<farkuhar>
Burna: there's no clang-32 in any of the published repositories, but some users might have a port of that name in their personal port collections. Indeed, one of the maintainers has recently turned his attention again to the compat-32 repo (https://git.crux.nu/ports/opt/issues/6) so it's certainly possible that he has clang-32 in his overlay while the issue is being investigated.
ivandi has joined #crux
Burna has quit [Remote host closed the connection]
tilman has quit [Ping timeout: 248 seconds]
tilman has joined #crux
ppetrov^ has joined #crux
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
<cruxbridge>
<tim> farkhuhar: wdym with "one of the maintainers has recently turned his attention again to the compat-32 repo" exactly?
<cruxbridge>
<tim> the line is there so that uses can create that port for themselves if they need it ;)
lavaball has joined #crux
ppetrov^ has quit [Quit: Leaving]
zorz has quit [Ping timeout: 252 seconds]
<cruxbridge>
<pitillo> mpark: Hi mark, which kind o embedded hardware? (Just out of curiosity here)
ppetrov^ has joined #crux
<farkuhar>
tim: I was referring to the new labels "bug" and "help wanted" added to opt #6 this week, drawing renewed attention to an issue opened almost a year ago (also involving compat-32 ports).
<cruxbridge>
<tim> Weird wording then, but what do I know, I am not a native speaker. fwiw I tried a lot in the context of that ticket but nothing helped/worked. Others are always free to join and see if they can solve it, these labels just make it extra clear.
ppetrov^ has quit [Quit: Leaving]
zorz has joined #crux
erik_ has joined #crux
serpente has joined #crux
<mpark>
pitillo: mostly rpi's and their irk - nothing professional :D
<mpark>
I'm doing home automation with them, mostly.
<mpark>
a couple are actually running servers
<mpark>
anyway, freedesktop's gitlab is back, so I'm resuming my desktop test build - while it compiles, I'm headed off to pick up a car part -- bad socket on the tail light ( nothing major )
<mpark>
I'll be back in a bit
<mpark>
fingers crossed no more build errors :D
<SiFuh_>
mpark: revdep ;-)
ppetrov^ has joined #crux
ppetrov^ has quit [Quit: Leaving]
ppetrov^ has joined #crux
<ppetrov^>
yep, all works :) However, is it a good idea that the device name is set in rc.d/wpa_supplicant? Can't be set froma conf file?
lavaball has quit [Remote host closed the connection]
<SiFuh_>
ppetrov^: Technically that is the conf file.
<SiFuh_>
ppetrov^: It does seem a bit weird thoug editing a rc script. Maybe something that can be modified in the future?
<mpark>
SiFuh_: yup, but not yet
<farkuhar>
SiFuh_: I think ppetrov^ expected the filesystem to respect the distinction between data and code. One file for the business logic, another file for the parameters that might change from one deployment to the next.
<SiFuh_>
Hmm pretty sure I said that farkuhar but it not so many fancy words ;-)
<farkuhar>
That's the same distinction that I think /etc/rc.functions was supposed to achieve. Keep the user-editable fields (KEYMAP, BLANKTIME, ...) separate from the code that the user is not supposed to modify.
<SiFuh_>
mpark: What do you think? Introduce a configuration file for ethernet Wi-Fi devices with a universal configuration file that they (The rc scripts) can read?
<SiFuh_>
farkuhar: Yeah I didn't like it myself but I did agree with that as I mentioned before. Rather than modifying every script you could just edit one file. But the other side issue, that I did mention was if I disabled colours then I have a junk file existing on my system which is not the CRUX way.
<ppetrov^>
farkuhar, you did some work on pkg-get, right?
<farkuhar>
SiFuh_: You'll notice that pkgutils ships a default pkgadd.conf with "UPGRADE ^etc/rc\.d/wlan$ NO", so they always expected the user to edit those startup scripts in-place, and rejmerge any differences if a new wpa_supplicant gets released (with non-trivial changes in the startup script).
<SiFuh_>
Another issue is apart from tim. Does colours actually work? I do get bold white font but I never saw any red errors.
<farkuhar>
ppetrov^: Yes, I did push a few changes to pkg-get.
