groovy2shoes has quit [Read error: Connection reset by peer]
groovy2shoes has joined #crux
groovy2shoes has quit [Ping timeout: 268 seconds]
groovy2shoes has joined #crux
dlcusa has joined #crux
lavaball has joined #crux
zorz has joined #crux
nekobit has quit [Ping timeout: 240 seconds]
tarxvfz has joined #crux
<cruxbot>
[opt/3.7]: poppler: updated to version 24.04.0
tarxvfz has quit [Remote host closed the connection]
tarxvfz has joined #crux
tarxvfz has quit [Remote host closed the connection]
tarxvfz has joined #crux
ppetrov^ has joined #crux
<SiFuh>
I guess revdep means reverse dependencies?
<r0ni>
scrap?
<SiFuh>
Scrap?
<zorz>
Scrape :Pp
nekobit has joined #crux
<farkuhar>
SiFuh: assuming you're asking sincerely and not as an April Fool's joke, I'd proffer the expansion "review dependencies", since that's what the tool does. But the definitive answer would have to come from jw or braewoods.
<SiFuh>
Cool
humky has quit [Read error: Connection reset by peer]
remiliascarlet has quit [Remote host closed the connection]
remiliascarlet has joined #crux
dlcusa_ has joined #crux
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux
<ukky>
SiFuh: My main CRUX system runs without linux-pam since September 2023. Removing linux-pam requires patches to these packages in my setup: openssh, libcap, shadow, cifs-utils, sudo, cups, cyrus-sasl, samba, pure-ftpd.
<ppetrov^>
ukky, do you have the patches hosted somewhere?
<SiFuh>
openssh, libcap, shadow and sudo shouldn't need to be patched.
<ukky>
ppetrov^: No, as I implemented patches via proxy git feature on CRUX. But I can post my 'git diff --staged' for all three main repos and you can extract all patches you need.
<ukky>
SiFuh: openssh v9.6p1 requires a patch to build without linux-pam. Unless you are talking about some other distro.
<ppetrov^>
thanks ukky. Also, any idea if the core team will reconsider pam in core? I, personally, do not mind, just curious
<ukky>
ppetrov^: I have no idea on the team' plans with regard to pam.
marlin has joined #crux
<SiFuh>
ukky: I was talking about CRUX
<SiFuh>
ukky: Are you talking about patching the Pkgfile or patching the source?
<farkuhar>
the git tag tarball for util-linux 2.40 differed from the "release" tarball, and attempting a meson build with the latter would have failed due to a missing file under pam_lastlog2 (https://github.com/util-linux/util-linux/issues/2875). We were lucky to avoid that issue by grabbing the git tag tarball, but meson builds of util-linux#2.40 had other problems unrelated to PAM.
<jaeger>
Presumably it was an oversight, not intentional
<farkuhar>
ppetrov^: the upstream authors are responsible for keeping the ./configure script in sync with the meson.build. If they start to diverge in behaviour, then it falls on the distro packagers to decide which system is less broken, and go with that one.
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
<SiFuh>
Could always make a shadow-pam sudo-pam openssh-pam libcap-pam
<SiFuh>
For those who wanted pam
<ppetrov^>
farkuhar, crazy...
<ppetrov^>
i get the impression meson is presumably better than autotools, the latter being old or sth
<ukky>
SiFuh: Good workaround, but unfortunately more work for Tim
<farkuhar>
SiFuh: I think it was FS#1410 "flexible footprint check" where beerman expressed strong disapproval of letting our ports collections fill up with so many near-duplicates, the way the Arch AUR had been trending. "Fluid Pkgfiles" is the ideal he preferred.
<SiFuh>
But that isn't the point. Someone has decided to stray away from CRUX being a lightweight distro and decided to dump what they want on others. I thought we are suppose to add to CRUX, rather than having to take away stuff we don't want or need.
<SiFuh>
Many of us use cryptsetup. But it isn't in core and it shouldn't be in core. Because it is not essential to core.
<ukky>
SiFuh: CRUX is far away from being minimalist. But this is the best lean distro available.
<SiFuh>
'The primary focus of this distribution is keep it minimally bloated'
<farkuhar>
correction: it was FS#1576, not FS#1410. Not that it matters, since the Flyspray instance is no longer active and its database not yet converted to Gitea.
<SiFuh>
farkuhar: I needed to change all the {core,opt,xorg,contrib}.git to rsync because it was no longer connecting.
<farkuhar>
SiFuh: what was the exact error? Something about an unreadable git config under /root is the message I saw recently, but it still managed to sync just fine.
