thread 'main' panicked at helix-core/src/transaction.rs:185:46:

[gobijan]

Summary

Hi, thanks for making helix. It's awesome and runs mostly stable. However today it crashed on me:

thread 'main' panicked at helix-core/src/transaction.rs:185:46:
internal error: entered unreachable code: (Some(Retain(18446744073709550406)))
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Reproduction Steps

I tried to switch to another open helix tab (had 3 window splits open) then it crashed. I was editing a Ruby on Rails application with ruby-lsp, solargraph and copilot as LSPs open.

Helix log

~/.cache/helix/helix.log

2024-12-09T09:42:08.335 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(429)), original request likely timed out
2024-12-09T09:42:10.587 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(430)), original request likely timed out
2024-12-09T09:44:15.720 helix_lsp::transport [ERROR] Tried sending response into a closed channel (id=Num(431)), original request likely timed out

Platform

macOS

Terminal Emulator

iTerm2 Build 3.5.10

Installation Method

brew

Helix Version

helix 24.7 (079f544)

[TornaxO7]

May I ask if you can reproduce this on the latest master?

[gobijan]

It rarely happens so I am not sure what triggers it. Maybe some language server died because of too many open files or something and then helix tries to send into a closed channel which is not handled gracefully. Did you tackle this area in the latest master? Can try to run it later but not sure if I can reproduce it.

[TornaxO7]

I didn't have any crashes of helix for a long time and I'm only using the master branch. So no, I can't provide any useful information here

[woojiq]

Have you researched github issues before creating this one?

Because I saw a bunch of such issues for the last few month (they all are fixed on the latest master).

[gobijan]

To be honest I just had a quick search on the string thread 'main' panicked at helix-core/src/transaction.rs and found a few potentially matching but with slightly different scenarios.

According to this it should have been fixed in April:
#10702

Because the latest release is from July 14 I opened the issue still.
But if it's fixed in master ofc just close it.