Consider other backend strategies

[ngirard]

Lolcate currently stores its data as a lz4-compressed list of path names, and recreates it each time lolcate --update is run. It's as simple as you can get.

Although it works well enough for my taste, I'd be glad to consider alternatives.

[dufferzafar]

I didn't want to create a separate issue for this, but I've found https://github.com/ozeidan/gosearch
Which seems to use "fanotify" - a Linux API that I was unaware of. It might be worth a shot to see if we can truly achieve real-time db update (and what the associated perf cost might be)

I'm aware of Everything (a Windows only tool) that used an NTFS specific API to update the index in near real-time.

[ohmree]

There's also Deepin's anything which I believe uses a kernel module and is meant to be similar to Everything (I've read so on Reddit, can't read the README since it's in Chinese).

[Lxtharia]

Hello, I hope there is still some live in this project, as it seems a bit dead super-stable.

I'd love to have an option to "watch for changes" that automatically updates (single entries in) the database when a file got created/moved/deleted, implemented using the inotify (or fanotify? i don't know the difference tbh) API.

I'd like to point to a small project kidex which is basically the counterpart to lolcate: It indexes files and directories but without a database, however, it sets up inotify watchers and updates the index on changes.

So for a backend, it needs some kind of database that can easily be updated (which i assume is more difficult with a compressed file).

I really want to try to implement this, but I'd be glad to hear any thoughts on this :)