| Age | Commit message (Collapse) | Author |
|---|
|
I don't want the machine to suspend after some time period, instead I want to
manually turn off the monitors.
|
|
I think this is the only "supported" nixos version now. But in any case scipy
seemed to be broken on the older version, and I couldn't build my code that uses
llm. Also, this allows me to get rid of the bild.os-unstable thing for
Beryllium, which was just a sitting timebomb of breaking stuff.
There are a lot of changes here because ruff updated to the very latest, and it
changed some minor lint things. Also with the new nixos I get a proper cgit
module, and some other breaking changes needed fixing.
|
|
This superceedes exllama and tabbyAPI which I could never get working fully.
Unfortunately I had to switch to NixOS unstable to get all the Go builder stuff
to work, so this is a cause of yet another version drift, but I guess it's
inevitable and I should just learn to mitigate it with my nixpkgs shenanigans.
|
|
This brings a bunch of improvements. I got rid of some custom packages, I can
now build exllama without using a non-default cuda version. Oh yeah and I get to
use GHC 9.6.2 now, a huge upgrade from 9.4. Unfortunately I also updated ormolu
and some unrelated formatting changed, but that's life I guess.
|
|
This change was motivated by my testing of tabbyAPI. I kept doing like
`nix-build -A pkgs.tabbyAPI` and I thought, can't bild just do this? So I wrote
a file called TabbyAPI.nix with the following contents::
{ bild }: bild.pkgs.tabbyAPI
and it worked, I just needed this change to Bild.hs to supply the `bild`
argument. The benefit of using bild here is that I can get the logging,
concurrency settings, and linking to _/nix etc all by default. Plus, using a
standalone nix file like TabbyAPI.nix might be a good way to pin some package
in the build system and make sure it continues to build, test, and so on.
Also, thie means I don't sprinkle relative paths to the Bild.nix library
throughout the repo, which is bad practice anyway.
Re: explicitly exposing refernces to stable: This keeps things a bit more tidy
and less confusing when working on the nix library.
|
|
This patch does a few things:
1. Switches from nixpkgs-unstable to nixos-unstable{,-small}, simply because
nixpkgs-unstable is not in cache.nixos.org, but nixos-unstable is, and -small
is the same but requires all tests to pass. So we should prefer
nixos-unstable-small, whenever possible.
2. Reorganizes the nixpkgs import code such that Nixpkgs.nix returns an attrset
of all the nixpkgs that I want to use, rather than putting other nixpkgs
branches into the main one as an overlay. This is much simpler and explicit,
but it meant I had to change a lot of usages throughtout the nix codebase.
3. As a consequence of 2, moves the overlays into separate files so they can be
re-used across nixpkgs branches.
|
|
This seems to be a bug in nixos, and disabling it is easy enough and doesn't
seem to break anything.
https://github.com/NixOS/nixpkgs/issues/180175
|
|
nixfmt is the soon-to-be official formatter for Nix code, as per the NixOS
GitHub group. So I figure I should just adopt it without worrying too much about
the specifics of the formatting. I just formatted everything in one go, hence
the huge diff, oh well.
|
|
|
|
This should probably be enabled everywhere... oh well.
|
|
I couldn't get my monitor to work properly in xmonad, so decided to use gnome
for a while because I'll be away and will be using it headless from my macbook
anyway.
|
|
Even though I have 64 cpus, when I use them all for compilation, the UI still
lags, so that's annoying. Until I start using this for inference full-time,
offload the UI stuff to the GPU.
|
|
Just for my daily usage.
|
|
I finally got everything setup for the new dev machine, but I ran into a
networking problem: I can't tell my home router to expose the ssh port 22 to
multiple hosts. I could have made beryllium use a different port, but instead I
decided to use tailscale, and this seems to work well. I still don't have
hostname routing working, but maybe that's a simple config in tailscale
somewhere.
Eventually I will get all intra-networking stuff to use a vpn, but for now just
using it for beryllium is fine.
|
|
Previously, if there was a problem with the inputs and bild failed to
determine the namespace, 'fromPath' would return 'Nothing' and then
'catMaybes' would drop the error-causing input altogether. In the one
time that I had a bad input, this made debugging incredibly difficult.
It's always a bad idea to swallow errors silently, so instead lets just
kill the program if we have bad inputs.
|
|
Lots of changes here but the code is much improved. The nix code is clearer and
structured better.
The Haskell code improved in response to the nix changes. I needed to use a
qualified path instead of the abspath because the BIZ_ROOT changes based on
whether bild runs in nix or runs in the user environment.