<SiFuh>
What I pasted the other day
<SiFuh>
I wonder which channel and it is difficult to grep because of so many linux-PAM rants. Hahah
<farkuhar>
Heh, at least 'ports -u' can speak git. We're still stuck with a pkgmk that relies on tarballs, and in the aftermath of the xz backdoor we might want to consider removing that restriction.
<SiFuh>
What would you consider replace it with?
<SiFuh>
I will run contrib.git again since I can't find it in the logs
<farkuhar>
Not necessarily replace pkgmk, just give it the ability to understand a git url in the source array.
<SiFuh>
I am wondering if there was a change when moving shit around and I didn't notice it
<farkuhar>
SiFuh: I was referring to /usr/bin/pkgmk seeing a git url in the source array of the Pkgfile, not /usr/bin/ports reading a syncfile under /etc/ports.
<SiFuh>
farkuhar: didn't Chinarulezzz do that already?
<SiFuh>
Also I know I and romster did one with aria2c so we could download files from torrents as well using pkgmk
<farkuhar>
What's nice about the Pkgfile format is that it's self-contained. The downside is that by mixing data and code, your tools sometimes have to spawn a shell to source the Pkgfile, and other times can get away with reading it themselves.
iNomad has quit [Quit: leaving]
<farkuhar>
KISS Linux had the interesting idea of splitting our Pkgfile fields and variables among multiple files: one for dependencies, one for sources, and another for the build() function. That way your tools can grab only what they need for the task at hand, not bothering to source an entire shell snippet if all they want is the urls of the sources.
marlin has quit [Quit: Client closed]
<farkuhar>
and although I can't seem to find any mention of it in the official KISS Linux documentation, each source url can be appended with a note like "?noextract" to prevent automatic unpacking of a tarball. With our tools you would have to redefine unpack_source(), or use the renames array to change the suffix.
<ukky>
kisslinux[dot]org has recently expired. It seems that the community has lost domain name. There is 'vocal remover' ad on the main page.
<farkuhar>
ukky: for me the site loaded fine yesterday. Maybe the DNS changes haven't propagated to your ISP yet?
<ukky>
farkuhar: It loads fine on my system. It was down a few days ago. A new owner, probably, reconstructed new site from archived version, adding a few ads.
tarxvfz has quit [Quit: tarxvfz]
<ppetrov^>
farkuhar, would be great if pkgmk could understand git links
lavaball has quit [Remote host closed the connection]
<jaeger>
^ hopefully shouldn't break anything, as far as I can nothing uses these anymore, last thing was nvidia-settings a while back
exark has quit [Quit: quit]
<jaeger>
s/can /can find /
exark has joined #crux
<ppetrov^>
:)
<ppetrov^>
jaeger, I have gtkglext which depends on these, but I'll just adopt the dropped ports in my repo, np
<jaeger>
IS gtkglext also in your own repo?
<ppetrov^>
yes
<jaeger>
OK.. yeah, feel free to grab them
brian__ has quit [Remote host closed the connection]
<SiFuh>
What is happening with this linux-PAM debacle? It is insane that it has been wrongfully sitting in core this long and no one has done anything about it.
<cruxbridge>
<tim> no matter how often you bring it up, it probably won't change a thing. fork it, let others use it, be happy. move on
<SiFuh>
cruxbridge: Not moving on. It doesn't belong in core, it belongs in opt. I use CRUX for CRUX, not because I want to use his system
<cruxbridge>
<tim> whatever
ppetrov^ has quit [Quit: Leaving]
<SiFuh>
Interesting how this distro changed from a minimalist distro to a distro dependent a single users decision to introduce bloat that no one needs. Hmm
<ukky>
SiFuh: It would take so little of your time to wipe out CRUX and install any non-CRUX distro of your choice. Slackware, OpenBSD, whichever best suits your needs.
<SiFuh>
ukky: Or stand up for what CRUX is suppose to be? You know that other distros moved to things like systemd and binary blobs because those who did stand up were ignored?
<ukky>
SiFuh: Not all CRUX users are minimalist. There are users who don't care about linux-pam. There are users who need linux-pam.
<SiFuh>
ukky: Then add to it, seems completely silly that others have to remove from it.
<SiFuh>
If you start sneaking in stuff that others don't need then you are contradicting what CRUX standards for. And where does it stop? Before you know it, developers have dictatorial powers and completely ignore what the users want. CRUX will be either officially a lie or officially dead.
