Unable to install any mods in CachyOS (Linux)

[4RCH1V3]

I tried installing BCML with the following method:

1. I'm using Hyprland with the env = QTWEBENGINE_DISABLE_SANDBOX,1 to prevent the white screen bug through my hyprland config.
2. pyenv with python 3.8 for the virtualenv.
3. And then pip install bcml==3.10.4
4. When I try to install any mods, it would give out this error. This appears no matter the python nor BCML version. I can confirm this still happens with python 3.10.19 using the latest BCML.

Traceback (most recent call last):
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/install.py", line 295, in install_mod
    tmp_dir = open_mod(mod)
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/install.py", line 99, in open_mod
    raise Exception(
Exception: No files were extracted. This may be because of an invalid or corrupted download.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/_api.py", line 33, in status_run
    data = func(*args, **kwargs)
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/install.py", line 264, in do_and_refresh
    res = func(*args, **kwargs)
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/_api.py", line 380, in install_mod
    mods = [
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/_api.py", line 381, in <listcomp>
    install.install_mod(
  File "/home/akita/.pyenv/versions/3.8.10/envs/bcml_env/lib/python3.8/site-packages/bcml/install.py", line 312, in install_mod
    raise util.InstallError(err) from err
bcml.util.InstallError: An error occured when installing your mod. No files were extracted. This may be because of an invalid or corrupted download.
Your mod is being removed, and no changes have been made.

[GingerAvalanche]

I dunno. I can't recreate it on Arch. Try using https://github.com/GingerAvalanche/ukmm instead

[NiceneNerd]

Could be worth toggling the system 7z setting.

[ppaul12]

@4RCH1V3 I was having the same issue on Ubuntu and this line is the cause.
It seems that without assigning TemporaryDirectory() to a variable, it will be deleted immediately. To solve this issue only, it is feasible to change this line into

temp = TemporaryDirectory()
tmpdir = temp.name

I was trying to use the converter, and the return value of this method caused another error inconvert_bnp. At this line, mod points to a invalid path. Since open_mod only returns a string, the temporary directory was collected right after open_mod ends.

And finally, I replace the original line (line 72 of install.py) with

tmpdir = path.with_name(path.stem)

which saves the unzipped .bnp file at the same place with the same name, and it solves the problem.

(But for my cases, I couldn't use BCML since its ability of converting mods is limited.)

Hope this can solve your problem.

[GingerAvalanche]

It seems that without assigning TemporaryDirectory() to a variable, it will be deleted immediately.

This is intended - that object doesn’t need to exist, since all we require is its name. The name is a string, which should retain the reference from tempdir and outlive the TemporaryDirectory object that is destructed at the end of that line. This also holds true for the return value: the reference held by tempdir is kept in the register and handed off to the calling function.

The disappearing TemporaryDirectory shouldn’t be a problem unless the tempfile package is non-Pythonic. Looking into the tempfile package, it’s pure Python and is Pythonic - when the object is destructed, the created folder is deleted, along with its contents, but we only want the name, which is created as a Python-allocated string, and will survive as long as the Python reference counter is greater than 0.

If the folder exists after the tempdir assignment line, we delete it anyway, rendering the survival of the object beyond that line superfluous: the name reference is held by the variable, ensuring it survives in either case, and the folder itself is destroyed in either case.

All these things taken together, I’m a bit stumped as to what your suggestion is meant to achieve, unless Python on Debian is… not Pythonic :P And I don’t know what it would solve for the OP, since they’re Arch-based, which I’ve already tested in my own computer.