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

The current "stable" distribution of Debian is version 13, codenamed trixie. It was initially released as version 13.0 on August 9th, 2025 and its latest update, version 13.3, was released on January 10th, 2026.

So as of today the latest "stable" release of Debian is a month old.

By contrast the last stable release of Fedora is Fedora 43, released on October 28, 2025 which four months old at this point.

Really once you get software that works all of this is pointless anyway, you have working software and you update once every year or so, or when you find you need to.

When you "need" to update is so personal that it cannot be predicted, but your FUD about Debian being universally old and outdated is clearly misleading at best and deliberately misleading at worst.


You are getting too worked up about this, not to mention cherry-picking.

Debian Trixie, to my knowledge, comes with Linux kernel 6.12 LTS. Many people with more modern hardware want the most modern Linux kernel -- currently 6.18 -- to support their devices. There are also countless stabilization patches (I heard some of my acquaintances praising their Linux kernel upgrades as finally giving them access to all features of various Bluetooth periphery but did not ask for details).

Having a modern kernel is important. With Debian though, it's a friction.

Can it still be done? Sure, or at least I hope so as I want to repurpose my gaming machine as a remote worker / station and the only viable choice inside WSL2 is Debian. I do hope I can somehow make Debian install a 6.18 kernel.

Furthermore, you putting the word "need" in quotes implies non-determinism or even capriciousness -- those two cannot be further from the truth.

Arch and Fedora can't come to WSL2 soon enough.

...and none of that is even touching on the issue of much older versions of all software in there. I want the latest Neovim, for example. For objective developer experience reasons.

Debian stable is for purists or server admins. Not for users.


> You are getting too worked up about this

No. I just see the same person in this discussion making multiple posts saying "Fedora is modern, fedora is good, stable Debian is broken, old, and wrong".

Of course my reply is a little mechanical and biased because I'm refuting a strawman.

Suggesting that Debian's stable release is no good for users, when I'm sat here using it, and many many other people do so is crazy hyperbole!

Anyway I guess arguing further is pointless.


Maybe you can show me that person and their claims so we can work with them?

Because I'm not that person.

Sure I said users and not programmers. Sue me.

I was criticizing Debian's model. I'll be getting Arch or a derivative on my main machine but for WSL2 (secondary machine that is for now stuck on Windows) I don't have much choice so I'll have to work with a distro where I'll have to actively work against how it normally operates. I'll handle it, but it doesn't need to be that way.


> I do hope I can somehow make Debian install a 6.18 kernel.

There’s the backports repository.

https://backports.debian.org/


So you lose the stable and have to deal with terminal... Or just use Fedora.

You don’t lose stable. It will only install the package you select and deps.

Also the terminal is the main interface for Linux and the BSDs. Why does having to learn it is a negative? A computer is not a toy. You don’t drive a truck with no training.


Or just understand that Debian stable can be moved to Debian testing (or even Debian unstable if even 2 weeks is too long) trivially. The best decision that Debian has ever made is not to distribute or advocate for testing as a rolling distribution, because if you're too ignorant to change your repo to testing, you're really too ignorant to be using testing.

Admitting that getting 6.18 on Debian is some sort of insurmountable mountain is not something I would do in public while trying to show off my expertise. I'm not running it, because I don't need a kernel that's been out for 5 minutes and offers me nothing that can't wait a month or two. I'm running what's current on testing, which is 6.17.13. It's about a minute of work to switch to testing. I run stable on all my servers, and testing on my laptops, it is a triviality. But to all you bleeding edge software people, it's somehow rocket surgery.

> Many people with more modern hardware want the most modern Linux kernel

To run the latest version of Progress Quest. Need biggest number available.

> Arch and Fedora can't come to WSL2 soon enough.

So, it's really still Windows, then. I assume you've moved from spending years ranting about how Linux people were purist server admins and Windows was for users and just worked, and now you've chosen the same posture after being pushed out of Windows.

> Debian stable is for purists or server admins. Not for users.

You're not a typical user. Most users want a functional computer, not the largest numbers they can find.


>Admitting that getting 6.18 on Debian is some sort of insurmountable mountain is not something I would do in public while trying to show off my expertise.

I genuinely don't care to show off expertise. I just want a distro that works.


I'm really not sure what made you so rude but I'm not participating. You're intentionally misrepresenting because I didn't say even one thing of those you so criticize, yet have the gall to speak about showing something in public.

All the best.


The Kernel was released in 2024...

