ncda VS tig

I'm working on Ncurses based disk usage analyzer. The only way I have found to get directories' sizes is to use lstat() and get it through the st_size field. So I made a recursive function (which is named get_dir_tree()), it reads the directories content and if it's a file it uses lstat() on it, if it's a directory it calls itself recursively. However I have found that it gives larger results than other disk usage analyzers like du, ncdu and baobab and I can't find out why, so any help is really appreciated. Here is the file on Github that contains the function and it is on line 282.

Related but I use tig, a TUI, a lot to examine the state of my working tree and index and stage/unstage/reset changes piecemeal. It works great.

I'd like to plug [tig](https://github.com/jonas/tig) for those who don't use emacs. I see lazygit recommended here too, but I've been using tig for years now and love it's simplicity.

Try tig

I'm normally a CLI git (and tig) user.

Sometimes when I work in command line I use tig (https://jonas.github.io/tig/). There is also similar tool lazygit (https://github.com/jesseduffield/lazygit)

And you accidently open a git TUI

tig +My custom command to fix MR comments by quickly editing an old commit's changes at the time when that commit was created. (Like a more controlled git-absorb that explicitly selects a commit to fixup and therefor avoids rebase-conflicts when squashing)

today I looked at tig which is a nice text based GUI, and I think I will never use git log again :-)

If you are looking for more interactivity while remaining on the commandline, have you looked at Tig? Tig has a view for browsing refs, and you can sort by date.