If you're using Debian stable, hopefully you fully expect and want not to get major software updates until long after they release, in exchange for a more predictable system.
I'm excited for Plasma 6 but I'm very willing to wait for it, and stick to 5.27 as a daily driver for the next year.
They might include it. Or they might not. If they don't have time to test it, they just won't, and you may wind up with 5.27 for longer than just the next year if you're waiting for debian's stable repos.
debian's neovim is on version 0.7.2 (even in trixie/sid, you have to go to experimental to get to 0.9.5, which is the current). If there are any bugfixes between 0.7.2 and 0.9.5 that aren't security backported... too bad. You aren't getting it any time soon, because it's not landing in Trixie, and it's not guaranteed to land in whatever is after that either.
Debian's "stable" refers to "predictable" like you said. Which includes bugs being predictable. Not resolved. Predictable. And if you have a bug that crashes your system, that bug will stay there unless it's a "security" issue. Predictable crashing. NOT the "doesn't crash" that people seem to think "stable" means.
I did. Something about the combination of Debian, KDE, Wayland and nvidia drivers made the system unstable.
After googling the issue for a bit, the consensus was that that just isn't a good combination.
So I switched distros and now everything works.
In about 2 years. By that time it will actually be stable and pretty much bug free. For me Debian is the only distro that provides a reliable experience.
Well I don't actually use KDE. I just am stating what I've noticed in the past. KDE doesn't have stability as a priority so using old versions is really the only way to not have bugs.
KDE hasnt had, but 5.27 is pretty good.
Though, I know of and have reported a ton of bugs "Resolved in Plasma 6", where backports would make no sense as these architectural changes where so big.
Can confirm, after 8 years of distro hopping i stopped at Debian, Debian is for home servers and tired ass mothafuckers
I tried Debian for my very first Linux install very long ago. Its installer formatted my windows partition despite me explicitly telling it not to.
Never touched it after. Not out of resentment, but because I just don't need it for anything.
I mean, the point of Debian is stability. If I'm running Debian then I'm not even gonna want to try and install the thing until after I've seen 100 people use it. I don't think they'll be looking for it in repos.
Plasma 6 is the most tested release EVER. There where at least 5 ways to test it, there was KDE Neon and a dedicated atomic Fedora image for it.
There are many bugs only fixed in Plasma 6.
So it is debateable
Both of you misunderstand the point of Debian's stability.
When I run Debian Stable I want to be sure nothing changes about how the system works, until I have time to plan an upgrade.
So KDE6 could have literally zero bugs and it still wouldn't make sense to push it into a current Debian release, because it has new features.
I think backwards compatibility is the keyword here. That would be the biggest requirement to allow updates.
New bugs, and maybe for example new hardening policies needed, could be another one. Maybe a future firefox implements feature x and you want to / have to disable that.
But at the same time Firefox is the best example of upstream doing the versioning. They know when to freeze features and likely backport every security critical issue. Thats not the case with many other packages debian ships, where it just doesnt ship updates whatsoever.
Debian sid user here. If it appears now or in two weeks in the repo does not change anything for me as I don't depend on the changes for my workflows. For Debian Stable I actually demand them to come much later, in a mostly bugfree version. What's the rush when it probably needs more field-testing?
For Debian Stable, it should actually never come, cause what's the point of KDE on a server?
Oh how quickly people forgot about Plasma 4. When a Debian stable is released with Plasma 6 then I'll know it's ready for me. I don't rush into major KDE releases anymore.
4 is two major versions back. For this statement to be fair, you should have evaluated it against 5. (Spoiler alert: that release was super smooth)
My point was more that it was a very rough release even for a .0 and the distros should have waited a bit longer before shipping it as a default. All said I <3 KDE and have been using it since 0.9.6 I think
As said: yes, 4.0 wasn't ready. But 5.0 was, so I think it's fair to assume that they learned from their mistake and 6.0 is fine too.
There is an atomic image ready ,it still uses GDM (display managers are more complex to implement than you think) and I have the feeling it cant be the latest stuff, because it feels very incomplete.
Things like a populated dock etc are all missing, its upstream COSMIC.
Also there is no SELinux support yet, you can run it in permissive mode and should get all the errors needed to create one.
Join the Matrix Chat to discuss Fedora Cosmic.
Wasn't it just released or something? It'll be there on the next release. Maybe....
You can license anything you want. Question is whether you can afford to assert it when it comes time to.
©solidgrue@lemmyworld, 2024
Its not a very funny joke, sadly.
The FOSS community was doing fine with BSD and GNU v1 & v2, maybe the odd MIT licenses for decades. Or, decade. Or both. Or neither, if thats your bent. They were substantially similar: here's the code, this is the license. You may modify the code hut not the license, and any derivative code must contain the original license. Oh and it has to include the code. Some allowed for commercial use, some did not. Not everybody liked that.
Then one day someone, I think it was Apache, decided no these terms don't work for us, we don't like to release the code and we want commercial use and to sell support. So they cut a new license. Not everybody cared, and Apache was happy.
Likewise, GNU foundation decided they wanted to compete in commercial space and allow for commercially supported releases, allowing .COMs to make money using and supporting what used to be FOSS. GNUv3 license got issued. Most of everybody who cared didn't like that, but at least IBM, Oracle and Amazon are happy.
And then suddenly everyone was cutting their own licenses, and people who cared kinda gave up on it and went with what works.
Eventually, the people behind Wikipedia and the mediawiki software decided the People needed a license, so the Creative Commons license was born. I'm not well read in it, but I gather it us favorable for content producers, and is aaserted in "arenas of public discourse" (my own term) where content derives from individual contributors mostly as prophylaxis against trawlers, scrapers, aggregators and (to be tested,) LLMs.
Of course, asserting the license and defending against contrary use of the content is incumbent on the licensor, and must be defended in civil court. Not everybody can afford a lawyer, and EFF and the Open Source Initiative aren't rich enough to litigate everything pro bono.
Any or all of the preceding might or might not be bullshit, but its my good faith read of the license shenanigans the last 10 years as I've bothered to pay attention to them.
EFL-2.0 just to be a dork about it.
I mean I get it might be because of being afraid of it being used to train LLMs. But I doubt that it would work, either because they won't be used regardless or because of how federation works, i.e. literally it'd be more efficient if all of the known network instances' operators somehow agreed to include/Lemmy, kbin and all of the microblogging platforms that can federate with Lemmy shipped a robots.txt that blocks known AI crawlers. Probably what would be more useful would be something that e.g. Akkoma and some other AP implementers offer, i.e. message autodeletion.
Also terrible if you want to retain any anonymity even if more people did it, because of stylometry.