parse5-html-rewriting-stream 8.0.1 → 8.1.1

package/README.md

@@ -19,6 +19,26 @@
   📖 <a href="https://parse5.js.org/modules/parse5-html-rewriting-stream.html"><b>Documentation</b></a> 📖
 </p>
 
+## Usage
+
+Modify HTML tokens while streaming the result. Tokens without a handler are
+passed through unchanged; handlers must emit the replacement tokens.
+
+```js
+import { Readable } from 'node:stream';
+import { pipeline } from 'node:stream/promises';
+import { RewritingStream } from 'parse5-html-rewriting-stream';
+
+const rewriter = new RewritingStream();
+rewriter.on('text', (token) => rewriter.emitText({ text: token.text.toUpperCase() }));
+let output = '';
+rewriter.on('data', (chunk) => {
+    output += chunk;
+});
+await pipeline(Readable.from(['<p>Hello, world!</p>']), rewriter);
+console.log(output); // <p>HELLO, WORLD!</p>
+```
+
 ---
 
 <p align="center">

package/dist/index.d.ts

@@ -49,6 +49,8 @@ import { SAXParser, type EndTag, type StartTag, type Doctype, type Text, type Co
  * ```
  */
 export declare class RewritingStream extends SAXParser {
+    /** Whether text is currently tokenized without decoding character references. */
+    private inRawText;
     /** Note: `sourceCodeLocationInfo` is always enabled. */
     constructor();
     _transformChunk(chunk: string): string;

package/dist/index.js

@@ -1,4 +1,4 @@
-import { html } from 'parse5';
+import { TokenizerMode } from 'parse5';
 import { SAXParser, } from 'parse5-sax-parser';
 import { escapeText, escapeAttribute } from 'entities/escape';
 /**
@@ -54,6 +54,8 @@ export class RewritingStream extends SAXParser {
     /** Note: `sourceCodeLocationInfo` is always enabled. */
     constructor() {
         super({ sourceCodeLocationInfo: true });
+        /** Whether text is currently tokenized without decoding character references. */
+        this.inRawText = false;
     }
     _transformChunk(chunk) {
         // NOTE: ignore upstream return values as we want to push to
@@ -69,9 +71,20 @@ export class RewritingStream extends SAXParser {
     }
     // Events
     emitIfListenerExists(eventName, token) {
+        if (eventName === 'startTag') {
+            // The tokenizer has already been switched to the state for this element's content.
+            const { state } = this.tokenizer;
+            this.inRawText =
+                state === TokenizerMode.RAWTEXT ||
+                    state === TokenizerMode.SCRIPT_DATA ||
+                    state === TokenizerMode.PLAINTEXT;
+        }
         if (!super.emitIfListenerExists(eventName, token)) {
             this.emitRaw(this._getRawHtml(token.sourceCodeLocation));
         }
+        if (eventName === 'endTag') {
+            this.inRawText = false;
+        }
         // NOTE: don't skip new lines after `<pre>` and other tags,
         // otherwise we'll have incorrect raw data.
         this.parserFeedbackSimulator.skipNextNewLine = false;
@@ -111,10 +124,7 @@ export class RewritingStream extends SAXParser {
     }
     /** Emits a serialized text token into the output stream. */
     emitText({ text }) {
-        this.push(!this.parserFeedbackSimulator.inForeignContent &&
-            html.hasUnescapedText(this.tokenizer.lastStartTagName, true)
-            ? text
-            : escapeText(text));
+        this.push(this.inRawText ? text : escapeText(text));
     }
     /** Emits a serialized comment token into the output stream. */
     emitComment(token) {

package/package.json

@@ -2,7 +2,7 @@
     "name": "parse5-html-rewriting-stream",
     "type": "module",
     "description": "Streaming HTML rewriter.",
-    "version": "8.0.1",
+    "version": "8.1.1",
     "author": "Ivan Nikulin <ifaaan@gmail.com> (https://github.com/inikulin)",
     "contributors": [
         "James Garbutt (https://github.com/43081j)",
@@ -31,9 +31,9 @@
         }
     },
     "dependencies": {
-        "entities": "^8.0.0",
-        "parse5": "^8.0.0",
-        "parse5-sax-parser": "^8.0.0"
+        "entities": "^9.0.0",
+        "parse5": "^8.1.1",
+        "parse5-sax-parser": "^8.0.1"
     },
     "repository": {
         "type": "git",
@@ -42,5 +42,8 @@
     "files": [
         "dist/**/*.js",
         "dist/**/*.d.ts"
-    ]
+    ],
+    "devDependencies": {
+        "parse5-test-utils": "1.0.0"
+    }
 }