configuration applied twice about kanshi

[tbm]

I just had:

applying profile
applying profile output 'eDP-1' on connected head 'eDP-1'
applying profile
applying profile output 'eDP-1' on connected head 'eDP-1'
configuration applied
configuration applied

but sway's log only contains it once:

2019-12-20 12:35:35 - [main.c:300] Found config * for output eDP-1 (Unknown 0x5B2D 0x00000000)

[jameswalmsley]

Did you resolve this?

What is your config file?
Which versions of wlroots / sway are you using?

[tbm]

Config:

{
	output eDP-1 enable
}

{
	output eDP-1 disable
	output "Samsung Electric Company SyncMaster HVGQ901399" mode 1680x1050 position 0,0
}

sway 1.2 and wlroots 0.7.0

[jameswalmsley]

I'll create a similar setup with my system and see what happens.

[zsugabubus]

It happens because after applying a profile, the display configuration will potentially change, and whenever the display configuration changes(*), kanshi tries finding a matching profile for the new configuration. But at the second run (hopefully) nothing will change and that's why there is two profile applies.

*: display plugged in/out, resolution, refresh rate...

[simoneruffini]

What do you mean with ''will potentially change'', I mean what types of change?

I have a similar issue but for me profiles are reapplied 3 times.

In addition to that, at startup, when kanshi starts, sway logs out this:

2020-10-21 20:51:31 - [swaybg-1.0/main.c:299] Found config * for output DP-3 (Dell Inc. DELL U2414H 4VKH665N5TRS)
2020-10-21 20:51:31 - [swaybg-1.0/main.c:299] Found config * for output eDP-1 (Sharp Corporation 0x148D 0x00000000)
00:00:04.790 [ERROR] [backend/drm/drm.c:364] Failed to page-flip output 'DP-3': a page-flip is already pending
2020-10-21 20:51:34 - [swaybg-1.0/main.c:167] Destroying output eDP-1 (Sharp Corporation 0x148D 0x00000000)

[lytedev]

In order to work around this issue, I moved my exec calls into a script that sets up a lockfile. Here's my example:

# kanshi config
profile desktop-H  {
  output ...
  output ...
  exec "setup-workspaces.sh"
}

And here's setup-workspaces.sh:

# kanshi will potentially run this more than once
LOCKFILE="/tmp/kanshi-setup-workspaces.lock"
if ! (set -o noclobber; echo > "$LOCKFILE"); then exit 1; fi
touch "$LOCKFILE"
# TODO: some way to ensure the lock file is cleaned up?
# maybe check if the file is older than a minute?

# do your stuff

rm -f "$LOCKFILE"

[Gochkin]

Well I have noticed the same thing when running kanshi. For me any profile applied twice (at least this is what kanshi output suggests).