RFE: Delay asking for a password

[gdamjan]

most screen lockers these days have an option to delay the password requirement, so that if your locker activates you have 2-3 seconds to just press a button and get your screen back without typing the password.

[emersion]

I bet this can be implemented with swayidle, or another client which would run swaylock after a timeout?

[gdamjan]

well if there was a passwordless swaylock that would work :)

or if there was a swaymsg "output * blank-and-keyboard-locked on" command ?

[emersion]

I'd like not to implement this in swaylock/sway, since this can be implemented by a third-party client.

[gdamjan]

I'd like not to implement

would you be ok with someone else implementing it?

[emersion]

Clarification: I'd like not to have this in swaylock/sway, since this can be implemented by a third-party client.

This means you can write the third-party client. ;)

[gdamjan]

I think that option is suboptimal, and probably error prone, but sigh

[gdamjan]

master...gdamjan:delay-password

there it is for anyone else

[snorkypie]

This should definitely be part of swaylock since we want to display (and abort) the actual lock screen. Feels like it would be a lot of work for a third party application to implement all of swaylocks display settings and render the exact same image before swaylock actually kicks in.

[emersion]

Nah, just open a layer shell surface that spawns swaylock after a timeout.

[drrossum]

Nah, just open a layer shell surface that spawns swaylock after a timeout.

This sounds great. Can this easily be achieved with currently available tools?

[emersion]

Could be a pretty simple C program.

[edrex]

this should be default behavior imo not something users have to manually configure. the current behavior is a usability fail.

[ddevault]

As the maintainer of the project, I disagree and will not be enabling this behavior by default.

[gdamjan]

As the maintainer of the project, I disagree and will not be enabling this behavior by default.

how about if it's optional? (my current branch has that)

[ddevault]

I don't really like it. I think that it would be worth having a more featureful fork of swaylock out there which is more open to including features like that.

[emersion]

I feel like we've already discussed about that, @gdamjan.

[mortie]

Sorry for commenting on an ancient thread, but i think it's relevant (and, as #139 shows, people are still discussing it). I just merged gdamjan/swaylock:delay-password in to swaylock-effects: mortie@c0f6b94

[sh1boot]

So I copied something like the following out of /etc/sway.conf:

exec swayidle -w \
         timeout 300 'swaylock -f -c 000000' \
         timeout 600 'swaymsg "output * dpms off"' resume 'swaymsg "output * dpms on"' \
         before-sleep 'swaylock -f -c 000000'

After reading this thread I decided the simplest thing to do was changed it thus:

exec swayidle -w \
         timeout 270 'notify-send -t 30000 "getting sleepy"' \
         timeout 300 'swaylock -f -c 000000' \
         timeout 600 'swaymsg "output * dpms off"' resume 'swaymsg "output * dpms on"' \
         before-sleep 'swaylock -f -c 000000'

I'm sure that'll work fine. I'm just too lazy to wait around to test it right now.

I think this works better than a grace period. If I see the notification then simply dismissing it is enough to reset the timeout, and in the mean time I can still see whatever I'm watching. And if I don't see the notification then it really doesn't matter what happens next.

[mensinda]

I wrote a manager script for swayidle / swaylock that manually handles the logic: https://gist.github.com/mensinda/94d0b05604bbb50f9b5bd4a5f1b48bfb

After finishing this script, a friend told me that swaylock-effects has this feature built in...