Investigate options for 'Daemonized' sessions

[jtrv]

I would like to investigate options for 'daemonizing' kakoune sessions through kampliment in order to improve the startup time of kamp edit.

My current proposal for this functionality would be to add the following in kampliment

• An option to behave similar to the functionality of the useGitDirSessions implementation in kks, this would also include a boolean environment variable KAMP_USE_GITDIR_SESSIONS that works in the same way.

• An option to behave similar to the functionality of the defaultSessions implementation in kks, this would also include a boolean environment variable KAMP_DEFAULT_SESSION that works in the same way.

• Alternatively a kampliment.toml etc might be useful if we intend to add more configuration options to kampliment in the future.

[jtrv]

Additionally I think KAMP_USE_GITDIR_SESSIONS and KAMP_DEFAULT_SESSION would help to daemonize other tools like kak-lsp. For example we can use these sessions with kak-lsp --daemonize. I'll see if I can learn some rust and submit a PR.

[vbauerster]

I think following script fulfills KAMP_USE_GITDIR_SESSIONS scenario:

#!/bin/sh

usage() {
    printf "Usage: %s: [-c client] args\n" $(basename $0) >&2
    exit 2
}

cflag=
while getopts 'c:' OPTION; do
    case $OPTION in
    c)
        cflag=1
        cval="$OPTARG"
        ;;
    ?)
        usage
        ;;
    esac
done
shift $(($OPTIND - 1))

[ $# -eq 0 ] && usage

directory="$(dirname $1)"

git_dir=$(git -C "$directory" rev-parse --show-toplevel 2>/dev/null)

if [ $? -eq 128 ]; then
    echo "not a git dir"
    exit 128
fi

session=$(echo "${git_dir##*/}" | sed "s/[^[:alnum:]]//g")

kak -clear
nohup kak -d -s "$session" 2>/dev/null &

if [ "$cflag" ]; then
    kamp -s "$session" -c "$cval" edit "$@"
else
    kamp -s "$session" edit "$@"
fi

[vbauerster]

I would like to investigate options for 'daemonizing' kakoune sessions through kampliment in order to improve the startup time of kamp edit.

I'm not sure if 'daemonizing' kakoune sessions will improve the startup time at all. However I decided to benchmark kamp edit with kks edit. To my surprise kks was faster. After some digging I found out that kamp is trying to auto focus corresponding client while kks is not. So I've added optional edit --focus switch in order not to focus by default.

Here is final benchmark invoking edit rc/lsp.kak. This file is about 2500K lines.

❯ hyperfine --warmup 3 -N 'kks edit rc/lsp.kak' 'kamp edit rc/lsp.kak'
Benchmark 1: kks edit rc/lsp.kak
  Time (mean ± σ):       1.6 ms ±   0.4 ms    [User: 1.5 ms, System: 0.2 ms]
  Range (min … max):     1.4 ms …  11.5 ms    797 runs

  Warning: The first benchmarking run for this command was significantly slower than the rest (3.8 ms). This could be caused by
 (filesystem) caches that were not filled until after the first run. You should consider using the '--warmup' option to fill th
ose caches before the actual benchmark. Alternatively, use the '--prepare' option to clear the caches before each timing run.

Benchmark 2: kamp edit rc/lsp.kak
  Time (mean ± σ):       1.5 ms ±   0.1 ms    [User: 1.3 ms, System: 0.2 ms]
  Range (min … max):     1.2 ms …   2.6 ms    2041 runs

  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet PC without any interferences from
other programs. It might help to use the '--warmup' or '--prepare' options.

Summary
  'kamp edit rc/lsp.kak' ran
    1.11 ± 0.27 times faster than 'kks edit rc/lsp.kak'
❯ hyperfine --warmup 3 -N 'kamp edit rc/lsp.kak' 'kks edit rc/lsp.kak'
Benchmark 1: kamp edit rc/lsp.kak
  Time (mean ± σ):       1.5 ms ±   0.1 ms    [User: 1.3 ms, System: 0.2 ms]
  Range (min … max):     1.2 ms …   5.5 ms    1416 runs

  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet PC without any interferences from
other programs. It might help to use the '--warmup' or '--prepare' options.

Benchmark 2: kks edit rc/lsp.kak
  Time (mean ± σ):       1.6 ms ±   0.1 ms    [User: 1.4 ms, System: 0.3 ms]
  Range (min … max):     1.3 ms …   3.9 ms    1808 runs

  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet PC without any interferences from
other programs. It might help to use the '--warmup' or '--prepare' options.

Summary
  'kamp edit rc/lsp.kak' ran
    1.08 ± 0.13 times faster than 'kks edit rc/lsp.kak'

~/.local/share/kak/cork/plugins/kak-lsp/repo master *1                                                                       5s