Manjaro Linux To Drop 32-bit Support

[starshipeleven]

32-bit as a PC standard is certainly dead, but this has never stopped anyone from still using it.

There is also nothing forcing anyone to support it for a couple thousand dudes worldwide that can't just let go of their obsolete stuff.

People that want to use obsolete stuff will have to go use distros that are aimed at that.

So why choose Manjaro when it's not exactly having a large user base, is cutting off support for 32-bit, doesn't very much attract but it rather looks like it's slowly dying?

I still don't understand how cutting support for an architecture not really used anymore and whose hardware is mostly obsolete giving anyone the feel that "it doesn't very much attract" or that "it looks like it's slowly dying".

[geearf]

They're doing this because Arch itself is dropping i686 support in November. And since Manjaro relies heavily on Arch packages, they kinda have no choice.

That said, there will be unofficial 32bit Arch packages available from https://archlinux32.org/. They've set up a similar infrastructure to what Arch Linux ARM has.

Exactly. Manjaro is Arch-based, that was definitely expected.

[sdack]

I still don't understand how cutting support for an architecture not really used anymore ...

I see. Well, I took you for smarter than this. Perhaps you only don't see a point for yourself in it. And I, too, carry out my garbage and don't hoard it in my basement. However it doesn't mean I get to dictate how others should deal with their existing 32-bit hardware and software. Rather do I respect that not everyone has the same options and that cutting off support for it is a loss for those who still need it and that it's only a win for those who don't. If it was always this easy then I sure wouldn't want to see Nvidia put Fortran support into LLVM *lol*. You simply got to have respect for what others need and not just when you need it, too.

[kon14]

I see. Well, I took you for smarter than this. Perhaps you only don't see a point for yourself in it. And I, too, carry out my garbage and don't hoard it in my basement. However it doesn't mean I get to dictate how others should deal with their existing 32-bit hardware and software. Rather do I respect that not everyone has the same options and that cutting off support for it is a loss for those who still need it and that it's only a win for those who don't. If it was always this easy then I sure wouldn't want to see Nvidia put Fortran support into LLVM *lol*. You simply got to have respect for what others need and not just when you need it, too.

I think you're missing some of his points here, I'll try to give my take below, but in any case I think you'd agree he's not just spouting random bs insults towards 32bit here a few others above. You'll have to give him that at least, right?

So let's consider a few things before we jump to conclusions. Whether we like it or not and whether it's going to impact things positively or negatively in the short term, 32bit support is dying and it's pretty much struggling at its death bed when it comes to pure 32bit systems and not just multilib support.
Manjaro has less man power than most distros, as do most other derivatives, and with Arch having killed 32bit installers and already deprecating i686 packages they'd have to put a lot of effort in maintaining 32bit support and providing these packages themselves. The important question to ask here is: Is that effort going to be worth it in the end?

Perhaps in doing so Manjaro devs could get some new users interested in their distro, but their goals and philosophies aren't about eventually becoming another distro for legacy users as far as I can tell. They're currently providing a nice and sleek out of the box arch experience for desktop users. Sometimes "dumb" users (exaggerating) who care mostly about simply using their os as a means, not an end, and not maintaining it. They're installing Manjaro cause it works for them, they like its eye-candy and don't wish to bother being tryhards for no good reason. They're likely to have a decent 64bit desktop anyway as I'm sure Manjaro devs would have considered their statistics before taking any such decisions.

32bit exclusive programs are still relevant ofc, whether for good reason or not, especially when it comes to proprietary software and light game ports from windows devs, but that won't prevent 32bit's decay. Now then, how soon is too soon? Most big desktop-oriented distros are either going to be dropping 32bit completely or declare it maintained, stop prioritizing it and letting its bugs prevent packages moving into stable and blocking releases and packages moving into stable branches. Again, Manjaro would either have to follow along or focus on becoming a distro targetting a legacy userbase.

Maintaining 32bit support is not that bad by itself, but you have to consider whether the time and effort spent doing that could have been spent better otherwise.
There are 2 obvious variables here: One is the amount of users requesting it, the other is how much hard it's going to do it (and it's going to be getting harder).
These are approximately known, but there's another factor too: inevitability.
So no matter how hard anyone would push for this, they'd probably already know that at some point soon all of this effort is going to end up trashed.

I know I've been repeating myself above, but I can't get my thoughts straight atm, gotta get some sleep really.
I wish Manjaro all the best, though now that they're discontinuing OpenRC and splitting support for it into another separate distro I'm losing more interest in them.

