Reimplement MD037/no-space-in-emphasis to better handle multiple fixes on a line and more scenarios (fixes #251).

helpers/helpers.js

@@ -234,17 +234,22 @@ module.exports.getLineMetadata = function getLineMetadata(params) {
     }
   });
   filterTokens(params, "list_item_open", function forToken(token) {
+    let count = 1;
     for (let i = token.map[0]; i < token.map[1]; i++) {
-      lineMetadata[i][5] = true;
+      lineMetadata[i][5] = count;
+      count++;
     }
   });
+  filterTokens(params, "hr", function forToken(token) {
+    lineMetadata[token.map[0]][6] = true;
+  });
   return lineMetadata;
 };
 
 // Calls the provided function for each line (with context)
 module.exports.forEachLine = function forEachLine(lineMetadata, handler) {
   lineMetadata.forEach(function forMetadata(metadata) {
-    // Parameters: line, lineIndex, inCode, onFence, inTable
+    // Parameters: line, lineIndex, inCode, onFence, inTable, inBreak
     handler(...metadata);
   });
 };

lib/md037.js

@@ -2,39 +2,67 @@
 
 "use strict";
 
-const { addErrorContext, forEachInlineChild } = require("../helpers");
+const { addErrorContext, forEachLine } = require("../helpers");
+const { lineMetadata } = require("./cache");
 
-const leftSpaceRe = /(?:^|\s)(\*\*?\*?|__?_?)\s.*[^\\]\1/g;
-const rightSpaceRe = /(?:^|[^\\])(\*\*?\*?|__?_?).+\s\1(?:\s|$)/g;
+const emphasisRe = /(^|[^\\])(?:(\*\*?\*?)|(__?_?))/g;
+const asteriskListItemMarkerRe = /^(\s*)\*(\s+)/;
+const leftSpaceRe = /^\s+/;
+const rightSpaceRe = /\s+$/;
 
 module.exports = {
   "names": [ "MD037", "no-space-in-emphasis" ],
   "description": "Spaces inside emphasis markers",
   "tags": [ "whitespace", "emphasis" ],
   "function": function MD037(params, onError) {
-    forEachInlineChild(params, "text", (token) => {
-      const { content, lineNumber } = token;
-      const columnsReported = [];
-      [ leftSpaceRe, rightSpaceRe ].forEach((spaceRe, index) => {
+    forEachLine(
+      lineMetadata(),
+      (line, lineIndex, inCode, onFence, inTable, inItem, inBreak) => {
+        if (inCode || inBreak) {
+          // Emphasis has no meaning here
+          return;
+        }
+        if (inItem === 1) {
+          // Trim overlapping '*' list item marker
+          line = line.replace(asteriskListItemMarkerRe, "$1 $2");
+        }
         let match = null;
-        while ((match = spaceRe.exec(content)) !== null) {
-          const [ fullText, marker ] = match;
-          const line = params.lines[lineNumber - 1];
-          if (line.includes(fullText)) {
-            const text = fullText.trim();
-            const column = line.indexOf(text) + 1;
-            if (!columnsReported.includes(column)) {
-              const length = text.length;
-              const markerLength = marker.length;
-              const emphasized =
-                text.slice(markerLength, length - markerLength);
-              const fixedText = `${marker}${emphasized.trim()}${marker}`;
+        let emphasisIndex = -1;
+        let emphasisLength = 0;
+        // Match all emphasis-looking runs in the line...
+        while ((match = emphasisRe.exec(line))) {
+          const matchIndex = match.index + match[1].length;
+          const matchLength = match[0].length - match[1].length;
+          if (emphasisIndex === -1) {
+            // New run
+            emphasisLength = matchLength;
+            emphasisIndex = matchIndex + emphasisLength;
+          } else {
+            // Already in a run
+            if (matchLength !== emphasisLength) {
+              // Looks like a run within a run, reset to embedded run
+              emphasisLength += matchLength;
+              emphasisIndex = matchIndex + emphasisLength;
+            }
+            // Extract emphasized content
+            const content = line.substring(emphasisIndex, matchIndex);
+            const leftSpace = leftSpaceRe.test(content);
+            const rightSpace = rightSpaceRe.test(content);
+            if (leftSpace || rightSpace) {
+              // Report the violation
+              const contextStart = emphasisIndex - emphasisLength;
+              const contextEnd = matchIndex + emphasisLength;
+              const context = line.substring(contextStart, contextEnd);
+              const column = contextStart + 1;
+              const length = contextEnd - contextStart;
+              const marker = match[2] || match[3];
+              const fixedText = `${marker}${content.trim()}${marker}`;
               addErrorContext(
                 onError,
-                lineNumber,
-                text,
-                index === 0,
-                index !== 0,
+                lineIndex + 1,
+                context,
+                leftSpace,
+                rightSpace,
                 [ column, length ],
                 {
                   "editColumn": column,
@@ -42,11 +70,15 @@ module.exports = {
                   "insertText": fixedText
                 }
               );
-              columnsReported.push(column);
+            }
+            // Update the run
+            emphasisLength -= matchLength;
+            if (!emphasisLength) {
+              emphasisIndex = -1;
             }
           }
         }
-      });
-    });
+      }
+    );
   }
 };

