Merge pull request #523 from mathjax/liteDOM-update

Fix LiteDOM comments and add support for DOCTYPE
mathjax · Jul 23, 2020 · 6bf0ac0 · 6bf0ac0
2 parents 15f82e1 + dfcac38
commit 6bf0ac0
Show file tree

Hide file tree

Showing 5 changed files with 45 additions and 4 deletions.
diff --git a/ts/adaptors/HTMLAdaptor.ts b/ts/adaptors/HTMLAdaptor.ts
@@ -36,6 +36,7 @@ export interface MinDocument<N, T> {
   head: N;
   body: N;
   title: string;
+  doctype: {name: string};
   /* tslint:disable:jsdoc-require */
   createElement(kind: string): N;
   createElementNS(ns: string, kind: string): N;
@@ -228,6 +229,13 @@ AbstractDOMAdaptor<N, T, D> implements MinHTMLAdaptor<N, T, D> {
     return doc.documentElement;
   }
 
+  /**
+   * @override
+   */
+  public doctype(doc: D) {
+    return `<!DOCTYPE ${doc.doctype.name}>`;
+  }
+
   /**
    * @override
    */

diff --git a/ts/adaptors/lite/Document.ts b/ts/adaptors/lite/Document.ts
@@ -41,6 +41,11 @@ export class LiteDocument {
    */
   public body: LiteElement;
 
+  /**
+   * the DOCTYPE comment
+   */
+  public type: string;
+
   /**
    * The kind is always #document
    */
@@ -56,5 +61,6 @@ export class LiteDocument {
       this.head = new LiteElement('head'),
       this.body = new LiteElement('body')
     ]);
+    this.type = '<!DOCTYPE html>';
   }
 }
diff --git a/ts/adaptors/lite/Parser.ts b/ts/adaptors/lite/Parser.ts
@@ -41,8 +41,8 @@ export namespace PATTERNS {
   export const OPTIONALSPACE = '(?:\\s|\\n)*';
   export const ATTRIBUTE = ATTNAME + '(?:' + OPTIONALSPACE + '=' + OPTIONALSPACE + VALUE + ')?';
   export const ATTRIBUTESPLIT = '(' + ATTNAME + ')(?:' + OPTIONALSPACE + '=' + OPTIONALSPACE + VALUESPLIT + ')?';
-  export const TAG = '(<(?:' + TAGNAME + '(?:' + SPACE + ATTRIBUTE + ')*' +
-    OPTIONALSPACE + '/?|/' + TAGNAME + '|!--[^]*?--|![^]*?)(?:>|$))';
+  export const TAG = '(<(?:' + TAGNAME + '(?:' + SPACE + ATTRIBUTE + ')*'
+                       + OPTIONALSPACE + '/?|/' + TAGNAME + '|!--[^]*?--|![^]*?)(?:>|$))';
   export const tag = new RegExp(TAG, 'i');
   export const attr = new RegExp(ATTRIBUTE, 'i');
   export const attrsplit = new RegExp(ATTRIBUTESPLIT, 'i');
@@ -277,6 +277,14 @@ export class LiteParser implements MinDOMParser<LiteDocument> {
   protected checkDocument(adaptor: LiteAdaptor, root: LiteDocument) {
     let node = this.getOnlyChild(adaptor, adaptor.body(root));
     if (!node) return;
+    for (const child of adaptor.childNodes(adaptor.body(root))) {
+      if (child === node) {
+        break;
+      }
+      if (child instanceof LiteComment && child.value.match(/^<!DOCTYPE/)) {
+        root.type = child.value;
+      }
+    }
     switch (adaptor.kind(node)) {
     case 'html':
       //
@@ -365,7 +373,7 @@ export class LiteParser implements MinDOMParser<LiteDocument> {
     return adaptor.childNodes(node).map(x => {
       const kind = adaptor.kind(x);
       return (kind === '#text' ? this.protectHTML(adaptor.value(x)) :
-              kind === '#comment' ? adaptor.value(x) :
+              kind === '#comment' ? (x as LiteComment).value :
               this.serialize(adaptor, x as LiteElement));
     }).join('');
   }

diff --git a/ts/adaptors/liteAdaptor.ts b/ts/adaptors/liteAdaptor.ts
@@ -133,6 +133,13 @@ export class LiteAdaptor extends AbstractDOMAdaptor<LiteElement, LiteText, LiteD
     return doc.root;
   }
 
+  /**
+   * @override
+   */
+  public doctype(doc: LiteDocument) {
+    return doc.type;
+  }
+
   /**
    * @override
    */
@@ -402,7 +409,8 @@ export class LiteAdaptor extends AbstractDOMAdaptor<LiteElement, LiteText, LiteD
    * @override
    */
   public value(node: LiteNode | LiteText) {
-    return (node.kind === '#text' ? (node as LiteText).value : '');
+    return (node.kind === '#text' ? (node as LiteText).value :
+            node.kind === '#comment' ? (node as LiteComment).value.replace(/^<!(--)?((?:.|\n)*)\1>$/, '$2') : '');
   }
 
   /**

diff --git a/ts/core/DOMAdaptor.ts b/ts/core/DOMAdaptor.ts
@@ -96,6 +96,12 @@ export interface DOMAdaptor<N, T, D> {
    */
   root(doc: D): N;
 
+  /**
+   * @param {D} doc     The document whose doctype is to be obtained
+   * @return {string}   The DOCTYPE comment
+   */
+  doctype(doc: D): string;
+
   /**
    * @param {N} node        The node to search for tags
    * @param {string} name   The name of the tag to search for
@@ -436,6 +442,11 @@ export abstract class AbstractDOMAdaptor<N, T, D> implements DOMAdaptor<N, T, D>
    */
   public abstract root(doc: D): N;
 
+  /**
+   * @override
+   */
+  public abstract doctype(doc: D): string;
+
   /**
    * @override
    */