[BUG] The --dev mode spawns more servers without clearing previous ones.

[sansyrox]

Description

The --dev mode is going in infinite recursion when a file is changed in the directory.

To check, you can do
lsof -i tcp:5000

Expected Behavior

Handle a base case that might have been missed here.

Environment:

• Host OS: every

---

[Optional] Additional Context

---

[guilefoylegaurav]

Hey @sansyrox I am interested in taking this up. Can you kindly elaborate more on this?

• What kind of result am I expected to see with --dev and lsof -i tcp:5000?
• It seems that hot reload is not really working on my system? I ran the base_routes.py with the --dev flag and cannot see changes being reflecting in the browser immediately when I alter anything in the integration_tests directory?

[sansyrox]

Hi @guilefoylegaurav ,

Firstly, thank you for your interest.

You will notice changes when you change integration_tests/base_routes.py file.

When you type lsof -i tcp:5000 , you should not see any output after the dev server has been killed. However, if you run pytest integration_tests, you will see many processes that have to be killed manually.

[guilefoylegaurav]

Hello @sansyrox

However, if you run pytest integration_tests, you will see many processes that have to be killed manually

• This problem is coming up because in the teardown in conftest.py, process.terminate() has been used which does not kill the child processes spawned

• Replacing process.terminate() with os.killpg(os.getpgid(process.pid), signal.SIGKILL) would solve the problem

• Also the pytest fixtures inside configtest.py have to be scoped to session level

Please feel free to let me know what your thoughts are regarding this

[sansyrox]

@guilefoylegaurav , great spot. I had fixed this issue thrice and it still appeared, so I assumed it would be the same this time too. 🤣

Do you want to create a PR for the same?

[guilefoylegaurav]

Sure! If you say so, I'm ready to raise a PR

[sansyrox]

@guilefoylegaurav , perfect then :D