I’m hoping NixOS figures out these inevitable growing pains. The problem they’re having has a ton to do with flakes and Eelco not wanting to accept the community’s pushback on this feature. So, he implemented the feature in his “upstream” project.

I use it all the time but I’m told it is unstable.

Now do Netanyahu

Personally, I’d love to see this feature.

Right now, I build Haskell using cabal because I found the IOHK Haskell.nix implementation broken when I reached for it for a project with pkg-config dependencies.

Similar deal with purs-nix when Purescript rewrote spago and broke purs-nix.

It would be nice to finally build these two languages in Nix again with that lovely determinism I’ve grown so comfortable with.

Found myself muttering, “OP should try Purescript” to myself with every point the author made.

Welcome!

I highly recommend forking an advanced config and refactoring it to do what you want.

I mostly agree but from the article, it’s clear that he only hatched his little stupid scheme when he got demoted.

Undoubtedly.

Still, he will go down in history both for his stupidity AND boldness.

Dude is an absolute legend!

Thanks! It’s funny because I’ve chosen the absolute most obscure languages and frameworks (other than postgresql). I’m seeing zero jobs available out there for them…but I just can’t bring myself to learn JS, Rust, or any other popular tech stack over Haskell and Purescript…and I’d hate to work in any other stack, TBH.

I’m basically unemployable but happy! ;)

I made a community for my “firehose” of links. Some people seem to hate it and one guy even called me a bot.

But I don’t care. It’s for me.

Maybe try that.

I’m attempting to teach myself full stack web dev using my preferred tooling (Purescript front end using Deku and Hyrule for modernized FRP, Haskell Servant back end talking to a PostgreSQL server with Nix flakes gluing at all together) by building a web app for managing cannabis dispensary inventory.

Lately, I’m finally at the point where I realize/experiment with making the app domain-agnostic. To do that, I’m attempting to extend lambdabuffers to programmatically generate my Types (and instances eventually) so I can make my whole app (front end Types, back end Types, and SQL schema) more generic in order to generate exactly the domain specific app that I currently have.

Here’s the lambdabuffer I wrote to describe my Types (for example): https://github.com/harryprayiv/cheeblr/blob/delete/backend/codegen/Inventory.lbf

And here’s my most active branch at the moment: https://github.com/harryprayiv/cheeblr/tree/delete

I also stopped working on a previous, ambitious project aimed at building a decentralized fantasy betting baseball DApp until some of the tech I’m waiting on (crypto oracles) matures enough: https://github.com/cardanonix/pelotero-engine

I welcome any and all critiques and assistance in either of these 100% FOSS projects.

deleted by creator

Thanks for the update. I did read down the page and saw that. The dev looked really receptive, cooperative, and pro about it.

Nixpkgs has been working on packaging it for more than 6 months now. That is a damning timeframe for both Zen and nixpkgs in IMO equal parts. In Zen’s case, that means that they are packaging their app in a very unconventional way and lagging with security updates, breaking standards (which usually means they are amateurs). In nixpkgs case, that means nixpkgs needs better

buildMozillaMach

support.

I used it in an appImage and came away unimpressed. Using it in a tiling WM, I loved the minimalism….but I just wish it had better security and rendered webpages better. On Brave, websites literally look sharper than on Zen.

Edit: most of my concerns were addressed by the dev themself just down the page a bit.

sephene

deleted by creator