@cjser/parse5 8.0.1-cjser.2 → 8.1.0-cjser.2

package/dist/index.d.ts

@@ -16,8 +16,8 @@ export declare namespace DefaultTreeAdapterTypes {
     type Node = DefaultTreeAdapter.Node;
     type DefaultTreeAdapterMap = DefaultTreeAdapter.DefaultTreeAdapterMap;
 }
-export type { TreeAdapter, TreeAdapterTypeMap } from './tree-adapters/interface.js';
-export { type ParserOptions, /** @internal */ Parser } from './parser/index.js';
+export type { ElementCreationContext, TreeAdapter, TreeAdapterTypeMap } from './tree-adapters/interface.js';
+export { type ParserOptions, /** @internal */ Parser, InsertionMode } from './parser/index.js';
 export { serialize, serializeOuter, type SerializerOptions } from './serializer/index.js';
 export { ERR as ErrorCodes, type ParserError, type ParserErrorHandler } from './common/error-codes.js';
 /** @internal */

package/dist/index.js

@@ -1,6 +1,6 @@
 import { Parser } from './parser/index.js';
 export { defaultTreeAdapter } from './tree-adapters/default.js';
-export { /** @internal */ Parser } from './parser/index.js';
+export { /** @internal */ Parser, InsertionMode } from './parser/index.js';
 export { serialize, serializeOuter } from './serializer/index.js';
 export { ERR as ErrorCodes } from './common/error-codes.js';
 /** @internal */

package/dist/parser/index.d.ts

@@ -4,8 +4,8 @@ import { FormattingElementList } from './formatting-element-list.js';
 import { ERR, type ParserErrorHandler } from '../common/error-codes.js';
 import { TAG_ID as $, NS } from '../common/html.js';
 import type { TreeAdapter, TreeAdapterTypeMap } from '../tree-adapters/interface.js';
-import { type Token, type CommentToken, type CharacterToken, type TagToken, type DoctypeToken, type EOFToken, type LocationWithAttributes } from '../common/token.js';
-declare enum InsertionMode {
+import { type Token, type CommentToken, type CharacterToken, type TagToken, type DoctypeToken, type EOFToken, type LocationWithAttributes, type Attribute } from '../common/token.js';
+export declare enum InsertionMode {
     INITIAL = 0,
     BEFORE_HTML = 1,
     BEFORE_HEAD = 2,
@@ -30,6 +30,10 @@ declare enum InsertionMode {
     AFTER_AFTER_BODY = 21,
     AFTER_AFTER_FRAMESET = 22
 }
+interface InsertionLocation<T extends TreeAdapterTypeMap> {
+    parent: T['parentNode'];
+    beforeElement: T['element'] | null;
+}
 export interface ParserOptions<T extends TreeAdapterTypeMap> {
     /**
      * The [scripting flag](https://html.spec.whatwg.org/multipage/parsing.html#scripting-flag). If set
@@ -134,7 +138,11 @@ export declare class Parser<T extends TreeAdapterTypeMap> implements TokenHandle
     /** @protected */
     _setDocumentType(token: DoctypeToken): void;
     /** @protected */
-    _attachElementToTree(element: T['element'], location: LocationWithAttributes | null): void;
+    _attachElementToTree(element: T['element'], location: LocationWithAttributes | null, insertionLocation: InsertionLocation<T>): void;
+    /** @protected */
+    _getInsertionLocation(): InsertionLocation<T>;
+    /** @protected */
+    _createElement(tagName: string, namespaceURI: NS, attrs: Attribute[], intendedParent: T['parentNode']): T['element'];
     /**
      * For self-closing tags. Add an element to the tree, but skip adding it
      * to the stack.
@@ -177,10 +185,7 @@ export declare class Parser<T extends TreeAdapterTypeMap> implements TokenHandle
     /** @protected */
     _shouldFosterParentOnInsertion(): boolean;
     /** @protected */
-    _findFosterParentingLocation(): {
-        parent: T['parentNode'];
-        beforeElement: T['element'] | null;
-    };
+    _findFosterParentingLocation(): InsertionLocation<T>;
     /** @protected */
     _fosterParentElement(element: T['element']): void;
     /** @protected */

