Simplify the idle timeout logic & don't short-circuit in handle_msg_send_datagram

[matheus23]

Intended for merging into #3676

• Makes idle_timeout always be a time::Sleep that is .reset once there is any activity. I find this much easier to reason about.
• Fixes a bug where handle_msg_send_datagram short-circuits when sending on one path fails. Instead, it should try all paths, even if some of them fail.

[github-actions]

Documentation for this PR has been generated and is available at: https://n0-computer.github.io/iroh/pr/3678/docs/iroh/

Last updated: 2025-11-18T17:19:44Z

[dignifiedquire]

how about using the std lib? https://doc.rust-lang.org/std/pin/macro.pin.html

[matheus23]

Fair. I wanted to match the rest of the code's style.

[matheus23]

Err.. Seems like that macro actually does something different. It doesn't seem to be a drop-in replacement for neither tokio::pin! nor n0_future::pin! in this situation.

[flub]

let mut idle_timeout = std::pin::pin!(time::sleep(...)) is the std version. I don't really know which I prefer, this version existed in n0_future so I assumed we liked it for some reason?

[github-actions]

Netsim report & logs for this PR have been generated and is available at: LOGS
This report will remain available for 3 days.

Last updated for commit: 0481183

[flub]

I prefer this version too.

[matheus23]

Not sure why netsim fails now :/ I've tried to spelunking in the logs, but couldn't find anything easily. I thought we'd fixed all netsim issues in this stack of PRs. :/ TBD.

[Frando]

Weird, it seems to complete successfully yet the logs say the process completed with error code 1. @Arqu any ideas?

[matheus23]

This log seems to come from chuck's util.py file, IIUC. And it seems like it defaults to exit code 1, if no exit code is provided?
Looking at the conditions when this function is called, it seems like it's called in eject, which is only called in main.py when monitor_short_processes returns any process errors, which can basically happen if the process returns a non-0 exit code or times out. However, these errors are printed, but I can't find a single "Process timeout" or "Process failed" in any log file. And as far as I can tell, there's no other place that eject, and thus transitively, logs_on_error are called.

help