Timeshift corrupts all snapshots when no disk space is available

[ygerlach]

Describe the bug
I run timeshift with btrfs and had a full SSD. I opend timeshift to see if i can delete any snapshot to gain some space. Timeshift did not open. (Many ENOSPC errors in the console). I couldnt safe a log or coredump, as my drive was full. I think the final error message was, that /etc/timeshift/timeshift.json was not writeable.
While i tried to open timeshift i also entered the subvolid=0 of my btrfs. I saw all snapshots. I then noticed, that the info.json of all snapshots was gone.

I solved the problem with the little space differently (deleting docker images) and open timeshift up again. This time it worked, but all my snapshots where gone. A look into the subvolid=0 mount confirmed this. Aparently timeshift had deleted all snapshots, because they had no valid info.json anymore.

To Reproduce
Steps to reproduce the behavior:
(I havent tested how reproduceable this is)

1. Configure timeshift with btrfs
2. Create some snapshots
3. Fill the drive to the brim (ex. sudo dd if=/dev/urandom of=big.file make sure the root-reserved space is also allocated.)
4. get massive errors from all programs that the disk is full
5. try to open timeshift (crash)
6. free space
7. open timeshift again
8. snapshots gone

Expected behavior
If no space is available, timeshift should not crash, but allow the user to delete snapshots. In the best case, the notification, that shows "no space left" should have a button to open timeshift, next to the "Analyze Disk Usage" Button.

Screenshots
In that moment i was not able to store any logs, screenshots or crashdumps :c i will try to reproduce this in a vm.

System:

• Linux Distribution Name and Version: Linux Mint 21
• Desktop: Cinnamon x11
• Application Version: 24.06.3
• btrfs mode