package/dist/parser/index.js

@@ -14,7 +14,7 @@ const HIDDEN_INPUT_TYPE = 'hidden';
 const AA_OUTER_LOOP_ITER = 8;
 const AA_INNER_LOOP_ITER = 3;
 //Insertion modes
-var InsertionMode;
+export var InsertionMode;
 (function (InsertionMode) {
     InsertionMode[InsertionMode["INITIAL"] = 0] = "INITIAL";
     InsertionMode[InsertionMode["BEFORE_HTML"] = 1] = "BEFORE_HTML";
@@ -273,7 +273,7 @@ export class Parser {
         }
     }
     /** @protected */
-    _attachElementToTree(element, location) {
+    _attachElementToTree(element, location, insertionLocation) {
         if (this.options.sourceCodeLocationInfo) {
             const loc = location && {
                 ...location,
@@ -281,13 +281,34 @@ export class Parser {
             };
             this.treeAdapter.setNodeSourceCodeLocation(element, loc);
         }
-        if (this._shouldFosterParentOnInsertion()) {
-            this._fosterParentElement(element);
+        if (insertionLocation.beforeElement) {
+            this.treeAdapter.insertBefore(insertionLocation.parent, element, insertionLocation.beforeElement);
         }
         else {
-            const parent = this.openElements.currentTmplContentOrNode;
-            this.treeAdapter.appendChild(parent !== null && parent !== void 0 ? parent : this.document, element);
+            this.treeAdapter.appendChild(insertionLocation.parent, element);
+        }
+    }
+    /** @protected */
+    _getInsertionLocation() {
+        var _a;
+        if (this._shouldFosterParentOnInsertion()) {
+            return this._findFosterParentingLocation();
         }
+        return {
+            parent: (_a = this.openElements.currentTmplContentOrNode) !== null && _a !== void 0 ? _a : this.document,
+            beforeElement: null,
+        };
+    }
+    /** @protected */
+    _createElement(tagName, namespaceURI, attrs, intendedParent) {
+        var _a, _b;
+        const element = this.treeAdapter.createElement(tagName, namespaceURI, attrs);
+        (_b = (_a = this.treeAdapter).onElementCreated) === null || _b === void 0 ? void 0 : _b.call(_a, element, {
+            intendedParent,
+            formElement: this.formElement,
+            isInTemplate: this.openElements.tmplCount > 0,
+        });
+        return element;
     }
     /**
      * For self-closing tags. Add an element to the tree, but skip adding it
@@ -295,37 +316,41 @@ export class Parser {
      */
     /** @protected */
     _appendElement(token, namespaceURI) {
-        const element = this.treeAdapter.createElement(token.tagName, namespaceURI, token.attrs);
-        this._attachElementToTree(element, token.location);
+        const insertionLocation = this._getInsertionLocation();
+        const element = this._createElement(token.tagName, namespaceURI, token.attrs, insertionLocation.parent);
+        this._attachElementToTree(element, token.location, insertionLocation);
     }
     /** @protected */
     _insertElement(token, namespaceURI) {
-        const element = this.treeAdapter.createElement(token.tagName, namespaceURI, token.attrs);
-        this._attachElementToTree(element, token.location);
+        const insertionLocation = this._getInsertionLocation();
+        const element = this._createElement(token.tagName, namespaceURI, token.attrs, insertionLocation.parent);
+        this._attachElementToTree(element, token.location, insertionLocation);
         this.openElements.push(element, token.tagID);
     }
     /** @protected */
     _insertFakeElement(tagName, tagID) {
-        const element = this.treeAdapter.createElement(tagName, NS.HTML, []);
-        this._attachElementToTree(element, null);
+        const insertionLocation = this._getInsertionLocation();
+        const element = this._createElement(tagName, NS.HTML, [], insertionLocation.parent);
+        this._attachElementToTree(element, null, insertionLocation);
         this.openElements.push(element, tagID);
     }
     /** @protected */
     _insertTemplate(token) {
-        const tmpl = this.treeAdapter.createElement(token.tagName, NS.HTML, token.attrs);
+        const insertionLocation = this._getInsertionLocation();
+        const tmpl = this._createElement(token.tagName, NS.HTML, token.attrs, insertionLocation.parent);
