _iterator.e(err);
-    } finally {
-      _iterator.f();
     }
   }
 };

helpers/micromark.cjs

@@ -26,9 +26,10 @@ const {
  *
  * @param {string} markdown Markdown document.
  * @param {Object} [options] Options for micromark.
+ * @param {boolean} [refsDefined] Whether to treat references as defined.
  * @returns {Object[]} Micromark events.
  */
-function getMicromarkEvents(markdown, options = {}) {
+function getMicromarkEvents(markdown, options = {}, refsDefined = true) {
 
   // Customize options object to add useful extensions
   options.extensions = options.extensions || [];
@@ -38,8 +39,10 @@ function getMicromarkEvents(markdown, options = {}) {
   const encoding = undefined;
   const eol = true;
   const parseContext = parse(options);
-  // Customize ParseContext to treat all references as defined
-  parseContext.defined.includes = (searchElement) => searchElement.length > 0;
+  if (refsDefined) {
+    // Customize ParseContext to treat all references as defined
+    parseContext.defined.includes = (searchElement) => searchElement.length > 0;
+  }
   const chunks = preprocess()(markdown, encoding, eol);
   const events = postprocess(parseContext.document().write(chunks));
   return events;
@@ -50,12 +53,14 @@ function getMicromarkEvents(markdown, options = {}) {
  *
  * @param {string} markdown Markdown document.
  * @param {Object} [options] Options for micromark.
+ * @param {boolean} [refsDefined] Whether to treat references as defined.
  * @returns {Token[]} Micromark tokens (frozen).
  */
-function micromarkParse(markdown, options = {}) {
+function micromarkParse(markdown, options = {}, refsDefined = true) {
 
   // Use micromark to parse document into Events
-  const events = getMicromarkEvents(markdown, options);
+  const events =
+    getMicromarkEvents(markdown, options, refsDefined);
 
   // Create Token objects
   const document = [];

lib/md034.js

@@ -3,7 +3,7 @@
 "use strict";
 
 const { addErrorContext } = require("../helpers");
-const { filterByPredicate, getHtmlTagInfo } =
+const { filterByPredicate, getHtmlTagInfo, parse } =
   require("../helpers/micromark.cjs");
 
 module.exports = {
@@ -11,9 +11,9 @@ module.exports = {
   "description": "Bare URL used",
   "tags": [ "links", "url" ],
   "function": function MD034(params, onError) {
-    const literalAutolinks =
+    const literalAutolinks = (tokens) => (
       filterByPredicate(
-        params.parsers.micromark.tokens,
+        tokens,
         (token) => token.type === "literalAutolink",
         (token) => {
           const { children } = token;
@@ -43,26 +43,33 @@ module.exports = {
             }
           }
           return result;
-        });
-    for (const token of literalAutolinks) {
-      const range = [
-        token.startColumn,
-        token.endColumn - token.startColumn
-      ];
-      const fixInfo = {
-        "editColumn": range[0],
-        "deleteCount": range[1],
-        "insertText": `<${token.text}>`
-      };
-      addErrorContext(
-        onError,
-        token.startLine,
-        token.text,
-        null,
-        null,
-        range,
-        fixInfo
-      );
+        }
+      )
+    );
+    if (literalAutolinks(params.parsers.micromark.tokens).length > 0) {
+      // Re-parse with correct link/image reference definition handling
+      const document = params.lines.join("\n");
+      const tokens = parse(document, undefined, false);
+      for (const token of literalAutolinks(tokens)) {
+        const range = [
+          token.startColumn,
+          token.endColumn - token.startColumn
+        ];
+        const fixInfo = {
+          "editColumn": range[0],
+          "deleteCount": range[1],
+          "insertText": `<${token.text}>`
+        };
+        addErrorContext(
+          onError,
+          token.startLine,
+          token.text,
+          null,
+          null,
+          range,
+          fixInfo
+        );
+      }
     }
   }
 };

test/bare-urls.md

@@ -83,8 +83,7 @@ Angle brackets work the same for email: <user@example.com>
 
 Links bind to the innermost [link that [is-a-valid] link](https://example.com) {MD034}
 
