Option to read a module playlist from a file/stdin?

[AliceLR]

Something I do a lot for getting aggregate data when testing fixes on more common module formats is I run checks on the ModLand or Mod Archive collections with a program that takes a list of newline-separated filenames. Doing a similar test with xmp --load-only would be useful to determine if anything new is rejected by libxmp after implementing a patch. If there's a way to do this with xmp without invoking multiple xmp processes, I missed it. (The ModLand 2016 dump I test with has >20k Impulse Tracker modules, for example, so while argv is fine for less common formats, it just isn't adequate in this case.)

Would a feature like this be generally useful?

[narodnik]

I'd like the ability to have a playlist. Right now I use this script:

#!/bin/bash

for i in $(fd --type f); do
    if rg $i played.txt; then
        echo "Skipping $i...";
    else
        xmp $i;
        echo $i >> played.txt
    fi
done

[iskamag]

I run "xmp ’shuf -n 100 modlist’"
You can probably save your playlist with "shuf -n 100 modlist > playlist" then run either "xmp ’cat playlist’" or "xmp ’head -n 1 playlist’ && sed -i 1d" for each line.
Scripting is awesome.

[narodnik]

I use these scripts:

#!/usr/bin/python
import os, glob, subprocess

HOME = os.environ["HOME"]
os.chdir(f"{HOME}/mus/shared/latest/")
files = glob.glob("**/*", recursive=True)
files = [f for f in files if os.path.isfile(f)]

played_files = open(f"{HOME}/mus/rec/all").read().split("\n")
files = [f for f in files if f not in played_files]
for file in files:
    with open(f"{HOME}/mus/rec/curr", "w") as f:
        f.write(file)
    process = subprocess.run(["xmp", file])
    with open(f"{HOME}/mus/rec/all", "a") as f:
        f.write(f"{file}\n")

#!/bin/sh
cat ~/mus/rec/curr >> ~/mus/rec/fav
echo >> ~/mus/rec/fav

[ClayCore]

Personally, I keep my module files organized by the tracker used, followed by the author, so I used the following command to generate the listing for all module files in subdirectories.

find . -type -f > modules.txt

I have a python script which launches xmp and shuffles it before running. You can use all the regular controls since subprocess.run waits for xmp to finish before advancing to the next track, or just ^C to kill the script.

#!/usr/bin/env python

import subprocess as sp
import shlex
import random


def main():
    fid = open("modules.txt", "r")

    li = fid.readlines()

    fid.close()

    random.shuffle(li)
    for line in li:
        path = line.strip()
        cmd = shlex.split(f'xmp -d alsa "{path}"')
        sp.run(cmd)


if __name__ == "__main__":
    main()

[AliceLR]

I was thinking the most useful thing would probably be to have each line in a playlist file contain one or more module names plus whichever command line options should be used only with those modules, but this might be too much complexity for the current implementation (and would not work with driver options, only player mode, interpolation, loop, etc.). I don't know if I want to refactor main() for 4.3.0... 😬

I also found xmp technically does support per-module configuration... reading a modules.conf file line-by-line and searching for each individual module's MD5. Ouch.