         const content = this.treeAdapter.createDocumentFragment();
         this.treeAdapter.setTemplateContent(tmpl, content);
-        this._attachElementToTree(tmpl, token.location);
+        this._attachElementToTree(tmpl, token.location, insertionLocation);
         this.openElements.push(tmpl, token.tagID);
         if (this.options.sourceCodeLocationInfo)
             this.treeAdapter.setNodeSourceCodeLocation(content, null);
     }
     /** @protected */
     _insertFakeRootElement() {
-        const element = this.treeAdapter.createElement(TN.HTML, NS.HTML, []);
+        const element = this._createElement(TN.HTML, NS.HTML, [], this.document);
         if (this.options.sourceCodeLocationInfo)
             this.treeAdapter.setNodeSourceCodeLocation(element, null);
-        this.treeAdapter.appendChild(this.openElements.current, element);
+        this.treeAdapter.appendChild(this.document, element);
         this.openElements.push(element, $.HTML);
     }
     /** @protected */
@@ -1217,6 +1242,7 @@ function aaObtainFurthestBlock(p, formattingElementEntry) {
 function aaInnerLoop(p, furthestBlock, formattingElement) {
     let lastElement = furthestBlock;
     let nextElement = p.openElements.getCommonAncestor(furthestBlock);
+    const intendedParent = p.openElements.getCommonAncestor(formattingElement);
     for (let i = 0, element = nextElement; element !== formattingElement; i++, element = nextElement) {
         //NOTE: store the next element for the next loop iteration (it may be deleted from the stack by step 9.5)
         nextElement = p.openElements.getCommonAncestor(element);
@@ -1230,7 +1256,7 @@ function aaInnerLoop(p, furthestBlock, formattingElement) {
             p.openElements.remove(element);
         }
         else {
-            element = aaRecreateElementFromEntry(p, elementEntry);
+            element = aaRecreateElementFromEntry(p, elementEntry, intendedParent);
             if (lastElement === furthestBlock) {
                 p.activeFormattingElements.bookmark = elementEntry;
             }
@@ -1242,9 +1268,9 @@ function aaInnerLoop(p, furthestBlock, formattingElement) {
     return lastElement;
 }
 //Step 13.7 of the algorithm
-function aaRecreateElementFromEntry(p, elementEntry) {
+function aaRecreateElementFromEntry(p, elementEntry, intendedParent) {
     const ns = p.treeAdapter.getNamespaceURI(elementEntry.element);
-    const newElement = p.treeAdapter.createElement(elementEntry.token.tagName, ns, elementEntry.token.attrs);
+    const newElement = p._createElement(elementEntry.token.tagName, ns, elementEntry.token.attrs, intendedParent);
     p.openElements.replace(elementEntry.element, newElement);
     elementEntry.element = newElement;
     return newElement;
@@ -1268,7 +1294,7 @@ function aaInsertLastNodeInCommonAncestor(p, commonAncestor, lastElement) {
 function aaReplaceFormattingElement(p, furthestBlock, formattingElementEntry) {
     const ns = p.treeAdapter.getNamespaceURI(formattingElementEntry.element);
     const { token } = formattingElementEntry;
-    const newElement = p.treeAdapter.createElement(token.tagName, ns, token.attrs);
+    const newElement = p._createElement(token.tagName, ns, token.attrs, furthestBlock);
     p._adoptNodes(furthestBlock, newElement);
     p.treeAdapter.appendChild(furthestBlock, newElement);
     p.activeFormattingElements.insertElementAfterBookmark(newElement, token);
@@ -2227,7 +2253,10 @@ function genericEndTagInBody(p, token) {
         const element = p.openElements.items[i];
         const elementId = p.openElements.tagIDs[i];
         // Compare the tag name here, as the tag might not be a known tag with an ID.
-        if (tid === elementId && (tid !== $.UNKNOWN || p.treeAdapter.getTagName(element) === tn)) {
+        // Only HTML elements match; a foreign element falls through to the special-element check.
+        if (tid === elementId &&
+            p.treeAdapter.getNamespaceURI(element) === NS.HTML &&
+            (tid !== $.UNKNOWN || p.treeAdapter.getTagName(element) === tn)) {
             p.openElements.generateImpliedEndTagsWithExclusion(tid);
             if (p.openElements.stackTop >= i)
                 p.openElements.shortenToLength(i);

