Doom Emacs meets most of my expectations of an IDE for software development and note taking. However, having to use it in a virtualized environment for a long time recently, I experienced slowdowns and high resource consumption over time. Doom does an amazing job loading only necessary features and lazily, optimizing startup speed and the overall feel. Nevertheless, the big number of packages that come bundled with it eventually make it slower with time. Many of the functionality these packages offer are now built into Emacs but cannot be easily replaced. This convinced me to start building my Emacs configs bottom-up rather than trying to make Doom faster by taking the bloat away.

Doom is complex, so many advice and macros are included that I for one could not wrap my head around many of them for modifications. More magic means less control. I decided to build my config as simple as possible, no advice, no macros, no complexity (also because I don't know how to use advice and macros).

Here I try to minimize the number of packages loaded into Emacs and constantly check the loaded ones to avoid slowdowns caused by them. As an example, I use built-in projects and tab-bar-mode instead of projectile and pers/perspective which handle most of what I expect from them alright.

You can clone this to ~/.config/emacs (or ~/.emacs.d):

mv ~/.config/emacs ~/.config/emacs.bak
git clone https://codeberg.org/haditim/dotemacs.git ~/.config/emacs

I use this configuration mainly for Python development with the following amenities:

• A custom virtual env selector which is a strong alternative to pyvenv, poetry and conda dependencies which works well over tramp with conda/poetry/python-venv environments. This can be called with my/python-venv-mode (also local leader binding , m e). The minor mode also shows env info in mode line.
• LSP support with eglot. Just install the LSP server in your environment and run eglot after setting the env (, e).
• Quick REPL interactions with my/restart-python-repl (also bound to , m r). If you do SPC u before calling the function, you can call the REPL with a custom interpreter (e.g. ipython). If you later want to restart REPL with the custom interpreter you chose before, you can call my/restart-python-repl-last-interpreter (, m R).
• Debugging using dape which is an amazing package dominating my Python interactions. Just call dape (SPC d d) and pick from available configs. I also have streamlit config as well as debupy-remote and debugpy-file-remote-forwarded-port which help with streamlit and tramp-managed projects.
• Docstring management with python-docstring and numpydoc.

As I like to be able to talk to ESP32 (and other micropython-managed micro controllers), I added some REPL interactions as I have for normal Python to the config. If you call my/start-micropython, you can select the device and send buffer or region to that device, even if it is connected to a remote machine.

I also have basic support for the following:

• C
• Typescript/Javascript (mostly via treesitter)
• Flutter/Dart (no treesitter available yet). Remember to run flutter config --disable-analytics to prevent it from dialing home!

CDLaTeX is enabled for org-mode and LaTeX buffers. Use C-c { to get completions for TeX elements.