Ahoj,
I noticed that on the AUR (https://wiki.archlinux.org/title/Arch_User_Repository) some
-nosystemd packages, that were made specifically for Artix, are flagged to be deleted because they depend on some package that is not awailable for Arch Linux and would conflict with Arch Linux'
systemd.
I got deletion requests for the packages
- libratbag-nosystemd-git (https://aur.archlinux.org/pkgbase/libratbag-nosystemd-git/) and
- miraclecast-nosystemd-git (https://aur.archlinux.org/pkgbase/miraclecast-nosystemd-git/)
with the following comment:
So, it seems the AUR does not accept packages that are Artix-specific in a way that they cannot be installed on an Arch system.
So, do we have or can set up something like a "Artix User Repository" (AxUR), so that it is easy for users to submit their Artix specific packages as easy as to submit to the AUR?
I also read in the above comment from "MarsSeed (https://aur.archlinux.org/account/MarsSeed)" that "
Artix [...] their repos are open to hosting even -git packages.".
Does Artix has repos where users can upload packages in an AUR manner?
Regards!
I also recently encountered it with noudev packages
I don't know for certain, but I think anyone can flag packages - that might not represent "official" policy, it could just be that user trying to cause trouble. An Arch / systemd fan applied for the Artix wikipedia page to be deleted for some reason a while ago, but it wasn't deleted after the request was reviewed. So don't fold yet, argue your case and see what the decision is first. After all, Artix devs and users often contribute in ways that help Arch too, it isn't just a one way relationship.
Artix has no user accessible repos as far as I know from previous discussions - you could make a PKGBUILD and host it on your own git* type site and tell people about it here I suppose. Or possibly someone would add your package to the repos by request.
Incidentally, the source code for the AUR is opensource, it would be possible to use this to make a clone website if it was really required, so long as it didn't have too many packages or too much user traffic then it could potentially be hosted by free online server space providers.
Space/ traffic I can provide, but I would not dare to maintain the infrastructure. Maybe the people power to properly maintain the infrastructure would be more the bottleneck?
And then we would need Artix AUR wrappers which also include those repositories; so that stuff like
yay -Ss <searchstring> and
yay -Su works also with the AxUR.
Regards!
Playing devils advocate you can't blame them.
It's the "Arch User repository" not the "Any-old-arch-alike User repository"
The idea is that if a package is on the AUR it's dependencies will be available in the Arch repos or the AUR.
An Artix User Repository would be nice but any AUR (AxUR) helpers would need modifying if AxUR was to be of any more use than just hosting the build files on github or the like.
I did download the AUR source code when this subject came up before out of general interest, it seems to be constructed so you can easily create your own instance by adding your own URL's, and then updated as new versions are released. I expect this facility is mainly used for development and testing. So it wouldn't be as difficult as it might seem. It would be easier still to use the existing AUR though! For AUR helpers, if it was a complete clone of the AUR then it might take little more than adding the new URL to search - if not, yes, a much more difficult problem.
This reasoning is bullshit. Arch ships with systemd, but you can do whatever you want with it after you install it. This includes removing systemd and replacing it with whatever you like. By this logic elogind, seatd and other alternatives should be completely removed not just from the AUR but from the Arch repos too. By this logic, Arch should force developers to rewrite wayland compositors so that seatd isn't needed because they already have logind. What is it about these RTFM Arch people that stops them from reading their own f*cking manual?
The arch crowd have always been myway or the highway nasties. Deviation from the one true faith is not allowed.
The OP specifically stated "some -nosystemd packages, that were made specifically for Artix"
Why should a different Distro have to host Artix user PKGBUILD's ?
If you want to start from a base Arch install and then create a few AUR PKGBUILD's to replace systemd with another init, which deal with the needed replaces= & conflicts=, that is a separate argument.
But this is about packages which do not and are not intended to work on an Arch install.
I cannot follow you with this argument.
Arch's
extra/seatd (https://aur.archlinux.org/packages/elogind-git) does not
conflict with any package in Arch's
core repository, and
aur/elogind-git (https://aur.archlinux.org/packages/elogind-git) has a pending request, maybe deletion?