<farkuhar>
SiFuh_: If you never saw any red errors, it could be because all of the commands succeeded. You have to deliberately trigger an error if you want to see the colours.
<ppetrov^>
and pkg-repo, i remember had some trouble when generating post-install scripts for the packages
<SiFuh_>
farkuhar: Yeah, but why edit the actual rc script? Wouldn't a configuration file be more logical?
<ppetrov^>
did you look into that?
<ppetrov^>
SiFuh_, I feel uncomfortable editing an rc script, but maybe that's just me
<SiFuh_>
farkuhar: No I deliberately made it error out. Not even bold white font. Just standard text. Maybe it is my machine?
<SiFuh_>
ppetrov^: No I do too.
<ppetrov^>
there are dozens of us! dozens!
<SiFuh_>
Heh. Actually been thinking about it since like the early days when Per was around.
<ppetrov^>
as for the wifi card name, before it was wlan0, recently names are pretty weird for network cards
<SiFuh_>
ppetrov^: Oh there is a kernel setting for that. I think ifname=1 or 0 or something. But if you have multiple interfaces you need not only specify that you need udev to recognise the order too.
<farkuhar>
ppetrov^: Yes, there were a few commits that changed the format of PKGINST. Now it doesn't create a separate function for each port (with the name normalized), but passes the port name verbatim as an argument to a run_script function.
<SiFuh_>
ppetrov^: I think it is this. net.ifnames=1
<SiFuh_>
ppetrov^: Then you should get eth0 eth1 wlan0 wlan1 and so on.
<farkuhar>
SiFuh_: I did see the bold red error message when testing the recent rc, so maybe it is your machine.
<ppetrov^>
SiFuh_, thanks. I think Pat uses this for his Slackware kernels, and when I used your config, I disabled these in /etc/default/grub
<ppetrov^>
farkuhar, great, thanks
<SiFuh_>
farkuhar: Cheers
<SiFuh_>
farkuhar: I think I will do a 3.8 fresh install this week and look at this rc error that zorz is talking about with lvm2.
<erik_>
I am currently attempting to install in a (gnome-boxes(I think it uses qemu/kvm behind the scenes)), but I have gotten stuck at the bootloader GRUB/UEFI part, I made a 100mib partition for uefi with cfdisk but I am getting strange errors when trying to '$grub-install' according to the handbook
<ppetrov^>
farkuhar, did you find any issues with findredundantdeps?
<SiFuh_>
Then if we figure it out, send a patch for jue and old timmy to fix it.
<ppetrov^>
old timmy?
<SiFuh_>
erik_: dpaste the errors.
<SiFuh_>
ppetrov^: Yeah beerman. Tim
<farkuhar>
SiFuh_: It's interesting that they folded the "UPGRADE ^etc/rc\.d/wlan$ NO" rule into pkgutils directly, rather than adding it to /etc/pkgadd.conf using the post-install script for wpa_supplicant.
<ppetrov^>
ah,... didn't know his name
<SiFuh_>
ppetrov^: I like the name beerman from beirman. But I like beer so :-P
<SiFuh_>
Or however it was spelled
<ppetrov^>
his family name is Beirman?
<SiFuh_>
Maybe Bierman?
<ppetrov^>
well, "Beerman" is the only logical choice
<SiFuh_>
Great choice to be honest. Kind of envious
<ppetrov^>
yep
<erik_>
SiFuh_: I can't select in the VM window..., ill see if I can temphost a screengrab
<SiFuh_>
erik_: A photo or a screengrab is fine.
<SiFuh_>
erik_: At least it will assist people in seeing what you are talking about.
<SiFuh_>
farkuhar: I think it should be a config file for logical reasons. rc scripts then can read it.
<SiFuh_>
erik_: Not a grub guy sorry. But for syslinux I can assist.
<SiFuh_>
Grub guys ^ erik_ needs an assist
<ppetrov^>
erik_, is it EFI?
<ppetrov^>
i mean, UEFI boot?
<ppetrov^>
have you formatted /dev/sda2 properly?
<SiFuh_>
ppetrov^: Google search says change to ext4.
<erik_>
it's a VM running in gnome boxes
<ppetrov^>
so, you formatted it and mounted it in /mnt/boot/efi febore install?