test/escaped-emphasis-markers.md

@@ -1,28 +1,42 @@
 # Heading
 
+## Single-character markers
+
+None are valid emphasis without spaces.
+
 Escaped asterisks \* should \* be ignored by MD037.
 
 Escaped asterisks \* should * be ignored by MD037.
 
 Escaped asterisks * should \* be ignored by MD037.
 
-Escaped asterisks \** should ** be ignored by MD037.
-
-Escaped asterisks *\* should ** be ignored by MD037.
-
-Escaped asterisks ** should \** be ignored by MD037.
-
-Escaped asterisks ** should *\* be ignored by MD037.
-
 Escaped underscores \_ should \_ be ignored by MD037.
 
 Escaped underscores \_ should _ be ignored by MD037.
 
 Escaped underscores _ should \_ be ignored by MD037.
 
-Escaped underscores \__ should __ be ignored by MD037.
+## Double-character markers, start
 
-Escaped underscores _\_ should __ be ignored by MD037.
+All should be reported because they are valid single-character
+marker emphasis without spaces.
+
+Escaped asterisks \** should ** be ignored by MD037. {MD037}
+
+Escaped asterisks *\* should ** be ignored by MD037. {MD037}
+
+Escaped underscores \__ should __ be ignored by MD037. {MD037}
+
+Escaped underscores _\_ should __ be ignored by MD037. {MD037}
+
+## Double-character markers, end
+
+All should be reported, but are ignored because they look like
+the start of an embedded emphasis.
+
+Escaped asterisks ** should \** be ignored by MD037.
+
+Escaped asterisks ** should *\* be ignored by MD037.
 
 Escaped underscores __ should \__ be ignored by MD037.
 

test/spaces_inside_emphasis_markers.md

@@ -44,9 +44,11 @@ One-sided *broken emphasis * {MD037}
 
 One-sided * broken emphasis* {MD037}
 
-Don't _flag on _words with underscores before them.
+Will _flag on _words with underscores before them. {MD037}
 
-The same goes for words* with asterisks* after them.
+The same goes for words* with asterisks* after them. {MD037}
+
+But not with escaped\* asterisks\* \_and \_underscores.
 
 * Emphasis* with left space is recognized as a list
 
@@ -95,3 +97,31 @@ ___Strong and emphasis ___ with right space {MD037}
 {MD037} Right space __strong __
 
 {MD037} Right space ___strong and emphasis ___
+
+**Multiple ** spaces **in ** emphasis **at ** once. {MD037}
+
+**Multiple ** spaces ** in** emphasis ** at ** once. {MD037}
+
+This is * an ambiguous * scenario {MD037}
+
+* List item *with emphasis* on the
+  first and *second lines*.
+* List * item* {MD037}
+* List *item * {MD037}
+* List * item * {MD037}
+* List item with
+  *hanging* emphasis
+  and * some* lines {MD037}
+  with *space * problems {MD037}
+  throughout * the * content {MD037}
+
+Uncommon scenarios from the CommonMark specification:
+***strong emph***
+***strong** in emph*
+***emph* in strong**
+**in strong *emph***
+*in emph **strong***
+
+```markdown
+Violations * are * allowed in code blocks where emphasis does not apply.
+```