DaViruz has quit [Ping timeout: 260 seconds]
<ukky>
SiFuh: CRUX devs are not paid for maintenance/development, they implement SW in a way they see fit, and it is up to them whether they want to listen to user opinions.
<SiFuh>
ukky: Why are you defending this?
DaViruz has joined #crux
<ukky>
SiFuh: Defending, because if devs decided linux-pam should be in core and enebled, then it will be like that. I can voice my opinion that linux-pam is not installed on my system, but I will not ask devs to disable or remove linux-pam from core. It is extra work for devs and CRUX maintenance in general.
zorz_ has joined #crux
<SiFuh>
ukky: It's completely uncrux though. CRUX = the user base. Not the devs.
zorz_ has quit [Client Quit]
zorz_ has joined #crux
zorz has quit [Quit: Lost terminal]
<SiFuh>
Do you know it took 18 years for me to get the CRUX guys to fix their broken init system? You can thank jaeger for that. He pushed it to the the others and it got solved in a couple of days. But because Per and I had a conflict it was 18 years that passed before it finally got fixed.
zorz_ is now known as zorz
<remiliascarlet>
ukky: Linux has had so many bloaty changes over the past decade, I really doubt you can even have a truly minimalist distro these days, unless you slightly alter its definition.
<SiFuh>
I stood true to a clean system and to a simple working system. I followed the CRuX motto. I care now about the work load for the CRUX developers. They chose it. I bring the issue forward and they look into it. It's not hard to see
<ukky>
SiFuh: CRUX requires redesign to support cases when user wants to change stuff in CRUX. And this redesign might compilcate package management.
<SiFuh>
ukky: So adding unneccessary junk makes it fine?
<ukky>
SiFuh: No, I am not saying that
<remiliascarlet>
How does moving PAM to a different collection require a redesign?
<SiFuh>
pkgmk, perl, python3. These are not core material but they are there. As much as it annoys me we accepted it because CRUX has the best package manager on the planet. It shouldn't be in core but it makes CRUX what CRUX is.
<SiFuh>
Linux-PAM has no business in core, just as cryptsetup and llvm2. Yet we all use them.
<ukky>
remiliascarlet: Redesign is required to support alternative packages instead of build-in SystemV, etc.
<remiliascarlet>
Except nobody was talking about that.
<ukky>
I was talking about redesign to support enabling/disabling linux-pam
<SiFuh>
ukky: But it wasn't enabled by default. It was introduced in 2019 and then ports were enabled for it.
<ukky>
SiFuh: I had no idea when linux-pam was enabled as default (using CRUX for about a year)
<SiFuh>
ukky: It was enable during Freenode days. I spoke against it from the start. But it happened not at the request of the users but at the request of the developers. Which is 100% unfair.
<SiFuh>
And every freaking time I have to update CRUX, I am constantly removing it from my system and modifiying ports. Why? It is uneccessary, it is not needed by any core package, so why is it in core? Because a handful of people (not even 1% of the user base) said they wanted it.
<SiFuh>
I am just curious, do you update regulary? Or do you manually remove it too?
brian|lfs has quit [Read error: Connection reset by peer]
DesRoin has quit [Ping timeout: 240 seconds]
<SiFuh>
And if so, why are you defending it when you have to uninstall and 'patch' as you say many ports to get rid of it. Minimalism should be erased from the front page of crux.nu.
DesRoin has joined #crux
<ukky>
I do not update regularly, but I have automated port modification process using proxy git repos for port collections. There is manual intervention, but not extensive, mostly just deleting signatures of modified ports.
<SiFuh>
I think I was informed about that. Based on message right?
<ukky>
SiFuh: message?
<SiFuh>
Email, mailing list, something like that
<SiFuh>
Unless you are the user who designed it
<SiFuh>
I was asked, can't you select to not install linux-pam during a new installation of CRUX.
<SiFuh>
Yeah, not happening. Packages are boot strapped
<ukky>
There was mailing list discussion about using svn to automate port patching. I have designed my own system, based on git. farkuhar has helped me.
<SiFuh>
Guarantee you won't be able to log in
<SiFuh>
SVN is a thing still?
<ukky>
Some people still use svn
<SiFuh>
Anyone want to fresh install CRUX and prohibit linux-pam from installing? Bet it won't work and if you force it, you won't be able to log in.
<SiFuh>
ukky: I did until 2 years ago. But that was my fault because I switched from OpenSSL to LibreSSL
<ukky>
SiFuh: I believe you, as istall ISO is built using linux-pam
<farkuhar>
SiFuh: if you're adamant against having linux-pam installed "every freaking time you have to update CRUX", then just create an overlay port collection that masks /usr/ports/core, and populate it with your customized replacements for linux-pam, shadow, sysvinit, openssh.
