Arch Linux Finally Rolling Out Glibc 2.27

[dungeon]

I'm so confused... I thought you were saying Testing is bad because of the human delay factor, and now you're saying only Stable and Sid have it?

Sid and Stable have human delay factor possibilty only. Testing have more than just human factor, it have planned delay factor, but also possibility to get stuck in time, even to be removed from repo, etc... that testing could possibly suffer a lot from a user POV

It might only looks fine if you dunno these kind of things and what happens elsewhere

Also, didn't you just say Sid doesn't have the delay factor?

Usually have no any obstacles so don't, but It have delay factor for part of the repo once Freeze is there. That is all normal and usual, human delay is something else... maintainer is not there, maybe on vacation, bad day, etc... you know

Because of these bad days and various human factors possibilities, you are better to learn compiling and packaging as you will needed that sooner or later

[schmidtbag]

Sid and Stable have human delay factor possibilty only. Testing have more than just human factor, it have planned delay factor, but also possibility to get stuck in time, even to be removed from repo, etc... that testing could possibly suffer a lot from a user POV

I have no clue where you're getting any of this from. Testing is basically just packages that trickled down from Sid, minus a few obscure ones here and there. What evidence do you have that packages get removed (where they aren't removed elsewhere) or "stuck in time" (often enough for this to be a problem worth pointing out, I'm sure it has happened at some point)? Seeing as you seem to treat Testing like the plague, have you ever actually extensively used it? Because I use it on a nearly daily basis on systems with roughly 2000 packages installed, and I've never had these problems.

[dungeon]

I have no clue where you're getting any of this from. Testing is basically just packages that trickled down from Sid, minus a few obscure ones here and there.

Yes, it is kind of recycle bin, where if you look deeper you can find garbage usually from 10 days ago, but sometimes months old ones
Sometimes some garbage is so old or broken or simply unsupported that it suddenly disappears

To me only Sid and Stable make sense to use, while Testing only when it is in Freeze.

Seeing as you seem to treat Testing like the plague, have you ever actually extensively used it?

Yes, of course i "use(d)" it but i just treat Testing for testing as that is its purpose and for pretty much nothing else

Because I use it on a nearly daily basis on systems with roughly 2000 packages installed, and I've never had these problems.

These are not "problems", these are normal things that are planned and are known to happen there What to say, less you know might be better for you - more you know, you are more ready for Sid

To be ready for rolling, one just need to read newspapers properly so in a way to not always believe exactly what has been written there... i don't recommend doing that if somebody asks me, but it is easy really

[monraaf]

People who pay attention to Arch already knew what the delay was.
https://lists.archlinux.org/pipermai...ry/029151.html
https://lists.archlinux.org/pipermai...il/029219.html

And? Debian is already on gcc-8 and glibc 2.27 by default.

Please keep in mind that while Arch is run by mostly hobbyists, lots of the core packages in Debian are actually maintained by paid developers from Canonical or other companies which then have much more resources available than Arch does. Thus, it is not surprising that Debian can quickly rebuild the whole archive for testing purposes whereas Arch cannot.

In fact, Debian has so much resources that they are actually helping Arch with their reproducible build project.

See, for example: https://jenkins.debian.net/view/rebo...r_archlinux_3/

Same goes for Fedora and openSUSE which also have paid developers working on that stuff. Thus, Arch simply doesn't have the resources to keep up. That's not a shame though with Arch being developed mostly by hobbyists.

[monraaf]

There's a huge difference betweeen pushing things in an recently unfrozen prerelease or to the unstable repo, where breakages are not only justified but expected, and pushing it to your "production-ready" branch which happens to be a rolling-release. Arch also has a experimental branch, they just took more time to test and I doubt they have the manpower of Ubuntu and Debian for that too.

The person who maintains glibc in Debian is also a very active glibc upstream developer. Thus, you have someone very competent maintaining the package.

The toolchain (openjdk, gcc, binutils) is maintained by people from Canonical who are paid for their work in Debian.

Arch doesn't have any paid developers as far as I know so and also not the same resources as Debian or Canonical it's not surprising they are having a hard time keeping up.

Really, Arch is a system mostly developed by hobbyists while Debian/Ubuntu, openSUSE and Fedora has big companies behind them who support them. Arch can't and will never be able to keep up with that.

[monraaf]

h

I have no clue where you're getting any of this from. Testing is basically just packages that trickled down from Sid, minus a few obscure ones here and there. What evidence do you have that packages get removed (where they aren't removed elsewhere) or "stuck in time" (often enough for this to be a problem worth pointing out, I'm sure it has happened at some point)? Seeing as you seem to treat Testing like the plague, have you ever actually extensively used it? Because I use it on a nearly daily basis on systems with roughly 2000 packages installed, and I've never had these problems.

The more you know: https://lists.debian.org/debian-deve.../msg00006.html

[schmidtbag]

Arch doesn't have any paid developers as far as I know so and also not the same resources as Debian or Canonical it's not surprising they are having a hard time keeping up.

Except Arch has a far more up-to-date, accurate, and complete Wiki than Debian... Also for the most part, they keep up a hell of a lot better than Debian in terms of having a good balance between cutting edge and stable (with Debian, you can only pick one or the other). Though I personally like Debian Testing, it is definitely less stable than Arch while often having more outdated packages.
Ubuntu for the most part is decent at being modern and stable, but, only if you don't tamper with the repos. But the thing about Ubuntu is it requires a lot of maintenance. Their repos change way too frequently, and their LTS stuff starts to drag behind about a year after the initial release.

The more you know: https://lists.debian.org/debian-deve.../msg00006.html

A lot of repos do that sort of cleanup to some degree... To my knowledge, even Debian Stable does that (note, I do mean the actual "stable" repo, not the current stable, like "stretch").

Testing doesn't tend to remove anything from its repo that anyone actually cares about.

[Charlie68]

I have 11 years old Sid install, but i don't use it nor update it so often - it still works fine... maybe that depends on luck

Joke aside, it is never luck. User needs to know how to manage breakages on any rolling distro, otherwise if you are not capable it is the best to not use it

I never recommend rolling distros, if someone asks me "should i use Sid?" i say No. If you don't ask me, then Yes

On rolling release distributions I thought so, but now it's a year that I use Tumbleweed and the system has never broken, it's still there and it works beautifully. This made me review my judgment on rolling releases. I do not consider myself an expert user, when I update the system I do not even watch what updates, at the end of the reboot and the system is always there.

[andrebrait]

I think you misunderstood something there... Ubuntu was not in _their_ Freeze yet at the time and since they importing from Sid new glibc (regardless that it was ready) can't be uploaded to Sid until Ubuntu stop importing from Debian Dunno how better to explain that.

Yes, I misunderstood lol
Sorry about that

But yeah. Arch and Debian are completely different. I was just coming from an angle that even if they were the same, the unstable repo on Debian is not directed towards the end-users whereas Arch's stable is. That's all.

[stefantalpalaru]

glibc-2.27 is still masked on Gentoo, due to these issue: https://bugs.gentoo.org/showdependen...ide_resolved=1