test cookie

yt_dlp/jsinterp/_deno.py

@@ -2,7 +2,6 @@ from __future__ import annotations
 
 import http.cookiejar
 import json
-import re
 import subprocess
 import typing
 import urllib.parse
@@ -16,7 +15,7 @@ from ..utils import (
     shell_quote,
     unified_timestamp,
 )
-from ._helper import TempFileWrapper, random_string
+from ._helper import TempFileWrapper, random_string, override_navigator_js, extract_script_tags
 from .common import ExternalJSI, register_jsi
 
 
@@ -36,15 +35,7 @@ class DenoJSI(ExternalJSI):
 
     @property
     def _override_navigator_js(self):
-        return '\n'.join([
-            'Object.defineProperty(navigator, "%s", { value: %s, configurable: true });' % (k, json.dumps(v))
-            for k, v in {
-                'userAgent': self.user_agent,
-                'language': 'en-US',
-                'languages': ['en-US'],
-                'webdriver': False,
-            }.items()
-        ])
+        return override_navigator_js(self.user_agent)
 
     def _run_deno(self, cmd):
         self.write_debug(f'Deno command line: {shell_quote(cmd)}')
@@ -137,21 +128,13 @@ class DenoJSDomJSI(DenoJSI):
             self._run_deno(cmd)
         self._JSDOM_IMPORT_CHECKED = True
 
-    def _parse_script_tags(self, html: str):
-        for match_start in re.finditer(r'<script[^>]*>', html, re.DOTALL):
-            end = html.find('</script>', match_start.end())
-            if end > match_start.end():
-                yield html[match_start.end():end]
-
     def execute(self, jscode, video_id=None, note='Executing JS in Deno', location='', html='', cookiejar=None):
         self.report_note(video_id, note)
         self._ensure_jsdom()
         callback_varname = f'__callback_{random_string()}'
 
-        inline_scripts = '\n'.join([
-            'try { %s } catch (e) {}' % script
-            for script in self._parse_script_tags(html)
-        ])
+        html, inline_scripts = extract_script_tags(html)
+        wrapper_scripts = '\n'.join(['try { %s } catch (e) {}' % script for script in inline_scripts])
 
         script = f'''{self._init_script};
         {self._override_navigator_js};
@@ -164,27 +147,33 @@ class DenoJSDomJSI(DenoJSI):
             }});
             Object.keys(dom.window).forEach((key) => {{try {{window[key] = dom.window[key]}} catch (e) {{}}}});
             delete window.jsdom;
+            const origLog = console.log;
+            console.log = () => {{}};
+            console.info = () => {{}};
             return () => {{
                 const stdout = [];
-                const origLog = console.log;
                 console.log = (...msg) => stdout.push(msg.map(m => m.toString()).join(' '));
                 return () => {{ origLog(JSON.stringify({{
                     stdout: stdout.join('\\n'), cookies: jar.serializeSync().cookies}})); }}
             }}
         }})();
-        await (async () => {{
-            {inline_scripts}
-        }})();
-        {callback_varname} = {callback_varname}();
-        await (async () => {{
+        {wrapper_scripts}
+        {callback_varname} = {callback_varname}(); // begin to capture console.log
+        try {{
             {jscode}
-        }})().finally({callback_varname});
+        }} finally {{
+            {callback_varname}();
+        }}
         '''
 
         location_args = ['--location', location] if location else []
         with TempFileWrapper(script, suffix='.js') as js_file:
             cmd = [self.exe, 'run', *self._flags, *location_args, js_file.name]
-            data = json.loads(self._run_deno(cmd))
+            result = self._run_deno(cmd)
+            try:
+                data = json.loads(result)
+            except json.JSONDecodeError as e:
+                raise ExtractorError(f'Failed to parse JSON output from Deno: {result}', cause=e)
         self.apply_cookies(cookiejar, data['cookies'])
         return data['stdout']
 

yt_dlp/jsinterp/_helper.py

@@ -1,7 +1,9 @@
 from __future__ import annotations
 import contextlib
+import json
 import os
 import random
+import re
 import string
 import tempfile
 
@@ -69,3 +71,31 @@ class TempFileWrapper:
 
 def random_string(length: int = 10) -> str:
     return ''.join(random.choices(string.ascii_letters, k=length))
+
+
+def override_navigator_js(user_agent: str) -> str:
+    return '\n'.join([
+        'Object.defineProperty(navigator, "%s", { value: %s, configurable: true });' % (k, json.dumps(v))
+        for k, v in {
+            'userAgent': user_agent,
+            'language': 'en-US',
+            'languages': ['en-US'],
+            'webdriver': False,
+        }.items()
+    ])
+
+
+def extract_script_tags(html: str) -> tuple[str, list[str]]:
+    script_indicies = []
+    inline_scripts = []
+
+    for match_start in re.finditer(r'<script[^>]*>', html, re.DOTALL):
+        end = html.find('</script>', match_start.end())
+        if end > match_start.end():
+            script_indicies.append((match_start.start(), end + len('</script>')))
+            inline_scripts.append(html[match_start.end():end])
+
+    for start, end in script_indicies:
+        html = html[:start] + html[end:]
+
+    return html, inline_scripts

yt_dlp/jsinterp/_phantomjs.py

@@ -16,7 +16,7 @@ from ..utils import (
     is_outdated_version,
     shell_quote,
 )
-from ._helper import TempFileWrapper, random_string
+from ._helper import TempFileWrapper, random_string, extract_script_tags
 from .common import ExternalJSI, register_jsi
 
 
@@ -135,19 +135,23 @@ class PhantomJSJSI(ExternalJSI):
         if 'saveAndExit();' not in jscode:
             raise ExtractorError('`saveAndExit();` not found in `jscode`')
 
+        html, inline_scripts = extract_script_tags(html)
+        wrapped_scripts = '\n'.join([
+            'page.evaluate(function() { try { %s } catch (e) {} });' % inline for inline in inline_scripts])
+
         html_file = TempFileWrapper(html, suffix='.html')
         cookie_file = TempFileWrapper(self._save_cookies(url, cookiejar), suffix='.json')
 
-        jscode = self._TEMPLATE.format_map({
+        script = self._TEMPLATE.format_map({
             'url': json.dumps(str(url)),
             'ua': json.dumps(str(self.user_agent)),
-            'jscode': jscode,
+            'jscode': f'{wrapped_scripts}\n{jscode}',
             'html_fn': json.dumps(html_file.name),
             'cookies_fn': json.dumps(cookie_file.name),
             'timeout': int(self.timeout * 1000),
         })
 
-        stdout = self._execute(jscode, video_id, note=note)
+        stdout = self._execute(script, video_id, note=note)
         self._load_cookies(cookie_file.read(), cookiejar)
         new_html = html_file.read()