[BUG] Storage display in performance tab shows wrong usage (possible solution included)

[mastermach50]

Step 1. Before you submit

• I have read the above instructions and am sure that this is supposed to be posted here.
• I've successfully updated to the latest versions following the updating guide.
• I've successfully updated the system packages to the latest.
• I've ticked the checkboxes without reading their contents

Step 2. Version info

Version info

Packages: not on Arch

fatal: not a git repository: '/home/mathew/dotfiles/.config/.git'
Caelestia: not installed

Shell: version helper not available

Quickshell:
    quickshell 0.2.1, revision tag-v0.2.1, distributed by: Nixpkgs

Step 3. Describe the issue

The storage display in the performance tab shows the wrong usage stat. I found this out when my system ran out of storage but the stat still showed that there was free space left.

The correct value is seen in btop

Reminder

• I agree that it's usually impossible for others to help me without my logs.

[mastermach50]

I don't know enough qml to rewrite it but i know enough to find the issue.

lsblk provides the total size of free blocks on the disk, which is not the same as the total size of free blocks that can be used by a non-root user.

btop calculates this using <sys/statvfs.h>, the cli version of it would be the stat command. It does not provide the no of used blocks (as far as i can tell), only total blocks and usable free blocks. The total usable free space needs to be calculated as (total_blocks - non_root_usable_blocks) * 4096 (4096 being my disk's block size).

[mastermach50]

I desperately need this fixed, but I don't know qml.

Now please forgive me for this but I used claude to generate the code I want and got it working as I wanted.
https://github.com/mastermach50/caelestia-shell

Now this is veeeery janky code. Firstly because stat always needs to take a single mountpoint as input, it cannot display the usage of multiple disks at the same time. Secondly the way mounted disks are found is by looking at /proc/mounts which also lists tmpfs, ramfs etc, so the command has to manually filter out these disks. It uses one single long piped, regexed, incomprehensible command to do this which I don't like at all.

I would much rather write a simple utility some other language that lists the usage of all disks at the same time, and calculates it based on usable space left instead of total disk space left, seeing that most other utilities like df, lsblk, du use actual free space instead or usable space.

[mastermach50]

I have fixed the logic now with some hand written code that used duf instead of lsblk or stat