Skip to content

fix: Normalize timeout of 0 to be "none" instead of instant termination#77

Open
Cruel wants to merge 1 commit intosteel-dev:mainfrom
Cruel:fix-timeout
Open

fix: Normalize timeout of 0 to be "none" instead of instant termination#77
Cruel wants to merge 1 commit intosteel-dev:mainfrom
Cruel:fix-timeout

Conversation

@Cruel
Copy link
Copy Markdown

@Cruel Cruel commented Apr 12, 2026

This closes #76

Connections instantly close when using local hosted instances due to timeout being 0.
The value 0 should instead be interpreted as "no timeout" instead of "instant timeout."

@Cruel
Copy link
Copy Markdown
Author

Cruel commented Apr 12, 2026

The fact that 0 is a bad default needs to be a addressed too. From my look at the api server source, it always sends back 0 for timeout. And --session-timeout seems to do nothing on the server, as it's only used by cli daemon.

I'm personally making some changes for this (namely adding a default timeout I can put in the config.json file). Not sure how maintainers want it set up, but it's definitely not good the way it is now.

@junhsss
Copy link
Copy Markdown
Contributor

junhsss commented Apr 13, 2026

LGTM, thanks! Just noting the continue on 0 subtly falls through to the next key instead of returning immediately, but shouldn't matter in practice 👍

@Cruel
Copy link
Copy Markdown
Author

Cruel commented Apr 13, 2026

Yeah. The continue behavior is so that it checks all the keys just in case one of them is greater than 0.

Honestly, I don't understand the purpose/origin of the different keys, so not sure if this is even necessary or desirable. 🤷‍♂️

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

browser start fails on self-hosted: timeout:0 causes immediate daemon shutdown

2 participants