<SiFuh>
ukky: I built a few ISOs with jaeger's help.
<SiFuh>
farkuhar: And how about the $versions?
<farkuhar>
or use prt-get.aliases, with lines like "linux-nopam: linux-pam"
<SiFuh>
Won't work
tilman has quit [Ping timeout: 256 seconds]
<SiFuh>
ukky's port list errors out when you set it to with-pam or with-libpam which is now CRUX default
tilman has joined #crux
<farkuhar>
SiFuh: I agree, the first removal of linux-pam must be done manually after a successful login to the newly-installed system. But then you can apply ukky's patches to the affected core ports, and put them in your overlay, masking /usr/ports/core.
<SiFuh>
farkuhar: I have a third option. a pkgmk.conf with the pam option set to YES
<SiFuh>
farkuhar: I didn't need ukky's patches. I did it along time ago. Then last time, I messed up and came on #crux and said something like "fscking linux-pam, why is it even in core?". No one replied. And when I did the update for ppetrov^ I disabled all the linux-pam and it made me angry again.
<SiFuh>
That is why I question his patching OpenSSH 1.6 or 1.9 when we were on 1.15 or somehting.
<SiFuh>
ukky: It sounded like, by your words you were patching the source. But I knew this could not be the case as I had patched no sources. I just edited every package file and disabled it.
<SiFuh>
I will repeat again. Why is this in core?
<SiFuh>
necessary evil
<SiFuh>
That is what linux-pam is
<SiFuh>
But it isn't necessary
<SiFuh>
Why is it in core?
<jaeger>
Because the decision that was made was to enable it by default along with core things that use it, simple as that
<SiFuh>
By whose decision jaeger?
<SiFuh>
In 2019 I saw that the decision was made by the developers and not the users
<jaeger>
As you've said many times, and nobody argued that
zorz has quit [Quit: Lost terminal]
<jaeger>
You can't please everyone, it's not possible.
<SiFuh>
So the developers is what we please?
zorz has joined #crux
<SiFuh>
That is kind of un-CRUX
<jaeger>
Without using some kind of system like use flags and rebuilding everything you can't have both easily. At the time it was decided by core devs that it was useful enough to include and enable by default. Obviously not everyone agrees, and you have full autonomy to change your systems in any way you want
<SiFuh>
So basically you are saying that your junk, must be offloaded onto us and we clean up the mess?
<jaeger>
I'm not trying to justify it, it's a subjective thing. I personally find PAM useful for some situations but not all. In the time you've been ranting about this you could have rebuilt everything that uses PAM on your system instead of investing so much effort into being angry
<SiFuh>
That's like arch linux, debian, redhat stuff with some freedom
<jaeger>
If PAM is enabled, someone won't like it. If PAM is disabled, somneone won't like it.
<SiFuh>
jaeger: But I did rebuild everything. Every time. Every upgrade. It is so un-CRUX
<jaeger>
I don't personally feel it's un-CRUX, so to speak. Nowhere does CRUX say "you can't have PAM"
<SiFuh>
That is kind of childish
<jaeger>
That's a very ironic thing to say :D
<SiFuh>
Thanks, I hope stay young too
<SiFuh>
It is not needed but it is forced upon us and we now have to physically remove this. Per Liden didn't do this. So who is making the new decisions?
<jaeger>
Personally (this is opinion, obviously) I don't think that PAM violates CRUX's tendency to be lightweight and keep it simple. Sure, you CAN complicate them if you need or want to, depending on your setup, but the defaults seem pretty reasonable to me
<jaeger>
It isn't forced on you. You can remove it at any time
<SiFuh>
jaeger: I know you are the one that installed it into CRUX. But I think it was a bad decision and you should have relied on the users and not the developers for and opinion of yes or no
<jaeger>
Duly noted. You've made that opinion quite clear
<SiFuh>
So reasonable relies on you jaeger?
<jaeger>
I was not the sole decision-maker on this and you know it. Also, I just stated that that was my opinion, not objective truth.
<SiFuh>
I said that. You got the opinion of the developers. NOT the users
<jaeger>
Also, you realize that we are also users
<jaeger>
Yes?
<SiFuh>
I was there. I know. I remember it
<SiFuh>
No
<SiFuh>
It was a decision made by developers. The vote applied to them.
<SiFuh>
In 2019 beerman was the only one in the group that was not a developer