[🐛 Bug]: Deadlock in Selenium Java bindings

[asolntsev]

What happened?

My Java test opens a browser and hangs forever.

How can we reproduce the issue?

I managed to create a simple reproducible example: https://github.com/asolntsev/selenium-deadlock

When running `mvn test`, it hangs immediately.

A sample thread dump is stored [here](https://github.com/asolntsev/selenium-deadlock/blob/main/thread-dump.txt).
It contains two threads blocking each other: "main" #1 and "CDP Connection" #57.

Relevant log output

https://github.com/asolntsev/selenium-deadlock/blob/main/thread-dump.txt

Operating System

MacOS (no matter)

Selenium version

4.11.0 - 4.9.0 (probably older versions too)

What are the browser(s) and version(s) where you see this issue?

Chrome 116, Edge 115 (versions don't really matter)

What are the browser driver(s) and version(s) where you see this issue?

automatically downloaded by SeleniumManager

Are you using Selenium Grid?

no

[github-actions]

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

---

Info for maintainers

Triage this issue by using labels.

If information is missing, add a helpful comment and then I-issue-template label.

If the issue is a question, add the I-question label.

If the issue is valid but there is no time to troubleshoot it, consider adding the help wanted label.

If the issue requires changes or fixes from an external project (e.g., ChromeDriver, GeckoDriver, MSEdgeDriver, W3C), add the applicable G-* label, and it will provide the correct link and auto-close the issue.

After troubleshooting the issue, please add the R-awaiting answer label.

Thank you!

[joerg1985]

I had a look at this and i could not find the root cause for this on the first look.
But i have done a fix to the code, the listeners where added and called multiple times for each event.

4641ff1

[diemol]

@asolntsev does this fix the issue? You can test it with the Nightly release. https://www.selenium.dev/downloads/#nightly

[asolntsev]

@diemol @joerg1985 No, it doesn't help. I updated my sample project to run on latest Selenium 4.12.0-SNAPSHOT:
https://github.com/asolntsev/selenium-deadlock

Just clone it and run mvn test. The deadlock still happens. :(

[joerg1985]

Sorry i did not post an update here... i had to undo the change as the CI was failing.
I had a closer look at the implementation and there are multiple things to fix:

1. NetworkInterceptor.close does set a flag in Network, all following uses of NetworkInterceptor will not longer log errors because of this.
2. NetworkInterceptor.prepareToInterceptTraffic does register the callbacks multiple times.
   The events are processed multiple times and raise a lot of warning logs.
3. The Network will block the thread handling the request until the response is comming to replace the response.
   This is not needed we should be able to instantly send fullfilRequest instead.
   PS:
4. The original request was not intercepted, there was only the response replaced.

A first fix is in work at https://github.com/joerg1985/selenium/tree/interceptor