Hello,

this is just a drive-by comment as this was linked on HN, so feel free to ignore as I'm unlikely to keep using it (my solution to this problem is regular filesystem snapshots)

I think trash info/display/flush commands could very well be valid filenames (in particular the two letter aliases, I use t* with short names a lot for temporary files)

In particular that can be bad with wildcards: for example rmt something ti will just do ti action so rmt * is likely to not delete stuff as expected. (For scripts obviously the same problem as normal rm and friends apply, one should use rm -- * or rm ./*. Note rmt does not seem to handle -- either. But for interactive use having commands prefixed with -- is more natural.)

By the way I got panic when restoring the same file multiple times and entering '..' in restore path (aren't I a good monkey):

Please enter a new absolute path to restore your element
>> /home/myuser/../
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Os { code: 13, kind: PermissionDenied, message: "Permission denied" }', src/trash_manager.rs:270:6
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

but more importantly it doesn't handle cross dev errors either (leaving aside the fact that one cannot enter a path from /...); it should probably fallback to copy:

>> /home/myuser/../../tmp
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Os { code: 18, kind: CrossesDevices, message: "Invalid cross-device link" }', src/trash_manager.rs:270:6
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

(oh, and max_size in the config doesn't seem to do what I think it should do, but I didn't check if it's actually meant to be used yet nor documented)

Anyway, nitpicks aside it's a good idea; I've seen other implementations but nothing seems to take so good luck :)

Removing a file from a different device (ie. ramdisk) causes a crash when calling the fs::rename function:

thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: Os { code: 18, kind: CrossesDevices, message: "Cross-device link" }', src/trash_manager.rs:83:46

Add an option to automatically compress the items in the recycle garbage can, allowing to reduce its size.

For this, we can use the existing configuration file.

If merged — if you feel this isn't too much of a Good Thing — values from Cargo.toml like name, author, version, and description are injected into the --help text using the env!() macro.

Unfortunately, we can't do this (yet) with const values. So const USAGE becomes let usage in the process.

Hi little brother, I made this pull request to trigger a GitHub action at each tag and compile rmt.rs for macOS and Linux on many architectures. I recommend you review the GitHub action during the merge of squash my commits because I made many to test the GitHub action directly.

The output artefact:

Make the config.yml file look something like so:

compression: true

trash:
    max_size: 10000
    max_element: 10
    automatic_flush:
      time: something 

The user should be able to set the cron either from a file, or via a cli option

Add the encryption of the data in the recycle garbage can if the config file asks for it. https://github.com/AmineZouitine/rmt.rs/blob/main/config/config.yml Use: If you sell your computer but you forget to delete the recycle garbage can, the items can not be read

File: https://github.com/AmineZouitine/rmt.rs/blob/main/src/data_manager.rs

Some sql queries are not optimized, example :

pub fn find_trash_item_by_id(connection: &Connection, is_test: bool, id: i32) -> TrashItem {
    find_all_trash_items(connection, is_test)
        .into_iter()
        .find(|trash_item| trash_item.id == id)
        .unwrap()
}

While a where could directly give us the item we want, without retrieving all of them

Another example in the file https://github.com/AmineZouitine/rmt.rs/blob/main/src/input_manager.rs line 29

        data_manager::find_all_trash_items(connection, is_test).len()

While a "get_element_count" function by spanking a select COUNT(*) could be faster

Add an option to automatically compress the items in the recycle garbage can, allowing to reduce its size.

For this, we can use the existing configuration file.

+ cargo build --release
    Updating crates.io index
error: failed to select a version for the requirement `clap = "^4.0.13"`
candidate versions found which didn't match: 3.2.22, 3.2.21, 3.2.20, ...
location searched: crates.io index

rmt 0.1.7 linux x86_64 cargo 1.59.0 rustc 1.59.0