[icat-mini.sh] file manager with icat-mini.sh breaks zsh

[Ignatius18652]

I use icat-mini in my file manager (lf). one thing that is really annoying is that I HAVE to wait for the image to load in the preview before I press "w" to chmod into that directory. Otherwise, zsh (bash and dash too) BREAKS and outputs this error:

zsh: suspended (tty output) lf

I tried this in kitten icat and it didn't happen there, so I assume it is an icat-mini.sh issue.

[sergei-grechanik]

Do you have a reproducer or some instructions? lf - I'm assuming it's https://github.com/gokcehan/lf ?

Could you try commenting out these lines?

stty_orig="$(stty -g < "$response_tty")"
stty -echo < "$response_tty"
# Disable ctrl-z. Pressing ctrl-z during image uploading may cause some
# horrible issues otherwise.
stty susp undef < "$response_tty"
stty -icanon < "$response_tty"

restore_echo() {
    [ -n "$stty_orig" ] || return
    stty $stty_orig < "$response_tty"
}

trap restore_echo EXIT TERM

[Ignatius18652]

Do you have a reproducer or some instructions? lf - I'm assuming it's https://github.com/gokcehan/lf ?

Could you try commenting out these lines?

stty_orig="$(stty -g < "$response_tty")"
stty -echo < "$response_tty"

Disable ctrl-z. Pressing ctrl-z during image uploading may cause some

horrible issues otherwise.

stty susp undef < "$response_tty"
stty -icanon < "$response_tty"

restore_echo() {
[ -n "$stty_orig" ] || return
stty $stty_orig < "$response_tty"
}

trap restore_echo EXIT TERM

I removed these lines and the problem seemed to be fine. that is for disabling Ctrl+Z suspend?
yes I use lf. if you want, this is the page on how to set up previews and it includes kitty: https://github.com/gokcehan/lf/wiki/Previews

my preview script is this:

image() {
icat-mini.sh --max-cols ${w} --max-rows ${h} "$1"
exit 1
}

file="$1"
w="$2"
h="$3"

[sergei-grechanik]

Yes, it's to disable suspend by ctrl+z and to disable echo. I couldn't reproduce any "horrible issues", so I think it's fine to remove it.

I also tried to use icat-mini in lf, and the main problem I encountered was that it couldn't determine the cell size with a python script, so it tried to use the csi XTWINOPS command. The response from that command went to lf, which interpreted it as user input, resulting in weird issues.

Since reading terminal response is brittle, I decided to remove this logic, and I also added the --cell-size-fallback option.