Arch-Based Manjaro Users Express Concern Over Update Strategy

[Ayceman]

Your argument falls on deaf ears.

Are you really suggesting that Antergos isn't easier or as stable as Manjaro?

Because the reality is that Antergos eclipsed Manjaro a long time ago and only people left skeptical are those who swear by Manjaro without realizing its flaws, lack of Arch compatibility, and lack of choice.

It's not my fault you haven't heard about Antergos. Maybe you just aren't imaginative enough in your choice of distros.

Antergos is stable if leave [testing] disabled. Maybe once or twice a year you have to check archlinux.org to check for a mandatory change that absolutely can't be avoided. Manjaro has to deal with the same exact changes and decide on whether they get included in an update or dist-upgrade.

Antergos IS Arch. It can't eclipse a damn thing, since it's Arch doing the eclipsing, but Manjaro has different objectives than Arch, so it can't be eclipsed by it.

And having to check is one of the things that Manjaro avoids in the stable repos. We will have to deal with it, but not the people running stable (and what does a dist-upgrade have to do with rolling releases?).

The stable repo (as I said, but the first post must not have been approved yet) gets updated normally every 2 weeks, but it's not set. It will take as long as it needs, and gtk3 updates and ISO releases tend to stretch that term. We're just about to release 0.8.10 UP13, which is 0.8.11's stable release.

[pjezek]

Really?

There is 4th option - switch to Arch and never look back

First, I use the both, Arch on desktop and TP X-61 and Manjaro on TP X-200. I use Arch for some 9 years with quite a satisfaction. On desktop, I have facing issue with Cups and Avahi dependency leading to failure of the both. Manjaro (pure without Arch repo packages) works as I expect with no issue. Well, I am more experimenting on my desktop installation but why there is no problem with Manjaro at all? I can mix packages from Manjaro repos with those from Arch repos, but why to do that when there is no practical reason to do that? Arch is more configurable and is based on skilled back-reporting of issues by users to developers and maintainers. Arch users are more less aware testers of the latest RC SW because many "stable" packages are customised git or RC versions. I wonder why the both communities doesn't cooperate in more polite, product improving and security guaranting envirnoment? No response on bugs and security issues by Manjaro developers and maintainers would be very risky for all the Manjaro community!

[TheOne]

Because people do not want the breakage and update-pace of Arch as well as the time needed to invest in order to fix things manually every time.
Because people DO NOT want Arch. Installing Arch easily - is only one side of the coin. How about a production machine without crashes or requirement for tremendous amount of effort just to keep it away from breaking?
Why not have a sane amount of reliability, freshness, easiness of installation and maintaince as well as friendly community?
For people who do not fall neither into grannies, nor in kids category.

On the 2-3 years I have been using arch I only had to re-install once because of gtk libraries giving issues, and since I have my home partition apart the re-install was pretty easy and fast, without lost.

After that re-install everything has been solid. Anyway, with other supposedly stable distros like ubuntu you have to upgrade every 6 months to get latest software and sometimes those upgrades break your system leaving you with the only option been a re-install. And a similar situation is occurring on the manjaro camp, so which are the gains of delaying stuff???

[brosis]

This is a good argument, however, I think most people who prefer stability over bleeding-edge will choose the mainstream bi-annual distros like Ubuntu, Mint, Debian, Fedora, etc.

Elaborate on "etc":
Fedora is not rolling.
Ubuntu is not rolling and has NIH issues.
Mint is not rolling.
Debian is rolling, but has systemd issues // server-folks oppositing desktop/workstation folks; as well as lacking ABS equivalent.

That said, Archlinux isn't unstable and only uses the latest stable packages by default. I haven't yet seen an Arch update bork my install and if major changes are made they are generally mentioned on the Archlinux homepage and forums with instructions for dealing with them, which everyone should check before issuing an upgrade.

Yeah, this above. I want a not-breaking machine without any checking of any news.
I am ready to pay time setbacks for this.

Also, I want a pre-built base instead of DIY distro. Top-down one, not bottom up. I have already done many bottom up systems and I don?t want to maintain the notices everywhere I touch, just on few softwares I use.

[pjezek]

arch linux is not stable itś brokes a lots of time, i give up from them because of this, when they upgrade to cinnamon 2.2 the system simply not log in. never happen to me with ubuntu, manjaro stable and testing or fedora, arch is good to learn linux and testing last things, not for working in daily baiises, i?m sorry arch users

Mm, how often you was reporting your issues with Arch to bugzilla or directly to meintainers? I have no serious issue in spite of my time by time experimenting with DE/WM and with some AUR stuff. Arch can be used on daily working base when there are no user-driven shifts between key factors and user thoroughly follow debate on issue reports. Be sure prevailing "issues" are caused by improper administrating of the system by users themselves. Your problem with Cinnamon is simply because you preffer DE with a very short history in Arch. Bleeding edge is not for everyone...

[brosis]

After that re-install everything has been solid. Anyway, with other supposedly stable distros like ubuntu you have to upgrade every 6 months to get latest software and sometimes those upgrades break your system leaving you with the only option been a re-install. And a similar situation is occurring on the manjaro camp, so which are the gains of delaying stuff???

The same that barrier option for ext4 does. Make sure stuff is not fragile by delaying the actual write into system, proving its not braking for others. Same thing Linux Mint Debian Edition tried to do. We are OKAY with delay, but its coming as one big complete non-fragile piece. And no, we don?t want to join the army of hackers by converting every single machine into "debugging console".
This is sure way path for windumbs guys look over the shoulder and laugh their socks off. This distro is for production/work/entertainment/gaming machines.

[TheOne]

Yeah, this above. I want a not-breaking machine without any checking of any news.
I am ready to pay time setbacks for this.

