Unreliable test for typing notifications

[neiljp]

Bug description

Intermittently in PR checks - and occasionally locally - a test will fail for typing notifications. On rerunning the test, it typically passes fine. Hence this is an unreliable (flakey) test.

This appears to be in test_show_typing_notifications, based on connected discussion in the channel.

#zulip-terminal>Flakey test (typing notifications) #T1501

How is the bug triggered?

This can likely only be reproduced by repeated runs in CI or locally.

What did you expect to happen?

Ideally this test passes reproducibly on every occasion.

How are you running the application?

Please include as many of the following as possible:

• Zulip-terminal version: 0.7.0+git approximately ref 440a3fad9 eg. a specific version (0.7.0), or if running from main also ideally the git ref
• Zulip server version(s): N/A eg. Zulip Cloud, the version you are running self-hosted, or the Zulip Community server (chat.zulip.org)
• Operating system (and version): Observed locally in Debian, and on GitHub Actions runners eg. Debian Linux, Ubuntu Linux, macOS, WSL in Windows, Docker
• Python version (and implementation): Multiple, possibly more likely with PyPy eg. 3.8, 3.9, 3.10, ... (implementation is likely to be eg. CPython, or PyPy)

If possible, please provide details from the About menu: (hotkey: Meta + ?) (this can provide some of the details above)