-But not if the [link [is-not-a-valid] link](https://example.com) {MD034}
-HOWEVER this scenario could have an invalid shortcut and IS reported
+But not if the [link [is-not-a-valid] link](https://example.com)
 
-Escaping both inner square brackets avoids the unwanted report:
+Escaping both inner square brackets avoids confusion:
 [link \[is-not-a-valid\] link](https://example.com)

test/links.md

@@ -25,3 +25,8 @@ Duplicate links in tables should be handled:
 | Link                 | Same Link            | Violation |
 |----------------------|----------------------|-----------|
 | https://example.com/ | https://example.com/ | {MD034}   |
+
+This is not a bare URL: [text [undefined] text](https://example.com).
+This is a bare URL: [text [defined] text](https://example.com). {MD034}
+
+[defined]: https://example.com

test/snapshots/markdownlint-test-scenarios.js.md

@@ -3247,26 +3247,6 @@ Generated by [AVA](https://avajs.dev).
             'no-bare-urls',
           ],
         },
-        {
-          errorContext: 'https://example.com',
-          errorDetail: null,
-          errorRange: [
-            45,
-            19,
-          ],
-          fixInfo: {
-            deleteCount: 19,
-            editColumn: 45,
-            insertText: '<https://example.com>',
-          },
-          lineNumber: 86,
-          ruleDescription: 'Bare URL used',
-          ruleInformation: 'https://github.com/DavidAnson/markdownlint/blob/v0.0.0/doc/md034.md',
-          ruleNames: [
-            'MD034',
-            'no-bare-urls',
-          ],
-        },
       ],
       fixed: `# Detailed Results Bare URLs␊
       ␊
@@ -3353,10 +3333,9 @@ Generated by [AVA](https://avajs.dev).
       ␊
       Links bind to the innermost [link that [is-a-valid] link](<https://example.com>) {MD034}␊
       ␊
-      But not if the [link [is-not-a-valid] link](<https://example.com>) {MD034}␊
-      HOWEVER this scenario could have an invalid shortcut and IS reported␊
+      But not if the [link [is-not-a-valid] link](https://example.com)␊
       ␊
-      Escaping both inner square brackets avoids the unwanted report:␊
+      Escaping both inner square brackets avoids confusion:␊
       [link \\[is-not-a-valid\\] link](https://example.com)␊
       `,
     }
@@ -23985,6 +23964,26 @@ Generated by [AVA](https://avajs.dev).
             'no-bare-urls',
           ],
         },
+        {
+          errorContext: 'https://example.com',
+          errorDetail: null,
+          errorRange: [
+            43,
+            19,
+          ],
+          fixInfo: {
+            deleteCount: 19,
+            editColumn: 43,
+            insertText: '<https://example.com>',
+          },
+          lineNumber: 30,
+          ruleDescription: 'Bare URL used',
+          ruleInformation: 'https://github.com/DavidAnson/markdownlint/blob/v0.0.0/doc/md034.md',
+          ruleNames: [
+            'MD034',
+            'no-bare-urls',
+          ],
+        },
       ],
       fixed: `# Link test␊
       ␊
@@ -24013,6 +24012,11 @@ Generated by [AVA](https://avajs.dev).
       | Link                 | Same Link            | Violation |␊
       |----------------------|----------------------|-----------|␊
       | <https://example.com/> | <https://example.com/> | {MD034}   |␊
+      ␊
+      This is not a bare URL: [text [undefined] text](https://example.com).␊
+      This is a bare URL: [text [defined] text](<https://example.com>). {MD034}␊
+      ␊
+      [defined]: https://example.com␊
       `,
     }