On 25 April 2017 at 14:49, Anton Aylward
On 24/04/17 11:31 PM, Felix Miata wrote:
Stefan Seyfried composed on 2017-04-24 20:13 (UTC+0200):
Johannes Kastl compose:
And the differentiation between openSUSE (the project), Tumbleweed and Leap does no longer work if we drop the "leap"...
Seems to work well enough for Fedora...
Fedora's "releases" are much more like rolling releases than releases in the sense most of us use the term. Its releases routinely update a number of high profile packages to latest upstream releases regardless whether any carry an LTS designation. e.g. F25 came out close to 42.2 release 5 months ago with kernel 4.8.6, but now is on 4.10.10; Xorg server was 1.19.0rc, now is at 1.19.3; Plasma was 5.8.1, now 5.9.4.
Somewhere between "Indeed", and "yes-but".....
I am notionally running 42.1 In the kernel side I am at 4.10.12, xorg-x11-server-7.6, and many other packages such as Mozilla, LibreOffice and of course the photography tools I use are up there with the 'latest and greatest'. As for "plasma", well that whole thing has me confused so I'm sticking with what I understand. (I'm sure I'm about to be deluged in advice...)
For me, "42.1" is like a rolling release .... as far as the packages that matter to me are concerned[1].
Regular readers will be aware that I've pointed out my use of the Kernel_Stable repository. All these other 'latest and greatest' are also from repositories that I've configured.
I'm sure there are people who will shout me down for using this approach, but lets face it, there are many of us who configure extra repositories for a variety of reasons, and I'm sure that getting 'the latest and greatest' for application packages rates high among them. I'm pretty sure that is the case for the few other photo enthusiasts I correspond with.
To my mind this makes the ability to point at extra repositories one of the winning facilities of the mainstream versions of Linux. it's not like downloading a random package out of the Microsoft domain for Windows users, these repertoires are either maintained or you can find the owners responsible. I've often contacted repository owners and had meaningful exchanges with them. And to be quite frank they have been more responsive than the supposed "support" of the Big Name/Big Iron support that is run on a Pay-For (and pay a LOT for at that) basis. On top of that this (and the other opensuse lists I've subscribed to) have been more use to me that the Big name/Big Iron support lines were when I was in that the context to be using them.
(That major firms are slow to grok this is disappointing.)
My feelings on additional repositories are well documented. TL;DR - They are a bad idea when not done properly. I do not feel we are doing them properly right now, and I'm not sure the effort of doing them properly is justified. https://speakerdeck.com/sysrich/distribute-or-die-arguing-against-additional... My feelings on how rolling releases are one of the best things ever is also well documented, especially how rolling releases done properly (like Tumbleweed) undermine a lot of the justifications for additional repositories. https://speakerdeck.com/sysrich/fosdem-2017-how-i-learned-to-stop-worrying-a... I really think the following rule applies: "In order to be able to move ANYTHING quickly, you need to be able to move EVERYTHING quickly" or in other words, if you want to look at a 11000+ codebase, close your eyes and randomly pick one package to keep up with upstream, you need to have the tooling, skills, capabilities, and technology to potentially move all 10999 other packages in order to facilitate that upgrade. The 'stable base PLUS additional repositories' model is proven, through hard experience through Pre 2014 Tumbleweed that it just_does_not_work. And given how bloody awesome we are with OBS and additional repositories, I think this is a lesson we can probably take as fact. Leap's unique model shows you can make something work, if you are not religious about the stable base - occasional, facilitatory changes to the SLE base helps us get Leap in the shape we want it to be. But also in Leap there is the tacit understanding that sometimes we just cannot upgrade something to the latest version in a way that will work with Leap. It's a fact of life for conservative distributions. SLE's model, with modules and extensions that are more modular than what we do in Leap and Tumbleweed, does work, but has significantly stricter controls in place - very specific packages only with very specific dependency chains can be delivered safely in this way. As a result, the scope of SLE plus it's modules is a tiny fraction of Leap or Tumbleweed in comparison, so I don't think it really addresses the scope of 'getting software in the hands of users' the way you suggest we should focus on. So, in short, I'm happy you are happy with your approach. I respect anyone who supports your approach and wish them a lot of luck & patience. But I won't support them beyond that, because I think they're walking a path we've been down, screwed up, broke everything, fixed everything, and moved on from. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org