<ppetrov^>
*before
<farkuhar>
erik_: Do you have efibootmgr installed on that VM?
<erik_>
I have only ever really done partitioning in gparted/k(de)parted, but this time I tried using cfdisk to do the partitioning stuff, and I __THINK__ it worked, seeing as lsblk returns what I expect
<SiFuh_>
I liked the bold white text though. I thought it was cool
<jaeger>
erik_: what errors do you get from grub-install?
<SiFuh_>
The problem though jaeger isn't the colour. It is the script being written wrong. 'It just works' mentaliity seems to be the issue.
<jaeger>
My understanding is that Tim uses shellcheck as part of his editor setup, so maybe the shellcheck warnings are new? I hadn't used it before so just guessing.
<SiFuh_>
It needs to be adressed and it will be.
<jaeger>
And they're warnings, not errors, right?
<erik_>
jaeger: I sent a screenshot link earlier, one sec
<SiFuh_>
jaeger: Nice, but shellcheck complained it was not correct.
<jaeger>
With that said, I find the function syntax a bit hard to read, my tests were using exported vars instead
<farkuhar>
jaeger: the issue tracker for ports/core? Why not the issue tracker for tools/rc? The latter is where a pull request makes more sense.
<SiFuh_>
jaeger: It's really just spam. And it is written incorrectly. No worries, we will have it fixed soon. We are focusing on the lvm2 crisis for now because it is much more serious and it may need a shutdown script modified too.
<jaeger>
"Installing for i386-pc platform" doesn't look right
<jaeger>
Does "efibootmgr" work?
<jaeger>
Anyway, if the colors are actually broken I would agree they need to be fixed. If they're working but you don't like the syntax, ok, but that's lower priority than broken
<SiFuh_>
jaeger: Actually the colours don't work on my machine and a couple of users. I actually don't know why.
<SiFuh_>
jaeger: How can it work on farkuhar and beermans machine and even yours but my PC (Laptop) and a few others not. Strange right?
<jaeger>
ok, I will add it to my list to test
<SiFuh_>
jaeger: We already fixed the colour issue. No need to ad to the list and the correct printf function
<SiFuh_>
Our priority is not lvm2
<SiFuh_>
not/now*
<jaeger>
My screenshots weren't the stock rc, for reference. I rewrite rc.functions entirely (ironically not using functions) and modified rc, rc.multi, and rc.shutdown
<farkuhar>
SiFuh_ means that we're satisfied with the patch to clean up the function syntax, not that we have colours working on all tested hardware.
<SiFuh_>
farkuhar: It works here farkuhar because it was written correctly
<jaeger>
I would personally prefer to move away from functions for color output but we can certainly discuss that
<jaeger>
OK, we get that, continually repeating it is just inflammatory
<SiFuh_>
What I mean that we are not bothered with the colour functions anymore because we fixed that issue and used the correct printf functions. What we are now concerned with is getting lvm2 primed. Fixt two problems with one patch.
<SiFuh_>
jaeger: I don't even see the point of the colour functions and if it has to exist then fine. Let it. But I am curious about a dead file that exists rc.function which is easily modifiable but now a junk file on my system if I decided to opt out on colours.
<jaeger>
Some people like colors, some don't. If you don't want them you are absolutely free to strip out that functionality
<SiFuh_>
jaeger: I am actually for it. Let it be.
<jaeger>
You just said you don't see the point, then you're for it. That's confusing.
<SiFuh_>
I may never use it, but lets move the functions file into rc.
<farkuhar>
Heh, woe to the user who insists on eradicating junk files by writing the pkgadd.conf rule "INSTALL ^etc/rc\.functions$ NO", and then sees the system refuse to start when /etc/rc tries to source a nonexistent file.
<SiFuh_>
Yes, I don't see the point doesn't mean I am against it.
<SiFuh_>
jaeger: What is the point of a pretty coloured screen when our goal is to boot the system up as fast as mechanically possible?
<SiFuh_>
jaeger: And you are correct, some users might want that. So let it be. Just do it right.