[Pentarctagon]

However it doesn't mean I get to dictate how others should deal with their existing 32-bit hardware and software.

Erm... it's their distro, they can do whatever they want with it.

[rpcameron]

This is essentially a non-news story. Arch is dropping i686 support, and since Manjaro is based on Arch, this makes perfect sense. Also, Arch stopped building i686 install images some months ago. The only surprise is that it took Manjaro so long to do the same.

32 bit support will still be available in the form of multilib packages, which are not going away.

(Long time reader, first time poster, and all that jazz ;-)

[leipero]

I don't see a problem with it, and I'd say it's about time. As long as we have multilib to support 32-bit applications (mainly games), there's really no need to be upset about it.

[starshipeleven]

However it doesn't mean I get to dictate how others should deal with their existing 32-bit hardware and software.

This is wrong. Distro maintainers get to dictate where their efforts will be better spent (i.e. will do most good to most people), random people of a non-paying userbase don't.

As others have already said above, there is an Arch 32-bit project https://archlinux32.org/ where other people who care about 32-bit decided to devote their time to that since upstream Arch maintainers will drop 32-bit support soon (which is the reason Manjaro also dropped now support, and why Antergos will also probably drop support

Really, your points have exactly 0 standing. Man up and stop thinking that everyone should pamper you and your special needs.

Rather do I respect that not everyone has the same options and that cutting off support for it is a loss for those who still need it and that it's only a win for those who don't.

This is bullshit, you can get a better 64-bit system for like 50$ or less, AND everyone knew that 32-bit was doomed since at least 5 years ago. That's why distros are starting to switch off 32-bit support.
The only reason you don't upgrade your PC is because you can't let go your stuff, which is a purely personal choice.

And I repeat, I'm not saying your personal choice is wrong per-se. I'm saying that your idea that the world should adapt to your personal choices even if they are wildly unpractical is wrong.

I mean, in Manjaro forum you see other 32-bit enthusiasts that leave posts like "thanks for everything, I'll switch to <distro of choice> I will surely come back after my current PC is dead and I'm forced to upgrade", which is what a true man would say in such situation when he acknowledges the reality and realizes that it's up to HIM to deal with his own special needs.

If it was always this easy then I sure wouldn't want to see Nvidia put Fortran support into LLVM *lol*.

Fortran isn't any less obsolete than modern C (and just like C it was updated through time), so your point here is invalid.

If you want to talk about a legacy language we can talk about Cobol (whose main reason to exist is that there are still large projects in it but the newest is more than 10 years old, newer stuff is usually in Java). You see NVIDIA adding Cobol to LLVM?

[cj.wijtmans]

This is wrong. Distro maintainers get to dictate where their efforts will be better spent (i.e. will do most good to most people), random people of a non-paying userbase don't.

As others have already said above, there is an Arch 32-bit project https://archlinux32.org/ where other people who care about 32-bit decided to devote their time to that since upstream Arch maintainers will drop 32-bit support soon (which is the reason Manjaro also dropped now support, and why Antergos will also probably drop support

Really, your points have exactly 0 standing. Man up and stop thinking that everyone should pamper you and your special needs.

This is bullshit, you can get a better 64-bit system for like 50$ or less, AND everyone knew that 32-bit was doomed since at least 5 years ago. That's why distros are starting to switch off 32-bit support.
The only reason you don't upgrade your PC is because you can't let go your stuff, which is a purely personal choice.

And I repeat, I'm not saying your personal choice is wrong per-se. I'm saying that your idea that the world should adapt to your personal choices even if they are wildly unpractical is wrong.

I mean, in Manjaro forum you see other 32-bit enthusiasts that leave posts like "thanks for everything, I'll switch to <distro of choice> I will surely come back after my current PC is dead and I'm forced to upgrade", which is what a true man would say in such situation when he acknowledges the reality and realizes that it's up to HIM to deal with his own special needs.

Fortran isn't any less obsolete than modern C (and just like C it was updated through time), so your point here is invalid.

If you want to talk about a legacy language we can talk about Cobol (whose main reason to exist is that there are still large projects in it but the newest is more than 10 years old, newer stuff is usually in Java). You see NVIDIA adding Cobol to LLVM?

You have not provided me an example yet where fortran is a better tool than C

[bug77]

I read this title and realized I didn't remember my last 32bit CPU. But it came back to me: it was an AthlonXP 2500+ (running at 3200+ speeds, of course).