Fedora 21 Has Been Delayed By Three Weeks

[phoronix]

Phoronix: Fedora 21 Has Been Delayed By Three Weeks

Due to many of the Fedora 21 changes/features not being ready in time, the release schedule has been pushed back by three weeks...

http://www.phoronix.com/vr.php?view=MTc0ODY

[Anvil]

for Fedoras sake i hope it aint delayed any longer than 3 weeks however i wouldnt mind having a bet it will be

[justmy2cents]

for Fedoras sake i hope it aint delayed any longer than 3 weeks however i wouldnt mind having a bet it will be

why would that matter? whenever it was delayed it was for good reason and if you really want bleeding edge, you have copr repositories for all things on 20.

for bleeding edge distro, fedora was always rock stable and delays played part in that. or do you prefer shipping on date and broken?

[tessio]

It's 21 or 22 that will have a major change in the way fedora QA works?

[Delgarde]

why would that matter? whenever it was delayed it was for good reason

The problem is, Fedora releases are *always* delayed. Every single release cycle in recent memory, we've gone through the same pattern of the release being delayed a few weeks, then another few weeks after that.

Sure, it's better to get things right than to release broken and early - but something's wrong with your project management when you're targeting a six-monthly release cycle, and you're consistently a month or more late in delivering it.

[Kostas]

The problem is, Fedora releases are *always* delayed. Every single release cycle in recent memory, we've gone through the same pattern of the release being delayed a few weeks, then another few weeks after that.

Sure, it's better to get things right than to release broken and early - but something's wrong with your project management when you're targeting a six-monthly release cycle, and you're consistently a month or more late in delivering it.

Expected to jump into the topic defending F21 from the usual shitposting but I can certainly agree with this reasoning.

[Delgarde]

Expected to jump into the topic defending F21 from the usual shitposting but I can certainly agree with this reasoning.

Well, for the most part, the Fedora folks do a really good job. It's just this one area where they have problems, again and again and again...

[blackiwid]

still the best distro if you want a cutting edge but kind of stable distro.

yes archlinux is close and they have a better wiki which I use much for fedora, but some stuff redhat gets more right in my opinion that arch linux.

- they dont always wait till something is stable before they deliver it, like mesa 9.2 in f19 or with xmbc 13 from rpmfusion while it was beta.
yes u find such stuff in AUR but its a pain in the ass to compile this 2 packages yourself. and it did break for me.

- nice installer, that brings a bootable install on machines that are not so easy than mine, I just use a btrfs without partitions and made grub easily boot, that did not work on the thinkpad of my brother even the bios is set to legacy only, so no efi shit should break it. The only way or the easiest way was to install fedora copied from the harddisk over usb2 (pain) again his archlinux installation and sligthly changed the grub.cfg so it boots now.

- good working pinning with a always availible bleeding edge rolling distro (rawhide)

- commitment to free software (of course some of you will see that as disadvantage for me its a advantage)

- technical top notch distro.

And to the delay what do u miss from the next distro? gnome 3.12? there is a copr for that.
new kernels, they get updated anyway in f20.
newer mesa? u could try to pin that from rawhide.

Of course its nice to have official new gnome packages. I dont use the new gnome copr for my dad.

newer xbmc u get updated versions from rpmfusion 13.1 at the moment.

I guess mesa is the main thing, for wayland here can every delay only help to get something really useful out finaly.

[AdamW]

It's 21 or 22 that will have a major change in the way fedora QA works?

Fedora 21 is the release that incorporates all the change going on under the Fedora.next banner. Those changes are in the way the distro is organized and developed, though. Changes in the QA process are just natural consequences of that; we haven't actually set out with the explicit goal of 'major change' in how QA works, but obviously there are changes to the QA process to account for the new Products and so on.

We've kind of run into a perfect storm of three issues with the Alpha. The .next changes themselves, not surprisingly, make the process of building and testing the Alpha much more challenging than usual; we had a pretty well-oiled process for building an 'old' Fedora release because we hadn't really changed the deliverables or the processes for several releases, but with a whole new slate of deliverables it's kind of a different story. Secondly, there happen to be some messy bugs that aren't really anything to do with Fedora.next but are throwing a spanner into the release composing works; there's a weird bug with live image creation which causes live images to just fail to build sometimes, something that causes large chunks of the filesystem on live images to have the wrong SELinux labels, and a few other things along those lines. And thirdly, the big Fedora conference, Flock, is coming up soon: the original F21 schedule was designed to have the Alpha release hit right before Flock. That's why we just jumped to a big delay: if we can't make the release right before Flock, then we basically burn a couple of weeks just on getting to Flock, doing Flock, and then getting out again before anyone actually gets back to working on the Alpha. So we can't just slip a week.

[AdamW]

The problem is, Fedora releases are *always* delayed. Every single release cycle in recent memory, we've gone through the same pattern of the release being delayed a few weeks, then another few weeks after that.

Sure, it's better to get things right than to release broken and early - but something's wrong with your project management when you're targeting a six-monthly release cycle, and you're consistently a month or more late in delivering it.

It's actually every release cycle ever, not just in recent memory, I don't think there's ever been a release without a slip.

It kinda becomes a philosophical question: if everyone knows the schedule isn't going to be hit perfectly, is it really a problem? I kinda don't honestly know the answer. So far as QA goes, we just aim to get the release up to the required quality level as fast as possible. It'd be nice to actually hit the schedule sometime, sure, but...it doesn't seem like it actually matters a whole lot when we're two weeks behind.

F21 is kinda a different kettle of fish, though, with all the major changes going on. It's still kinda too early to get a read on whether we're going to have serious issues getting all the work done in a reasonable timeframe or not. But it's fun finding out!