Have you used Fedora? After using Fedora I was actually offended how bad Windows was and how bad Debian family was.

Its the best OS I've used in my lifetime by what feels like an order of magnitude.

My only terminal commands were to unblock some minecraft ports for my kid. You won't find a Ubuntu/Debian user with that experience.


> you update once every year or so

We're so fucked from a security perspective.


I had meant to say that you update the system when security updates come out, automatically.

But upgrade to the next stable release once a year or so, when one becomes available. (Be a point release, or a real update.)


I read a story recently, which to be honest sounded fake, but it involved a guy donating a book to the children's hospital once every week.

That helped one person, chosen at random, in a way that is simple to repeat and scale.


It's a business park "out of the way", so it's not really prime real estate for local people at the best of times.

Then build houses. Giving away real estate for basically 10 more jobs and a highly automated facility is bad bargaining.

I think it was being purchased

Back in the day I used livejournal and for a couple of years in a row I setup a matchmaking site that paired users up.

You'd login to my site and see a list of all the blogs you followed, then you could nominate five of them as people you were interested in.

If they did the same, you'd both get a notification.

It was a cute system and because it was restricted to selecting only from people you already followed it was nice and local. The code was released at the time, but has now become lost in the winds.

I could almost imagine setting it up again for instagram, facebook, or similar, but .. getting users would be hard I imagine, and I'm sure the companies would try to sue or prohibit it.


GitLab is no improvement over github, their features are frequently half-baked, their site is slow, and outages are just as common.

I used to like Gitlab, and I've self-hosted enterprise versions of both github and gitlab, and strongly believe migration from one of them to the other for "improved reliability" will be utterly underwhelming and pointless.

Gitlab used to be able to take the high-ground due to the open-core model, but these days I'm not even sure if that makes an appreciable difference.


This pretty much, and it's also more expensive.

If you're considering moving away from github due to problems with reliability/outages, then any migration to gitlab will not make you happy.

Thanks for the heads up.

Yeah here in Finland the archive site seems to come and go on a monthly basis.

I suspect, for many, that implementing a forth is more interesting than using a forth.

Once you start writing really complex programs the system gets painful and hard. But trivial things are easy, and the consistency is so appealing.


It is the bootstrap that makes it interesting.

Creating the required primitives in Assembly, and then the remaining userspace out from them.

Afterwards it is programming like most languages.

I have done it with Lisps though.

Also on 8 bit home computers it provided the feeling to be coding close to Assembly while being close enough to BASIC as high level language.


  > I think you're the one who missed the point
Yes, I would like to know more ..

Ha, caught that reference. That brought back memories.

I'm doing my part!

Only oddity was the reference to the "router cache". I agree if your browser tried to lookup example.com the local cache would be used, but then it would be the system's configured DNS server - and that would most likely be an ISP, rather than your local router.

(Assuming a typical home connection, your router is _probably_ not a DNS server with local cache, it probably is a DHCP server which will hand out the upstream/ISPs' nameservers.)


I think this is probably quite dependent on what’s normal for ISPs in the region. In the UK for example, every ISP router I’ve had runs a DNS server and it’s that which is given out via DHCP. It then forwards onto the ISPs DNS platform.

In Scotland I was with Telewest, then Virgin, and my memory is always that the DHCP pushed out the external IP of the ISP's DNS servers.

Nowadays I'm in Finland and definitely the router runs no DNS service, the DHCP service advertises the ISP resolvers.

Probably depends on the region/ISP I guess, but I had no expectation that it would be the more common option.


American here, most of ISPs here do it as well. With modern router hardware, there is plenty of hardware available to run tiny DNS server that caches and forwards all requests to ISP upstream. Memory overhead is probably about 50MB and CPU overhead is trivial, probably .1% or less.

I would argue the contrary - most home routers are running a DNS server of some kind. They forward to upstream, but will resolve local names like your printer and whatnot.

dnsmasq is the defacto tool on these embedded devices for dhcp+dns. probably a billion deployments. it's up there with sqlite for most used tech.


IIRC a resolver is what people would think of as a DNS server only it's not an authority for any domains. Like you said, they're used to get load off of authoritative servers and are very common. I think dnsmasq is mentioned explicitly in the O'Reilly locust book but it's been a while.

My parents are with Bell (the biggest ISP in Canada) and use the Bell Gigahub (Router/AP/Switch in one). It does have a DNS cache and the its set as the DNS resolver in their DHCP configuration.

The system's configured DNS resolver is usually your router.

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

Search: