[🐛 Bug]: Issue with shadowroot access

[dornech]

Description

I tried to access the shadow DOM. I have trouble to use the identified shadow root.
See code below.

When I try to access via shadowroot / shadowroot2 I get an error
AttributeError: 'ShadowRoot' object has no attribute 'id'. Did you mean: '_id'?
Actually, this is right because wether I use the direct javascript or the shadow_root property, I get an object with just a property "_id". I am pretty sure it worked in the past.

If I do a shadowroot.id = shadowroot._id it works again.
I guess evaluation fo shadowroot is buggy (or is it a webdriver issue ?)

Selenium was updated.

Reproducible Code

from selenium import webdriver as WebDriver
from selenium.webdriver.common.by import By

webdriver = WebDriver.Chrome()
webdriver.get("https://www.consorsbank.de/home")
shadowhost = webdriver.find_element(By.XPATH,"//wb-website-header")

# works
queryselector = "arguments[0].shadowRoot.querySelectorAll('div > button#header-login-button')"
elt1 = webdriver.execute_script(f"return {queryselector}", shadowhost)

# works not any more
queryselector = "arguments[0].querySelectorAll('div > button#header-login-button')"
shadowroot = webdriver.execute_script("return arguments[0].shadowRoot", shadowhost)
try:
    elt2a = webdriver.execute_script(f"return {queryselector}", shadowroot)
except:
    pass
shadowroot2 = shadowhost.shadow_root
try:
    elt2b = webdriver.execute_script(f"return {queryselector}", shadowroot2)
except:
    pass

---

ℹ️ Last known working version: 4.25 but earlier Chrome / Chromedriver

[selenium-ci]

@dornech, thank you for creating this issue. We will troubleshoot it as soon as we can.

Selenium Triage Team: remember to follow the Triage Guide

[cgoldberg]

Your code fails for me on this line:
elt1 = webdriver.execute_script(f"return {queryselector}", shadowhost)

with this error: javascript error: Cannot read properties of null (reading 'querySelectorAll')

Your code also swallows exceptions, which isn't very helpful.

Can you please provide:

• valid reproducible code
• an explanation of what you are expecting to happen
• the full error message you receive

[dornech]

You may have the issue that the website opens one of these cookie acceptance boxes. If you debug, and click it away, the code should work. I created a scratch form my complete code and did not include this part because it is not really relevant.

The issue here is that doing the shadowroot access WITHIN the javascript (i. e. ... arguments[0].shadowRoot.querySelector ... and passing the shadowhost as argument) it works, if I do it OUTSIDE javascript (i. e. ... arguments[0].querySelector ... and passing the shadowroot directly as argument) it does not work. That is something I would not expect and I am sure it worked before.

The relevant error I described above: the shadow root object lacks an attribute id but has only an attribute _id. I guess it is linked with the JSON return from the webdriver, so it maybe an error in the Chrome webdrive ritself.

[titusfortner]

This is the recommended way to interact with Shadow DOM content. I've verified it works.

    driver = webdriver.Chrome()
    driver.get("https://www.consorsbank.de/home")

    shadow_host = driver.find_element(By.TAG_NAME, 'wb-website-header')
    shadow_root = shadow_host.shadow_root
    shadow_content = shadow_root.find_element(By.ID, 'wemo_theme')