Hacker Newsnew | past | comments | ask | show | jobs | submit | kurante's commentslogin

Have you seen GhostCell[1]? Seems like this could be a solution to your problem.

[1]: https://plv.mpi-sws.org/rustbelt/ghostcell/


Yes. There's an implementation at https://github.com/matthieu-m/ghost-cell

Not clear why it never caught on.

There have been many attempts to solve the Rust back reference problem, but nothing has become popular.


The qcell crate is perhaps the most popular implementation of GhostCell-like patterns. But the ergonomics is a bit of a challenge still.


Right. The whole problem with all this is ergonomics, from the point of view of programmers who don't want to obsess over ownership and type theory. We sort of know how to make this work. It works fine with enough Rc/Weak/etc. But it's a huge pain.

I appreciate people arguing over this. It helps. We've seen proposals from people who are too much into the type theory and not enough into ease of use. I used to do proof of correctness work, where the problem is that proof of correctness people are too into the formalism and not enough into killing bugs.


This isn't a problem. The article says that this is only if you're looking to trade-off performance for less noise:

> In other words, we would only recommend upgrading to the NF-A12x25 G2 if you seek to lower noise levels as much as possible and if you are willing to sacrifice the maximum performance headroom in worst-case scenarios that the G1 HS-PWM fan provides.


That's for switching to the NF-A12x25 G2 fan (from the G1), which has a lower max RPM. The improved side panel appears to be a strict improvement.


It doesn't say that:

> In addition to redesigning and testing the Noctua fan grill, we also evaluated various other scenarios. These included replacing the NF-A12x25 with its G2 variant and incorporating an additional 8cm fan for exhaust purposes.


This is a bad faith comment in so many layers:

- The provided reason given was due to user accessibility concerns complicated by what likely is a breaking change.

- Even if you don't agree with the claim, a reserved name isn't unreasonable at all. Not to make a standard of GitHub, but the `admin` username is reserved there too.

- Dismissing an entire product based on a single non-critical technical limitation while simultaneously not contributing to the solution (unless you have a different username there, happy to be corrected) is fundamentally toxic.

- All the while conflating two separate products (Anubis and Forgejo) that aren't related at all.

- And that Anubis offers a non-anime girl solution, and is MIT Licensed if you really don't care for supporting the author.

I'm not going to prod at the "and everything else" part either.


As long as they don't use `#[tokio::main]` or any other attribute that wraps main into an async function!


There is more to North America than the United States.


And he got 71M votes, not even half of the US population.


What do folks think about authentik[0]?

I tried to set up Keycloak but after fiddling with it for awhile before giving up and trying something else. It felt really weird that I was just extracting a tar and running a jar instead of some pre-packaged solution, but that might just be me.

authentik was pretty easy to set up for my homelab, but maybe I'm missing something given all the positive recommendations for Keycloak?

[0]: https://goauthentik.io/


Authentik dev here, AMA


Using Authentik as a part of my selfhosted setup, mostly positive things to say. I tried with Keycloak first but had too much trouble getting the Docker image to work, so switched to Authentik.

I also checked out some other options along the way, and ultimately realized that pretty much all of the options come with enterprise-oriented features that are just added complexity for the self-hosting use case.

Ultimately, I've gotten at least somewhat familiar with all the complexities of Authentik, so I'd have a hard time switching off. Would definitely love to see a solution geared towards selfhosting that's more barebones, though.


I set up Keycloak using Docker and it was very simple to do.

I did not really try authentik yet since all the advanced features I needed worked with Keycloak, but I do have it running in a container to play with at some point in time.


AFAIK it has a bus factor of one, keep that in mind if you are going to build anything "serious" on top of it. I found it easier to configure than Keycloak, it lets you write short Python snippets for custom authorization logic, but it's about as heavy on resources as Keycloak. I hope he develops the project into a successful business, for users' sake if nothing else.


I was considering using authentik, but I'm not very keen towards having a Django application taking over SSO authentication.


why is that? too heavy or is there another reason?


Yes, Django is a quite some decently sized CMS framework, with it's occasional security quirks and attack surface. I expected it to be some lightweight go application to be honest, based on their domain.


Weissman scores are effectively fictitious and meaningless for any real application. Why are you looking for one instead of benchmarking data?


You're responding to a low-effort joke based on a plot line in HBO's Silicon Valley TV show.


I'm aware of the reference, but I'd rather try and inform on the off-hand chance that they're serious about the question. I figured that there's little harm in making sure, other than looking a little silly :)


Python has never followed semver. It's saner to treat minor releases as semver-major, and patches as semver-minor.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: