FS#62510 - [tmux] 2.9-1: coredump at start

Attached to Project: Community Packages
Opened by ilf (ilf) - Tuesday, 30 April 2019, 09:52 GMT
Last edited by Christian Hesse (eworm) - Friday, 17 May 2019, 15:52 GMT
Task Type Bug Report
Category Packages
Status Closed
Assigned To Christian Hesse (eworm)
Architecture x86_64
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

tmux 2.9-1 curedumps at every start, downgrade to 2.8-1 makes it work as expected again:

% sudo coredumpctl info 811
PID: 811 (tmux: server)
UID: 1234 (user)
GID: 1234 (user)
Signal: 11 (SEGV)
Timestamp: Tue 2019-04-30 10:30:36 CEST (1h 15min ago)
Command Line: tmux new-session -d -s name
Executable: /usr/bin/tmux
Control Group: /user.slice/user-1234.slice/session-1.scope
Unit: session-1.scope
Slice: user-1234.slice
Session: 1
Owner UID: 1234 (user)
Storage: /var/lib/systemd/coredump/core.tmux:\x20server.1234.54d03a393e694c7e971f9e4622eea7f5.811.1556613036000000.lz4
Message: Process 811 (tmux: server) of user 1234 dumped core.

Stack trace of thread 811:
#0 0x000055cff01ecb21 n/a (tmux)
#1 0x000055cff01d429b n/a (tmux)
#2 0x000055cff01d4587 n/a (tmux)
#3 0x000055cff01d12a9 n/a (tmux)
#4 0x000055cff0202776 n/a (tmux)
#5 0x000055cff01f7182 n/a (tmux)
#6 0x000055cff0202e74 n/a (tmux)
#7 0x000055cff01c8098 n/a (tmux)
#8 0x000055cff01c44e5 n/a (tmux)
#9 0x00007f976003ace3 __libc_start_main (libc.so.6)
#10 0x000055cff01c468e n/a (tmux)
This task depends upon

Closed by  Christian Hesse (eworm)
Friday, 17 May 2019, 15:52 GMT
Reason for closing:  Works for me
Comment by Christian Hesse (eworm) - Wednesday, 01 May 2019, 20:32 GMT
Is this fixed with tmux 2.9_a-1?
Comment by ilf (ilf) - Wednesday, 01 May 2019, 22:45 GMT
It seems fixed, yes.

Seems weird, since the upstream changelog doesn't read like relating to this. https://github.com/tmux/tmux/blob/master/CHANGES#L32
Comment by ilf (ilf) - Tuesday, 14 May 2019, 20:09 GMT
  • Field changed: Percent Complete (100% → 0%)
Nope, after a reboot, it appears again in 2.9_a-1:

% sudo coredumpctl info 1010 2> foo
PID: 1010 (tmux: server)
UID: 1234 (user)
GID: 1234 (user)
Signal: 11 (SEGV)
Timestamp: Mon 2019-05-06 16:36:36 CEST (3min 39s ago)
Command Line: tmux new-session -d -s main
Executable: /usr/bin/tmux
Control Group: /user.slice/user-1234.slice/session-1.scope
Unit: session-1.scope
Slice: user-1234.slice
Session: 1
Owner UID: 1234 (user)
Storage: /var/lib/systemd/coredump/core.tmux:\x20server.1234.e201410ec8b2445fb1717938a3132b69.1010.1557153396000000.lz4
Message: Process 1010 (tmux: server) of user 1234 dumped core.

Stack trace of thread 1010:
#0 0x000055680ead1b21 n/a (tmux)
#1 0x000055680eab929b n/a (tmux)
#2 0x000055680eab9587 n/a (tmux)

Again, fixed by downgrade to 2.8-1.
Comment by Christian Hesse (eworm) - Wednesday, 15 May 2019, 19:49 GMT
I have not seen this myself... No idea how to reproduce.
Possibly you should report upstream.
Comment by Christian Hesse (eworm) - Friday, 17 May 2019, 12:26 GMT
Is this fixed with tmux-2.9_a-2?
Comment by ilf (ilf) - Friday, 17 May 2019, 15:31 GMT
No. I finally found some time to debug this. It was my tmux.conf

The options "window-status-current-bg" and "window-status-current-fg" worked in 2.8, but are invalid in 2.9. With an adjusted config, 2.9 works nicely, thanks.

I still think upstream should provide a clear error message, and not coredump.
Comment by Christian Hesse (eworm) - Friday, 17 May 2019, 15:52 GMT
This should be fixed now:
https://github.com/tmux/tmux/issues/1736

If you have a reproducible issue (a segmentation fault is always an issue) you should still report upstream.

Closing this for now.

Loading...