File manager nnn takes half the time taken by ls to list
wow this bottleneck always annoyed me so much
>pointlessly complex GNU software is slow
shocker
I use exa instead of ls, it has nicer output. I wonder if it is faster?
mc
Does nnn have an image previewer? like luuuug smifff has for ranger in his ebic youtube vids?
developed by pajeets and you say they are incompetent
Makes sense. ls runs at least one extra syscall per file to get file attributes so it can color and mark the output with the additional information. It looks like nnn just doesn't have any of that information available because it never fetches it.
I wonder are there a set of flags you can use with ls that would make the test fair?
Sure, sure. Fast is good for some things. But can you look at thumbnails of reaction faces and dragon drop them on the upload button?
>using typewriter emulation from 70s so you can use image-representation hack for typewriters from 80s
the only reason this is a thing is because X and GTK are fucking disaster
typewriter != teletype. Teletypes are awesome, I have arch running on mine
>faster grep
>faster ls
so when are we getting faster mlocate and man-db?
there is a performance comparison in one of the comment threads here: news.ycombinator.com
(it's also where I came across the news)
Looks like the developer is more focused on writing a tool for other developers. No image previews but it has a plugins to invoke sxiv (which can do a lot of image magic) and preview pdf as text. Works fine for me.
In the detail mode nnn is showing me all the details - attributes, file type, permissions etc.
wut
nnn is a file manager unlike ls. I discovered it today. So far, I am enjoying the speed and trying to get accustomed to its workflows. Generally I am a thunar guy but I like how responsive nnn is.
>mlocate
you can make something faster yourself
sudo find / > ~/.db/files
grep file ~/.db/files
or use ripgrep
rg file ~/.db/files
create a systemd timer that updates db once a day and it's the same as mlocate
mlocate is ancient and optimized for old hardware, nowadays it's faster to just have a textfile and parse that if you have hdd that spins above 3k rpm.
I tried it for 10 mins and I must say it's an amazing piece of software. Thanks for sharing!
For anyone who wants a glimpse, here's the list of features:
GNUToddlers BTFO
I just tried it on orange pi.
Is it supposed to be better than mc for example?
okay this is epic.
really surprised to hear this is faster. In this case your searching unsorted data, clearly you could get way faster if you were searching sorted data by doing a binary search or something similar. is there a faster way again of doing this?
I am not an mc user. But this is far better than ranger. There are a bunch of things you can do with selection, the navigate-as-you-type mode is extremely smooth. And take a look at the cool plugins - github.com
Thanks for this. One of the best tips I've gotten from Jow Forums in a while. One thing, remember to use -N with rg to remove the line numbers. I created this handy bash function to replace locate:
function locate() {rg -N "$1" ~/.file_db;}
Just use Emacs.
Also another thing I've noticed is that updating takes WAY longer this way and really reads a lot of data. I reckon updatedb from mlocate must do incremental updates by looking at timestamps etc. This is one thing that I will need to find a solution to before I completely abandon mlocate, can't have this thing straining my drive every time the cron runs.