There is people that installs archlinux and doesn't upgrade their system unless they want the latest and greatest of newer software versions. Anyway, sticking with old software doesn't makes it stable by default, newer software versions usually ship with better features and bug fixes which you will never get without upgrading.

Also, I want a pre-built base instead of DIY distro. Top-down one, not bottom up. I have already done many bottom up systems and I don?t want to maintain the notices everywhere I touch, just on few softwares I use.

Antergos does that for you.

[zanny]

I'm an all-in Arch user. I don't even touch Antergos, and I have four PCs right now in my home running Arch in some form. Which includes an Arch server.

1. [testing] is not a catch all. Few users are on it, not all packages are being actively pounded on equally, and not all packages even enter testing when new versions release - especially when they are niche products in community or some such. Due to its reduced surface area of users, you can also get interaction bugs between software, especially when you run a lot of stuff from the AUR, that just did not turn up in testing, or were not considered important enough to hold a package back.

2. The idea behind Manjaro is great - take the best rolling release, and turn it into a staggered release product. Problem is the Manjaro team itself is catastrophically understaffed and its userbase way too small for a two week grace period to work. Additionally, having such a grace period simply does not work for security vulnerabilities, which should not be following any staggered model - if something is exploitable, the patched version needs to be pushed immediately. Again, Manjaro just does not have the size or scale to do proper package management from upstream.

3. Arch does break. Sometimes testing leaks, sometimes packages never hit testing, and like I said often not enough people are in testing to catch nascent bugs or interactions between all the software.

4. The AUR is an imperfect solution. Arch itself is still too small, and because its not a corporate business entity it does not really have the man power to pull in the majority of software users want. The AUR is a great compromise, but it remains just that - you are getting unsanctioned packages, and often a lot of them (my main desktop has 102 AUR packages, for example). It is "at your own risk", but if software you want is literally not in the community repo, it is that or the even worse option of downloading arbitrary .runs or binaries or compiling it yourself. The AUR is an amazing convenience as a result, but has none of the expected integration aspects "normal people" want from package management.

5. It is still the best implementation of Linux there is. It does not hesitate to make transitions that are necessary and proper - it switched to systemd once it had substantial benefits over sysv, even if the transition was hard. It was second to relocate binaries and libraries to singular directories in /usr, vastly simplifying package management. The AUR is the easiest resource to use to install software not in official repositories on any distro, and it is the most comprehensive. It still balances bleeding edge and stability as well as it can, where most other distros give up and end up sitting on KDE 4.10 or CUPS 1.x or Samba 3 for years on end rather than getting the breaking changes over with to have significantly better software.

If someone could do what Manjaro is concept is meant to do, that would be revolutionary. Monthly updates with constant security fixes (note, pacman has no mechanism to distinguish a "security" update and that is intentional and by its nature, so a derivative would *have to* have some utility that maintains that information outside the pacman program or just push security updates as constant regular updates users should auto-install). I've always wished Netrunner and Blue Systems could rise to the challenge - both on the Kubuntu and Manjaro fronts, where they are deriving Netrunner from both, but in theory should eclipse and consume both. It just takes more than a community effort to provide the security guarantees to make Arch "mainstream friendly" and it requires a comprehensive commitment to it.

[brosis]

Arch is not rolling Fedora. Arch [testing] users are banging their heads at bugs, rest of users get stable software.
Everyone understands hazards of 0day releases and enables [testing] repo at their own will.

Manjaro is rolling Fedora with advantage of ABS.
Also:
"Arch [testing] users are banging their heads at bugs, rest of users get stable software."
is completely breaking all laws of logic.
That you use "Arch / testing" does not mean, you actually bang your head at bugs.
It does not mean you have bugs.
It does not mean that others get stable software.
It does not mean others get stable software because Arch/testing users bang their heads at bugs.

If you use Arch/testing, then you have increased chance at software breakages, problems, deficiences.
This is multiplied by constant rolling nature. This is your choice, so its your challenge how you deal with that and how you see it integrated into your workflow.

But this is not everyones choice, claiming that is criminal.

Judging from what they do: Manjaro is about using vulnerable binaries for two weeks.

Again three logical fallacies:
1) Set-off in time does not necessary mean there is a vulnerability issue. Some less actual software may be more secure than cutting edge/git/9999
2) There is also increased chance of getting trivial bugs/problems, like segfaults in software/due to compilation in constantly rolling ecosystem. Likewise it does not guarantee that time-offset binaries are reliable/not breaking, but there is a better chance.
3) Fact of getting newer build does not mean that one becomes less vulnerable software. In fact, for security one uses old, well-tested/proven pieces of software - and even then one risks having a major unseen vulnerability, so an independent research as well as fast security patch service is needed. Yes, this basically blows away the argument that Arch is more secure. So the full sentence is "Manjaro is using vulnerable binaries for two weeks, Arch uses much more vulnerable binaries right now".

One can directly upgrade part of his system to newer version, omitting Manjaro upstream; for those in a hurry or dire need. This is possible with pacman.

[brosis]

There is people that installs archlinux and doesn't upgrade their system unless they want the latest and greatest of newer software versions. Anyway, sticking with old software doesn't makes it stable by default, newer software versions usually ship with better features and bug fixes which you will never get without upgrading.

Partial upgrades are not supported in Arch. So, you have to -Syuu the system anyway, dealing with all that is coming a top, while leaving the possibility of breakage here and there, which is fixed the next few days requiring to repeat the loop. This is the top of the constantly evolving volcano, and all I want is hot latte coffee.

Antergos does that for you.

Antergos is just an instant-o-arch. After installation, you have to face what regular Arch user faces.

This is not what Manjaro does. Manjaro promises more stability at price of delay from upstream.