Devs as well as sysadmins will appreciate this stuff. I didn’t know about grep performing poorly with utf. A few opinions:
- homegrown dedup scripts feel less reliable than rmlint
- bash compgen will list known commands, it’s neglected because it’s weirdly named
- if you have Python, you could have magic wormhole rather than a script that self-hosts ftp. Opportunistic encryption, doesn’t get firewalled, workalike alternatives exist
- fzf improves file selection and shows a preview. Consider previewing PDFs with fzf, and use it for shell history if you can
- ncdu can cache its results in a file for later. I find this more valuable than the older ‘locate’. It can’t read this cached representation from a file descriptor. An excellent example of where named pipes make something useful
- homegrown dedup scripts feel less reliable than rmlint
- bash compgen will list known commands, it’s neglected because it’s weirdly named
- if you have Python, you could have magic wormhole rather than a script that self-hosts ftp. Opportunistic encryption, doesn’t get firewalled, workalike alternatives exist
- fzf improves file selection and shows a preview. Consider previewing PDFs with fzf, and use it for shell history if you can
- ncdu can cache its results in a file for later. I find this more valuable than the older ‘locate’. It can’t read this cached representation from a file descriptor. An excellent example of where named pipes make something useful