dotfiles

inspired by https://github.com/mathiasbynens/dotfiles

A simple setup to automate and document the setup of my development machine as much as reasonable.

Full automation doesn’t make a lot of sense as it is a lot of work, needs to be adjusted every now and then, and won’t be used much more often than every few years. The invested time just doesn’t justify.

But at least this framework gives some tools to recreate a development setup on a new machine with not too much effort.

Currently this repo is optimized for macs. But it is also possible to adjust it for usage with different Linux distributions. Even a parallel setup and synchronization between a Linux and a MacOS system proved maintainable in the past.

Concepts

• The dotfiles are stored in git. They can be checked out to any folder on the target workstation.
• Via symlinks the dotfiles are “activated”.
• The repo is copied into place via bootstrap.sh (rsync) instead of managing per-file symlinks.
• Complex configuration setups from 3rd parties are managed via git submodules.

Installation

Setup package manager and install

On MacOS we use homebrew and install:

• awk
• aws-console
• aws-google-auth
• aws-iam-authenticator
• awscli
• bash
• bash-completion
• bash-language-server
• cmake
• coreutils
• emacs-plus@29
• gettext
• git
• git-lfs
• gnu-getopt
• gnupg
• graphviz
• helm
• kubectx
• kubernetes-cli
• make
• markdown
• node
• nvm
• pandoc
• plantuml
• rbenv
• scalastyle
• sops
• terraform-ls
• the_silver_searcher
• wget
• yaml-language-server

If possible a brew list on the old machine should be used to get a list of all installed packages

Usage

Run ./update.sh from time to time

Essential Tools

bootstrap.sh

Copies the tracked configuration into $HOME using rsync.

Dry run first (recommended):

./bootstrap.sh -n

Apply the changes once the output looks right:

./bootstrap.sh

AWS

Contains credentials, needs gpg

Optional But Helpful And Fun Tools

• byobu
• thefuck
• fasd
• jq
• pass
• watch
• Ollama (local LLMs for code/chat)

Local LLM (Ollama)

Run powerful local models for coding and chat.

Quick start:

bin/ollama-install.sh --yes --start
bin/ollama-coder.sh --repl              # interactive coder chat (qwen2.5-coder:7b)
bin/ollama-coder.sh "Write a Bash script..."

Tips:

• Use -m qwen2.5-coder:14b for higher quality if you have RAM/VRAM.
• General models (non-coding): llama3.1:8b, qwen2.5:7b.
• Server host: set OLLAMA_HOST or pass --host to scripts.

brew

sdkman

https://sdkman.io/

Tools

ssh

aws

kubectl + kubectx + kubens

Not Yet Documented

gnupg

password-store