So it looks that things are consistent and your argument here does not buy.
But in general, I agree with you, that
should be supported by Arch.
I
assume that Arch assumes (and requires, to be an "Arch" system) that all packages from the
core repository are installed always. (I faintly remember having read something somewhere, that packages from
core are always assumed to be installed and thus should not be mentioned in the
depends or
makedepends arrays of
PKGBUILDs.)
I also cannot follow that argument.
The logic I cited is that packages that
conflict (directly or indirectly) which packages in the
core repository are not allowed. Packages that need extra packages that duplicate functionality of
core packages, but do not conflict with them, do not fall under this.
Regards!
Where
exactly is that stated? Their precious wiki makes no mention of it.
If an update to the PKGBUILD would make it appropriate for the AUR then the person who submitted the deletion request is not following their own rules. There is no section for implications, inferences, or speculation about the intent of the package in their precious wiki.
Edit: I want to be clear that I am being deliberately pedantic the way I expect to be treated in the Arch forums - after
years of being told to RTFM I will absolutely jump at the chance to impugn an Arch user who doesn't follow the rules
@andyscott - Would I prefer it if Arch had never switched to systemd ? Yes.
- Given the switch would I prefer it if they'd made it easy to use alternative inits ? Yes.
1. They did
2. They didn't.
It's history. It's why we are here.
Not worth getting angry about.
And as it's their AUR they can run it how they see fit.
I think we're going to have to agree to disagree on whether history should evoke an emotional response. Although while we're on the subject I think we should be accurate that there was no "they" - the decision to switch to systemd was made by a single person without consulting the community. At least Debian had a damn vote.
It is the Arch User Repository and they can and do run it how they see fit. It's precisely why they have the AUR submission guidelines, and it's why they have the AUR maintainer guidelines. The deletion request does not follow those guidelines. What was the point of writing all that up and harassing people about it if they're just going to do whatever they want on a whim?
Actually, I was asking back via email
and in that email I also mentioned the discussion here for reference.
It was the argument given in the deletion request message.
That's fair, I realize I was asking some pointed questions and I don't expect you to defend the person that wants your packages deleted. In the meantime I'll take some deep breaths. I don't know... if I'm repeatedly told to RTFM don't be surprised when I memorize it then quote it back ;D (again not directed at you personally).
I don't expect this answer to change anytime soon. We're still working on improving our build process/infrastructure, which is more important.
libratbag is in Arch's extra repo, making it a good candidate for inclusion in Artix's world repo should a maintainer wish to do so.
I do not want to defend anyone; with regards to persons I am neutral.
I only want to have clarity about the arguments and their scope.
I see that the argument of the person who wants my package deleted seems not to be backed up with the Arch Wiki, but much of the discussion here I was racting upon was based on the arguments the person gave -- and so I tried to clarify that scope and how far this argument reaches (or rather, does not reach).
I did not understand anything targeted at me.
I find the whole situation totally annoying, yes, since the AUR is a convenient way to maintain packages and get some search, dependency resolution and update automatic (with wrappers like
yay).
I am curoius what the reaction to my question
back to MarsSeed and the AUR requests list is.
Well I'm glad you don't feel like I was too aggressive. I'm curious to see what the response is as well. It's not like there's never been a package in the AUR that displaced packages in core or even the base group. SELinux PKGBUILDs, ARM PKGBUILDs... Those can't be installed with packages from core. They can't even be used with the mainline kernels. The only difference is that the SELinux PKGBUILD is written in such a way that it handles the dependency issues by installing other packages (from the AUR again, not core). Then we have Arch ARM which is no more officially affiliated with Arch than we are to my knowledge, so to me it would require some very special mental gymnastics to say that those packages belong in the AUR when yours don't.
There could be a problem - but with MarsSeed getting a little over-enthusiastic....
https://bbs.archlinux.org/viewtopic.php?id=291449 (https://bbs.archlinux.org/viewtopic.php?id=291449)
Looks like some Arch AUR maintainers have already run into this issue! :D
Edit: It seems you won't get a reply from MarsSeed as his account was suspended, according to the AUR mailing lists.
