From 4fc14026e9bfffeedf317e9b61c7cda8509091ba Mon Sep 17 00:00:00 2001
From: Alex Rudenko <OrKoN@users.noreply.github.com>
Date: Fri, 2 Feb 2024 13:17:06 +0100
Subject: [PATCH] feat!: generate accessible PDFs by default (#11778)

---
 docs/api/puppeteer.pdfoptions.md              |  2 +-
 .../puppeteer-core/src/common/PDFOptions.ts   |  2 +-
 packages/puppeteer-core/src/common/util.ts    |  2 +-
 test/TestExpectations.json                    | 12 ++++++++++++
 test/src/page.spec.ts                         | 19 +++++++++++++++++++
 5 files changed, 34 insertions(+), 3 deletions(-)

diff --git a/docs/api/puppeteer.pdfoptions.md b/docs/api/puppeteer.pdfoptions.md
index eacffd30500..5015d0b2cc2 100644
--- a/docs/api/puppeteer.pdfoptions.md
+++ b/docs/api/puppeteer.pdfoptions.md
@@ -30,6 +30,6 @@ export interface PDFOptions
 | preferCSSPageSize   | <code>optional</code> | boolean                                   | Give any CSS <code>@page</code> size declared in the page priority over what is declared in the <code>width</code> or <code>height</code> or <code>format</code> option.                                                                                                                                                                                                        | <code>false</code>, which will scale the content to fit the paper size.  |
 | printBackground     | <code>optional</code> | boolean                                   | Set to <code>true</code> to print background graphics.                                                                                                                                                                                                                                                                                                                          | <code>false</code>                                                       |
 | scale               | <code>optional</code> | number                                    | Scales the rendering of the web page. Amount must be between <code>0.1</code> and <code>2</code>.                                                                                                                                                                                                                                                                               | <code>1</code>                                                           |
-| tagged              | <code>optional</code> | boolean                                   | Generate tagged (accessible) PDF.                                                                                                                                                                                                                                                                                                                                               | <code>false</code>                                                       |
+| tagged              | <code>optional</code> | boolean                                   | Generate tagged (accessible) PDF.                                                                                                                                                                                                                                                                                                                                               | <code>true</code>                                                        |
 | timeout             | <code>optional</code> | number                                    | Timeout in milliseconds. Pass <code>0</code> to disable timeout.                                                                                                                                                                                                                                                                                                                | <code>30_000</code>                                                      |
 | width               | <code>optional</code> | string \| number                          | Sets the width of paper. You can pass in a number or a string with a unit.                                                                                                                                                                                                                                                                                                      |                                                                          |
diff --git a/packages/puppeteer-core/src/common/PDFOptions.ts b/packages/puppeteer-core/src/common/PDFOptions.ts
index 0271faf7a5f..3d32294e3ed 100644
--- a/packages/puppeteer-core/src/common/PDFOptions.ts
+++ b/packages/puppeteer-core/src/common/PDFOptions.ts
@@ -158,7 +158,7 @@ export interface PDFOptions {
   omitBackground?: boolean;
   /**
    * Generate tagged (accessible) PDF.
-   * @defaultValue `false`
+   * @defaultValue `true`
    * @experimental
    */
   tagged?: boolean;
diff --git a/packages/puppeteer-core/src/common/util.ts b/packages/puppeteer-core/src/common/util.ts
index f8648d6d332..98d61d7425b 100644
--- a/packages/puppeteer-core/src/common/util.ts
+++ b/packages/puppeteer-core/src/common/util.ts
@@ -349,8 +349,8 @@ export function parsePDFOptions(
     pageRanges: '',
     preferCSSPageSize: false,
     omitBackground: false,
-    tagged: false,
     outline: false,
+    tagged: true,
   };
 
   let width = 8.5;
diff --git a/test/TestExpectations.json b/test/TestExpectations.json
index 0af8981ed3b..8ddac7c9197 100644
--- a/test/TestExpectations.json
+++ b/test/TestExpectations.json
@@ -955,6 +955,12 @@
     "parameters": ["webDriverBiDi"],
     "expectations": ["FAIL"]
   },
+  {
+    "testIdPattern": "[page.spec] Page Page.pdf can print to PDF without accessible tags",
+    "platforms": ["darwin", "linux", "win32"],
+    "parameters": ["webDriverBiDi"],
+    "expectations": ["SKIP"]
+  },
   {
     "testIdPattern": "[page.spec] Page Page.pdf should respect timeout",
     "platforms": ["darwin", "linux", "win32"],
@@ -3094,6 +3100,12 @@
     "parameters": ["cdp", "firefox"],
     "expectations": ["FAIL"]
   },
+  {
+    "testIdPattern": "[page.spec] Page Page.pdf can print to PDF without accessible tags",
+    "platforms": ["darwin", "linux", "win32"],
+    "parameters": ["cdp", "firefox"],
+    "expectations": ["SKIP"]
+  },
   {
     "testIdPattern": "[page.spec] Page Page.removeExposedFunction should work",
     "platforms": ["darwin", "linux", "win32"],
diff --git a/test/src/page.spec.ts b/test/src/page.spec.ts
index 8e585df9993..271296b5dba 100644
--- a/test/src/page.spec.ts
+++ b/test/src/page.spec.ts
@@ -1940,6 +1940,25 @@ describe('Page', function () {
       }
     });
 
+    it('can print to PDF without accessible tags', async () => {
+      const {page, server} = await getTestState();
+
+      const outputFile = __dirname + '/../assets/output.pdf';
+      const outputFileAccessible =
+        __dirname + '/../assets/output-accessible.pdf';
+      await page.goto(server.PREFIX + '/pdf.html');
+      await page.pdf({path: outputFile});
+      await page.pdf({path: outputFileAccessible, tagged: false});
+      try {
+        expect(fs.readFileSync(outputFileAccessible).byteLength).toBeLessThan(
+          fs.readFileSync(outputFile).byteLength
+        );
+      } finally {
+        fs.unlinkSync(outputFileAccessible);
+        fs.unlinkSync(outputFile);
+      }
+    });
+
     it('can print to PDF and stream the result', async () => {
       const {page} = await getTestState();