Merge pull request #1366 from pjkaufman/master

Fix CJK Characters at Start of Heading Potentially Getting Starting Space Removed

Author: pjkaufman

__tests__/remove-space-around-characters.test.ts

@@ -92,5 +92,17 @@ ruleTest({
         includeCJKSymbolsAndPunctuation: true,
       },
     },
+    { // accounts for https://github.com/platers/obsidian-linter/issues/1280
+      testName: 'Make sure that heading text is handled properly and keeps the heading as valid',
+      before: dedent`
+        ## 「example here」 More text here
+      `,
+      after: dedent`
+        ## 「example here」More text here
+      `,
+      options: {
+        includeCJKSymbolsAndPunctuation: true,
+      },
+    },
   ],
 });

src/rules/remove-space-around-characters.ts

@@ -2,7 +2,7 @@ import {Options, RuleType} from '../rules';
 import RuleBuilder, {BooleanOptionBuilder, ExampleBuilder, OptionBuilderBase, TextOptionBuilder} from './rule-builder';
 import dedent from 'ts-dedent';
 import {ignoreListOfTypes, IgnoreTypes} from '../utils/ignore-types';
-import {updateListItemText} from '../utils/mdast';
+import {updateHeaderText, updateListItemText} from '../utils/mdast';
 import {escapeRegExp} from '../utils/regex';
 
 class RemoveSpaceAroundCharactersOptions implements Options {
@@ -53,9 +53,10 @@ export default class RemoveSpaceAroundCharacters extends RuleBuilder<RemoveSpace
       return text.replace(fullwidthCharacterWithTextAtStart, '$2').replace(fullwidthCharacterWithTextAtEnd, '$1');
     };
 
-    let newText = ignoreListOfTypes([IgnoreTypes.list], text, replaceWhitespaceAroundFullwidthCharacters);
+    let newText = ignoreListOfTypes([IgnoreTypes.list, IgnoreTypes.heading], text, replaceWhitespaceAroundFullwidthCharacters);
 
     newText = updateListItemText(newText, replaceWhitespaceAroundFullwidthCharacters);
+    newText = updateHeaderText(newText, replaceWhitespaceAroundFullwidthCharacters);
 
     return newText;
   }

src/utils/ignore-types.ts

@@ -19,6 +19,7 @@ export const IgnoreTypes: Record<string, IgnoreType> = {
   math: {replaceAction: MDAstTypes.Math, placeholder: '{MATH_PLACEHOLDER}'},
   inlineMath: {replaceAction: MDAstTypes.InlineMath, placeholder: '{INLINE_MATH_PLACEHOLDER}'},
   html: {replaceAction: MDAstTypes.Html, placeholder: '{HTML_PLACEHOLDER}'},
+  heading: {replaceAction: MDAstTypes.Heading, placeholder: '{HEADING_PLACEHOLDER}'},
   // RegExp
   yaml: {replaceAction: yamlRegex, placeholder: escapeDollarSigns('---\n---')},
   wikiLink: {replaceAction: wikiLinkRegex, placeholder: '{WIKI_LINK_PLACEHOLDER}'},

src/utils/mdast.ts

@@ -24,6 +24,11 @@ type PositionPlusEmptyIndicator = {
   isEmpty: boolean,
 }
 
+type PositionPlusText = {
+  position: Position,
+  text: string,
+}
+
 export enum MDAstTypes {
   Link = 'link',
   Footnote = 'footnoteDefinition',
@@ -38,6 +43,8 @@ export enum MDAstTypes {
   Blockquote = 'blockquote',
   HorizontalRule = 'thematicBreak',
   Html = 'html',
+  Heading = 'heading',
+  Text = 'text',
   // math types
   Math = 'math',
   InlineMath = 'inlineMath',
@@ -149,6 +156,31 @@ function getListItemTextPositions(text: string, includeEmptyNodes: boolean = fal
   return positions;
 }
 
+function getHeaderTextPositions(text: string): PositionPlusText[] {
+  const ast = parseTextToAST(text);
+  const positions: PositionPlusText[] = [];
+  visit(ast, MDAstTypes.Heading as string, (node) => {
+    // @ts-ignore the fact that not all nodes have a children property since I am skipping any that do not
+    if (!node.children || node.children.length === 0) {
+      return;
+    }
+
+    // @ts-ignore the fact that not all nodes have a children property since I have already exited the function if that is the case
+    for (const childNode of node.children) {
+      if (childNode.type === (MDAstTypes.Text as string)) {
+        positions.push({
+          position: childNode.position as Position,
+          text: childNode.value as string,
+        });
+      }
+    }
+  });
+
+  // Sort positions by start position in reverse order
+  positions.sort((a, b) => b.position.start.offset - a.position.start.offset);
+  return positions;
+}
+
 // mdast helper methods
 
 /**
@@ -1169,3 +1201,30 @@ export function ensureFencedCodeBlocksHasLanguage(text: string, defaultLanguage:
 
   return text;
 }
+
+export function updateHeaderText(text: string, func:(text: string) => string): string {
+  const positions = getHeaderTextPositions(text);
+
+  // for the best performance, we want to grab all places that need updating and then
+  // at the end we want to update the text in one go because otherwise we get a lot of
+  // instances of the file text in memory
+  const updateLocations: {startIndex: number, endIndex: number, newText: string}[] = [];
+  for (const position of positions) {
+    const updatedText = func(position.text);
+    if (updatedText !== position.text) {
+      const headerText = text.substring(position.position.start.offset, position.position.end.offset);
+      const startIndex = position.position.start.offset+ headerText.indexOf(position.text);
+      updateLocations.push({
+        startIndex: startIndex,
+        endIndex: startIndex + position.text.length,
+        newText: updatedText,
+      });
+    }
+  }
+
+  for (const headerUpdate of updateLocations) {
+    text = replaceTextBetweenStartAndEndWithNewValue(text, headerUpdate.startIndex, headerUpdate.endIndex, headerUpdate.newText);
+  }
+
+  return text;
+}