Interesting, and I see that he spammed aur-general (https://lists.archlinux.org/hyperkitty/list/
[email protected]/thread/GOOJIO56E2GGTBPF4FHTYXSUKPBJ3BIQ/#VNB4MRHIMG4PONLJBAN6JOC7CM4MHV7Y) earlier today demanding that his account be reinstated. If nothing else the numerous complaints about him as well as Foxboron's reply on that thread helped me smile and left me feeling a little vindicated.
I emailed that person directly, too.
I now also asked for it here on the Arch forum (https://bbs.archlinux.org/viewtopic.php?pid=2143428).
Maybe that also helps to escalate the issue high enough so that clarity (regardless to which side) will be reached.
Regards!
Don't worry about escalating, that's just playing into his plan - Mars = God of war, MarsSeed = sowing the seeds of conflict, ie it's just some trouble maker, (possibly even some crazed AI bot) from another OS or distro out to cause trouble amongst Arch and derivatives. You see this from time to time, some apparently super keen volunteer worms their way in by making themselves invaluable then starts ruining things and causing trouble for whatever reason. It was only after he had been exposed on the Arch forum that he then launched his deletion requests at you, it was a final salvo because he knew he was on the way out. It's like spam / scam / junk mail - report it, ignore it, delete it, don't reply to it or click on the links! ;D
MarsSeed doesn't speak for Arch or the AUR as they suspended his account and no-one else asked for your packages to be deleted so that seems pretty clear to me.
Can we please have a reality check?
The AUR is for PKGBUILD's that are compatible with Arch.
The whole place is a mess and littered with old / unmaintained / broken PKGBUILD's
So it's not a great surprise that packages which don't belong there have been able to stay there for a long time.
It surprises me a little the ALARM packages are not allowed (ALARM is in part sanctioned by Arch. I can't find the link now but they at least have permission to use the name & logo etc) but by their guidelines the only architecture allowed is 'x86_64'
It surprises not at all that the guidelines don't specifically state that packages should be compatible with 'Arch Linux'
Because that would be considered implicit in a document called "
Arch package guidelines" which is used by
Arch User Repository Maintainers" to help decide if packages are compliant and can be included in the
Arch User Repository.
The devs here have said enough times "Artix is not Arch". Which it isn't.
So why get indignant when user packages clearly aimed at Artix and not compatible with Arch are deleted from the AUR.
The behaviour of the maintainer in question is another issue. It seems they have been ruffling feathers for quite a while but reading through the main thread about it on their mailing list it was full of 'Newspeak' and contrived drama imho. Let them get on with it!
But reading other AUR mailing list threads going back some years it appears the consensus is they only want x86_64 packages which are installable on
Arch
Do you have an alternative in mind or is it just your turn to get worked up?
I'm cool either way, I let myself get angry after all, it's just not clear to me what you're trying to accomplish by reiterating all that
It seems like a horsepill for the smal development team to swallow. Could these few packages not just be added to gremlins or something?
It so happens I have an AUR package https://aur.archlinux.org/packages/mongodb-runit it explicitly mentions artix in the info. When I suggested it after discussion in the Artix forum I was explicitly advised not to have the name mention artix.
Out of curiosity I find many AUR packages mention runit and artix see
https://aur.archlinux.org/packages?O=0&SeB=nd&K=runit&outdated=&SB=p&SO=d&PP=50&submit=Go
https://aur.archlinux.org/packages?O=0&SeB=nd&K=artix&outdated=&SB=p&SO=d&PP=50&submit=Go
so perhaps the AUR isn't as closely controlled after all.
Is it possible to install something from github with pacman?
Just to reiterate the stance, no artix specific AUR will happen any time soon, if at all.
Reason, we are fully busy with the amount of work already present, and we won't open our gitea instance for non artix team members, ie users.
An Artix AUR would be another gitea instance, we are not overstaffed to do this and maintain it.
I am aware with certain inconveniences using AUR, but it won't change the current stance.
Please, dear arch team, time to bury this myth, this is simply not true.