package/dist/tokenizer/index.d.ts

@@ -1,6 +1,6 @@
 import { Preprocessor } from './preprocessor.js';
 import { type Token, type CharacterToken, type DoctypeToken, type TagToken, type EOFToken, type CommentToken, type Attribute, type Location } from '../common/token.js';
-import { EntityDecoder } from '@cjser/entities/decode';
+import { HtmlEntityDecoder } from '@cjser/entities/decode';
 import { ERR, type ParserErrorHandler } from '../common/error-codes.js';
 declare const enum State {
     DATA = 0,
@@ -118,9 +118,9 @@ export declare class Tokenizer {
     state: State;
     protected returnState: State;
     /**
-     * We use `entities`' `EntityDecoder` to parse character references.
+     * We use `entities`' `HtmlEntityDecoder` to parse character references.
      *
-     * All of the following states are handled by the `EntityDecoder`:
+     * All of the following states are handled by the `HtmlEntityDecoder`:
      *
      * - Named character reference state
      * - Numeric character reference state
@@ -129,17 +129,28 @@ export declare class Tokenizer {
      * - Decimal character reference state
      * - Numeric character reference end state
      */
-    protected entityDecoder: EntityDecoder;
+    protected entityDecoder: HtmlEntityDecoder;
     protected entityStartPos: number;
     protected consumedAfterSnapshot: number;
     protected currentLocation: Location | null;
     protected currentCharacterToken: CharacterToken | null;
     protected currentToken: Token | null;
     protected currentAttr: Attribute;
+    private currentAttrLocation;
+    /**
+     * `[start, end)` of `preprocessor.html` still to be appended to the attribute value, comment data or character
+     * token being built (`_flushRun`): input that is used as is is appended as one slice, not a code point at a time.
+     */
+    protected runStart: number;
+    protected runEnd: number;
     constructor(options: TokenizerOptions, handler: TokenHandler);
     protected _err(code: ERR, cpOffset?: number): void;
     protected getCurrentLocation(offset: number): Location | null;
     protected _runParsingLoop(): void;
+    /** Emit buffered character tokens without completing the current token or emitting EOF. */
+    flushCharacters(): void;
+    private getCharacterTokenEndLocation;
+    private getCharacterTokenEndOffset;
     pause(): void;
     resume(writeCallback?: () => void): void;
     write(chunk: string, isLastChunk: boolean, writeCallback?: () => void): void;
@@ -152,19 +163,35 @@ export declare class Tokenizer {
     protected _createEndTagToken(): void;
     protected _createCommentToken(offset: number): void;
     protected _createDoctypeToken(initialName: string | null): void;
-    protected _createCharacterToken(type: CharacterToken['type'], chars: string): void;
+    protected _createCharacterToken(type: CharacterToken['type']): void;
     protected _createAttr(attrNameFirstCh: string): void;
     protected _leaveAttrName(): void;
     protected _leaveAttrValue(): void;
+    /**
+     * Append the pending run and `text` to the attribute value, comment data or character token being built.
+     * The run is a string of its own: V8 copies a slice of fewer than 13 characters but lets a longer one share
+     * the input's buffer, which would keep that buffer alive for as long as any attribute value or text cut from
+     * it is; a concatenation is flattened into a buffer of its own when it is sliced.
+     */
+    protected _flushRun(text: string): void;
+    /** Extend the pending run by the current input code point, read as `cp`, or flush it and start a new one. */
+    protected _extendRun(cp: number): void;
+    /**
+     * Extend the pending run by `text`, the input that ends `cpOffset` code points after the current one, or flush
+     * the run and append `text` where the input differs (a NUL replacement, a decoded reference, a dropped chunk).
+     */
+    protected _extendRunWith(text: string, cpOffset: number): void;
     protected prepareToken(ct: Token): void;
     protected emitCurrentTagToken(): void;
     protected emitCurrentComment(ct: CommentToken): void;
     protected emitCurrentDoctype(ct: DoctypeToken): void;
     protected _emitCurrentCharacterToken(nextLocation: Location | null): void;
     protected _emitEOFToken(): void;
-    protected _appendCharToCurrentCharacterToken(type: CharacterToken['type'], ch: string): void;
+    /** Make the current character token one of `type`, emitting the current one if it has another type. */
+    protected _ensureCharacterToken(type: CharacterToken['type']): void;
+    /** Emit the current input code point, read as `cp`. */
     protected _emitCodePoint(cp: number): void;
-    protected _emitChars(ch: string): void;
+    protected _emitChars(ch: string, cpOffset?: number): void;
     protected _startCharacterReference(): void;
     protected _isCharacterReferenceInAttribute(): boolean;
     protected _flushCodePointConsumedAsCharacterReference(cp: number): void;