<jaeger>
The point of it is aesthetic preference. I personally like having a bit of contrast and I do not care and cannot even notice nanoseconds of difference in boot time. This argument is nonsensical.
<farkuhar>
SiFuh_: You can get to a login prompt even faster, if you rewrite /etc/rc.d/sysklogd to call start-stop-daemon with OPTS="-ss -b". I'm surprised that's not the default.
<SiFuh_>
jaeger: I was never against beerman introducing colours. I am against the amateur code. It needs to be cleaned up. Which we are working on. But we hit a stump which was brought to our attention about lvm2.
<jaeger>
OK, then forget the colors for now. What's wrong with LVM?
<erik_>
I am honestly considering just trying to installing it on my spare-spare-laptop, a lenovo thinkpad X250, just hoping I can get the touchscreen and fingerprint scanner working (touch works under arch, and I remember the fingerprint scanner worked when I was still on debian-based). it is getting a bit late now(swedish time. 19:35) so it's time for dinner and some gaming/something before bed, good thing I don't have work until monday, so I can
<erik_>
actually put some time in
<SiFuh_>
jaeger: We are not entirely sure. Sorry but we are trying to find that out
<jaeger>
erik_: for what it's worth I suspect your system there was booted in BIOS mode rather than EFI mode. If you want to investigate that later let us know. Have a good evening
<jaeger>
What are the symptoms, then?
<farkuhar>
jaeger: From what I understand, the test [ -x /sbin/lvm ] needs to be accompanied by something else before running the vgchange or vgscan commands, because a user might have lvm2 installed without having configured any logical volumes.
<SiFuh_>
jaeger: And it appears the issue with lvm2 goes back to jue. But as I said, we are not entirely sure. But it was pointed out by a user who is hard to understand and often away.
<jaeger>
vgchange and vgscan do something stupid when there's nothing configured?
<SiFuh_>
So let it be. Lets consider the beerman patch ready but the lvm2 is still in the works for now.
<SiFuh_>
jaeger: We don't entirely know. We need that user to have an in depth conversation explaining it to us.
<erik_>
one quick question; how do I set the keyboard layout to swedish:no-dead-keys in the installation enviroment, the arch-wiki page talks about some systemd-thingy, which I dont expect crux to have installed
<SiFuh_>
setxkbmap?
<farkuhar>
As for the faster bootup times, I need to dig up my old modifications of /etc/rc.d/sysklogd because the default script produces a noticeable delay, which disappears when I called start-stop-daemon with the -b option (not -ss, that option is passed to /usr/sbin/syslogd).
<jaeger>
'loadkeys sv-latin1' would be my guess
<erik_>
SiFuh_: setxkbmap:can't open display
<SiFuh_>
jaeger: Wasn't jue the rc guy? Or am I wrong?
<jaeger>
Not sure about the no-dead-keys variant but setxkbmap is an X command rather than console
<SiFuh_>
jaeger: I'd prefer once we have the patch someone qualified to review it.
<jaeger>
I'm not sure what you mean by "rc guy", maybe?
<jaeger>
Anyone can write rc scripts. If you mean that he does a good job of it, sure
<SiFuh_>
jaeger: Pretty sure, maybe wrong, that jue is the one that took my patch for the broken init system that was broken for over 18 years then he reviewed it and tweaked it and made it better?
<SiFuh_>
And if it is jaeger then I trust him.
<jaeger>
Check the git history if you want to be sure
<SiFuh_>
I actually don't know how to do that. Everything has changed in the last 25 years.
<erik_>
jaeger: thanks, loadkeys worked, i'm not sure about the nodeadkeys thing either, but atleast I don't have to write lisp-code in the live-env, would be a good thing to put in the handbook tho, seing as it is a (primarily)swedish developed distro, that assumes everyone installing it knows how to use a US keyboard layout
<jaeger>
I don't think any of us still working on it are swedish for what that's worth
<farkuhar>
erik_: `find /usr/share/kbd/keymaps -name "*nodeadkeys.map.gz"` only returns two hits. Can you try `loadkeys de-latin1-nodeadkeys` and see if a German keyboard layout sans-deadkeys is close enough?
<SiFuh_>
jaeger: What the fsck? It says you did it, but when I clicked it, it is not my patch.
