News for hackers. Stuff that matters.
Re:Real World Haskell by anentropic on Friday November 06, 2020 @09:01AM (#25007625)
Step 1: install nix

hmm, no thanks

Re:Real World Haskell by _query on Friday November 06, 2020 @10:30AM (#25008285)
We did a blog post a while ago sharing why we decided to use nix for all package management aspects of the framework. Take a look: https://ihp.digitallyinduced.com/blog/2020-07-22-why-ihp-is-...
Re:Real World Haskell by anentropic on Monday November 09, 2020 @02:19PM (#25038613)
From the blog post...

> Additionally Nix is very hard to install. This is a stumbling block for a lot of people. Especially when using certain macOS systems you cannot even install nix at all. Luckily most of the nix problems, like the macOS issues will be solved with time.

Great choice

Re:Real World Haskell by justmyname on Friday November 06, 2020 @09:46AM (#25007921)
Install nix? Already done. What's the second step?
Re:Real World Haskell by wtetzner on Friday November 06, 2020 @10:02AM (#25008047)
Why is installing nix any worse than installing some language-specific package manager?
Re:Real World Haskell by whateveracct on Friday November 06, 2020 @10:37AM (#25008347)
Nix causes a visceral negative reaction in some people.
Re:Real World Haskell by theCodeStig on Friday November 06, 2020 @11:49PM (#25013277)
Why?
Re:Real World Haskell by otabdeveloper4 on Saturday November 07, 2020 @03:45AM (#25013879)
It's a complete end-to-end solution with an opinionated approach and unstructured documentation.

That is, the learning curve is very steep.

That said, for new projects I'm doing packaging and containerization in nix, because I don't have the time or mental energy to deal with half-assed solutions anymore.

Re:Real World Haskell by whateveracct on Saturday November 07, 2020 @10:03PM (#25022452)
> I don't have the time or mental energy to deal with half-assed solutions anymore.

Exactly. I'd rather solve a problem in Nix then try some working, but half-baked alternative.

Re:Real World Haskell by labawi on Sunday November 08, 2020 @06:35AM (#25025052)
I looked at nix a while ago, concluded the versioning/chesumming is - a given sum should represent kind-of the same thing, but can mean really anything.

There was even a post on how to have a stable version for a tracking branch.

Now, I wouldn't mind, if it were explained exactly what guarantees the hashing provided, but there seemed to be pervasive misconception on what it provides.