Rather than passing every argument into Builder.nix, now I just pass the json
from bild and deconstruct it in nix. This is obviously a much better design and
it only came to be after sleeping on it the other night.
|
|
|
|
|
|
Mostly thid required packaging up some deps, but also had to recompile
stuff with cuda support.
|
|
|
|
|
|
This works when I route from lithium, including with 'dig', but when I try to
'dig @lithium router.home' from helium, for example, it times out. So my thought
is that the firewall is blocking, but that doesn't seem to be the problem. So
maybe my router is doing something? Hopefully when I migrate this to my APU
router this will all just work, but idk.
|
|
|
|
I don't like TODOs in my codebase, I'd rather keep them in org files. Eventually
I need a linter that prevents all TODOs from getting into code.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Rebuilt email server, started wireguard setup.
|
|
This also makes some changes to the build tooling to clean up the environment a
bit, and get us closer to 'bild -s'.
|
|
|
|
|
|
|
|
Getting me closer to the latest GHC. This release also includes my own packages
that I submitted some time ago.
GHCJS is not present in 21.05 for some reason, but I think it's back in master,
so I might do another upgrade soon, but for now I just disabled my GHCJS
support. I'm not really using it anyway.
I also had to bring it string-quote, update nixos-mailserver, and a few other
things.
|
|
|
|
|
|
|
|
I also upstreamed this to nixpkgs-dev
|
|
My router's DNS service likes to die, then I can't lookup any names, so let's
just use the public ones.
|
|
|
|
Every key is just a new line in the $USER.pub file. This is not automatically
reflected to gitolite, which uses a separate config, so I'll need to come up
with a way to replace gitolite someday.
|
|
|
|
|
|
Now bild knows how to determine between modules that require ghcjs and ghc. It
also knows what *not* to build, meaning it won't try to build non-buildable nix
targets, for example (unfortunately this is just hardcoded for now), but it also
won't build scm or py targets that I haven't implemented yet. It just silently
fails, which is fine, because it means I can do `bild **/*` and everything just
works.
Of course, if I want to build scm code then I will have to implement that, but
that's not a priority right now.
|
|
I had to refactor Biz/Bild/Rules.nix. I also had to checkin my patched
hoogle.nix file, but I also upstreamed the patch to nixpkgs-dev so it shouldn't
stick around for too long.
|
|
This includes deployment and implementation.
As part of sprint-49, here are the startup progress questions:
- Are you on track?
- Yes? I'm making progress toward a proper launch.
- Are you launched?
- No
- How many weeks to launch?
- I would say 4 but it's probably more like 8
- How many (prospective) users have you talked to in the last week?
- 2, Kyle and his manager, see below
- What have you learned from them?
- Kyle thought the metrics were interesting.
- His manager thought the metrics were kinda useful but didn't think they
really helped people ship higher quality code faster. So that's the rub: I
have to show how this can make devs ship higher quality code faster; or,
develop a set of features that improve those things.
- Kyle pointed out that the clustering feature of devalloc will find optimal
pairings *and* identify team silos that could be improved, so that's
important to remember and might be a good angle in the future.
- On a scale of 1-10, what is your morale?
- 6 maybe
- What most improved your primary metric?
- Well I was able to deploy something within in the week, whereas before I had
zero deploys per week. So that's an improvement.
- What is your biggest obstacle?
- Finding customers to talk to.
- Also the thing isn't really built yet, I just have a python script. I need
to build the real SaaS product
- What are your top 1-3 goals for next week?
- Find a single customer I can work with on an ongoing basis
- I should ask around my network to see if I have any second-order
connections that would be willing to work with me (Asher, Chad,
previous bosses, etc)
- Build out the front-end of the website (it's very simple, would just need a
basic miso module and deployment)
- Figure out how to connect/auth to the Github API so I can start building the
SaaS version of the product
Some user feedback from my friend Kyle. This comes from his engineering manager:
> "Looks neat. If it were priced low enough I could see using it to run reports
> as part of an overall package. A lot of those metrics don't matter too much to
> me as a manager though A lot of these code quality tools are handy info but I
> don't feel like they make people ship code any faster or any higher quality
> Things like CodeClimate work well for Jrs though to avoid obvious static type
> mistakes"
Kyle provided some additional comments:
> he might have been an unusual case. Jared's not big into metrics, Pivotal
> Tracker point estimates, or things like that... He's far more into qualitative
> feedback, like retrospectives and 1:1s
>
> I think it's definitely neat data! I certainly like the collaboration analysis
>
> It's interesting, we recently had a pair where two devs didn't work well
> together, that could be represented here. Though, we didn't want to avoid
> having them work together, we wanted them to find a work style that worked for
> both of them
And that's a good point: devalloc will find optimal pairings *and* points where
you could improve team cohesiveness.
|
|
|
|
|
