And do what when tmp gets full and services & applications start crashing?
When someone implements that undo, someone's going to need to write a tool that really does remove files, really. Will the next guy wonder why really-rm doesn't have undo built in?
You could do what is more or less an industry standard when it comes to this and warn the user that there's not enough room to allow an undo when deleting said file(s) and prompt if they want to delete it permanently instead.
I am similarly not worried on my Linux system, which also has frequent snapshots turned on. Instead I worry about bugs in btrfs, but that is another story.
And do what when tmp gets full and services & applications start crashing?
When someone implements that undo, someone's going to need to write a tool that really does remove files, really. Will the next guy wonder why really-rm doesn't have undo built in?