In the spirit of Marie Kondo, I'm tidying up!
TL;DR:
- Prefer .envrc `use_nix` and delete all dir-locals.nix files
- Remove ~all references to <nixpkgs>, <unstable>, <depot> and prefer
referencing each with briefcase.third_party.{pkgs,unstable,depot}
- Delete nixBufferFromShell function since I was only using that in
dir-locals.nix files
Ensure that $NIX_PATH is empty using a .envrc file. nix.dev considers references
to <nixpkgs>, <briefcase>, <depot>, <unstable> to be an anti-pattern because it
hinders reproducibility by introducing an implicit environment.
This repository is chock-full of references to
<{nixpkgs,unstable,briefcase,depot}>, so right now everything is broken.
1. I should be using NixOS/nixpkgs-channels instead of NixOS/nixpkgs
2. Instead of refactoring everything, I'm supporting <unstable> and pointing it
to NixOS/nixpkgs-channels
I needed <unstable> to get a more recent version of the Data.Time Haskell
package.
Lorri does not cleanly integrate with my corporate device, which cannot run
NixOS. To expose dependencies to Emacs buffers, I will use nix-buffer.el, which
reads its values from dir-locals.nix. To easily expose dependencies from my
existing shell.nix files into dir-locals.nix, I wrote a Nix utility function.
From what I currently understand, lorri is a tool (sponsored by Target) that
uses nix and direnv to build and switch between environments quickly and
easily.
When you run `lorri init` inside of a directory, lorri creates a shell.nix and
an .envrc file. The .envrc file calls `eval "$(lorri direnv)"` and the shell.nix
calls `<nixpkgs>.mkShell`, which creates a shell environment exposing
dependencies on $PATH and environment variables. lorri uses direnv to ensure
that $PATH and the environment variables are available depending on your CWD.
lorri becomes especially powerful because of Emacs's `direnv-mode`, which
ensures that Emacs buffers can access anything exposed by direnv as well.
I still need to learn more about how lorri works and how it will affect my
workflow, but I'm enjoying what I've seen thus far, and I'm optimistic about the
road ahead.
Renaming my mono-repo briefcase.
I first introduced this commit in master, but it introduced a bug where one of
two things would happen:
1. Emacs wouldn't start and would crash X.
2. Emacs would start but my keyboard wouldn't work.
I learned some valuable debugging skills in the process. Here are some of them:
When my keyboard was broken, I wanted to control my computer using my
laptop. Thankfully this is possible by using `x2x`, which forward X events from
the SSH client to the SSH host.
```shell
> # I'm unsure if this is the *exact* command
> ssh -X desktop x2x -west :0.0
```
Git commit-local bisecting. I didn't need to do a `git bisect` because I knew
which commit introduced the bug; it was HEAD, master. But -- as you can see from
the size of this commit -- there are many changes involved. I wanted to binary
search through the changes, so I did the following workflow using `magit`:
- git reset --soft HEAD^
- git stash 1/2 of the files changed
- re-run `nix-env -f ~/briefcase/emacs -i`
- restart X session
- If the problem persists, the bug exists in the non-stashed files. Repeat the
process until you find the bug.
In my case, the bug was pretty benign. Calling `(exwm/switch "Dotfiles")` at the
bottom of `window-manager.el` was failing because "Dotfiles" is the name of a
non-existent workspace; it should've been `(exwm/switch "Briefcase")`.
There may have been more problems. I changed a few other things along the way,
including exposing the env vars BRIEFCASE to `wpcarros-emacs` inside of
`emacs/default.nix`.
The important part is that this was a valuable learning opportunity, and I'm
glad that I'm walking away from the two days of "lost productivity" feeling
actually productive.
At the moment, all of the Nix repositories that I'm consuming exist in ~. To
keep things consistent, I ran:
```shell
> hub clone nixos/nixpkgs ~/nixpkgs
```
I recently changed my hostname for my desktop and laptop from
wpcarro.lon.corp.google.com -> zeno.lon.corp.google.com
wpcarro2 -> seneca
If you're curious, the names Zeno and Seneca come from famous Stoic
philosophers. As you can see from this commit, my configuration depends on the
values of these hostnames.
Immediately impacted:
- .profile
- device.el
Not immediately impacted:
- configs/install
- configs/uninstall
- .ssh/config
- .zshrc*
* As a side note, I should stop supporting ZSH.
Using an .envrc file helps me DRY up some of my configuration. Ideally I should
only need to make changes to the .envrc file and then expect everything to work
as expected. Let's see how that goes.
Update code that depends on my mono-repo being named "mono". I've renamed it to
"universe", which explains the changes in this commit.
TODO: Merge dotfiles into universe.
When Emacs starts it's called from xsessionrc.shared, which is called outside of
direnv's .envrc scope. Because of this variables defined therein, like
ORG_DIRECTORY, are undefined and prevent Emacs from initializing.
I'm hard-coding the `org-directory` variable for now and removing references to
`(getenv "ORG_DIRECTORY")`.
Instead of keeping this in my ~/.profile, I'm going to define it in .envrc.
What I still don't know is how functions like `getenv` are supposed to interact
with direnv. I suppose maybe they aren't? Right now, when I call
`(getenv "DOTFILES")` from Emacs, it's `nil`, which I understand. Hopefully the
more I use direnv, the more reasonable expectations I'll have.