<SiFuh_>
No, I dont have an account. Everything is beyond me now jaeger. Best get another to do it for me. I am old school I still don't even know how to use a mobile phone :-P
lavaball has joined #crux
<cruxbridge>
<pitillo> Really registering to the development section of a project you are using is beyond you SiFuh_ ?
<jaeger>
I just installed lvm2 on a system with no LVM usage. It boots fine, just shows some warnings, for reference
<SiFuh_>
jaeger: Not sure, But I head shutdown is a problem. As I said. We are awaiting the guy to explain his issue but there is a problem. He switched to runit
<erik_>
farkuhar: well, it _IS_ a german qwertz keyboard I have on my thinkpad-laptop, but I think I will stay with muscle memory swedish-qwerty, or I will start typing zes to answer yes/no questions..., the lack of nodeadkeys would be more anoying if I were planning on writing code with a bunch of single-quotes '`´~ are all "dead" keys if I recall correctly, and will wait for another keypress to actually type the character, to allow for "strange"
<erik_>
letters like `e, which gets __really__ anoying if you write code, (I had a problem with this at my last job, I had to use a windows device to write (b/d)ash-code, windows does not have a nodeadkeys/programmer variant of the swedish layout, or a way to download/install one), I expect the x-keymap to have a nodeadleys variant. PS sorry for the ranty-ness
<SiFuh_>
ptillo I am not going to register. What is wrong with here? All the devs are here
<jaeger>
I saw no issues with shutdown but to be fair my sample size is 1
<SiFuh_>
ptillo if I have a serious issue or a patch. Can't I just sent it to #crux-devel? Do not forget, I am the olderst runnin CRUX user on all channels. I was there before it was officially released by Per. I was using CRUX many years before many of you all even knew or learned about what CRUX was. Even before jaeger and jue's time.
<SiFuh_>
jaeger: Thanks. We will probe more.
<SiFuh_>
jaeger: If nothing we will push the beerman fix but if we find there is an lvm2 issue then we will patch accordingly.
<cruxbridge>
<pitillo> For me isn’t a problem, but it could be a lack for the project. You are free to register or don’t register…. But actually many projects are moving to a better ways to manage improvements, fixes, contributions…. And CRUX IMHO is doing great in that sense.
<cruxbridge>
As said, we use IRC and mailing list to keep track of these project related problems, which are fine too of course.
<cruxbridge>
But saying is beyond you doesn’t seem right for me (sorry my lack of English)
<SiFuh_>
ptillo IRC was the original place. It works, it still works, I ain't bothering with a change.
<cruxbridge>
<pitillo> SiFuh_: It is, but like I said, for me it’s a lack for the project with current resources.
<cruxbridge>
I know who you are, even though I’m a noob here :)
<SiFuh_>
ptillo I am am in the mailing list. I think maybe because I am old, it is just confusing.
<SiFuh_>
ptillo, the old ways function fine. I like that.
<SiFuh_>
ptillo the old ways are still here. I am sticking with them. I like them they are easy to vent steam and it gets done. You guys do your fancy stuff. But this is tried and true and works. Want a patch? dpaste in IRC
<SiFuh_>
ptillo and besides. jaeger is here and he is old school. He has been here a long time and he gets shit done.
<cruxbridge>
<pitillo> It’s fine SiFuh_ I think in the project, how it evolved and we have many ways to do things. If you are more comfortable doing in the old days ways, it’s fine. I prefer to try to evolve while keeping my preferences, but thinking in the project itself, as a maintainer, IMHO it needs to adapt to the current days to get improved in many ways
<SiFuh_>
ptillo I think it is de-evolving. As many of us have seen and speak up about it in other CRUX channels.
<cruxbridge>
<pitillo> De-evolving in management sense? In software updated changes? In quality senses?
<SiFuh_>
I prefer to say in quality as my recent vent on #crux about the absurdity to do with amateur coding in /etc/rc
<cruxbridge>
<pitillo> So if there are more people who think in that way, why you don’t open a discussion on the mailing list as we did 20 years back
<SiFuh_>
Then only to be told by the guy who screwed it up, I should provide a patch.
<cruxbridge>
<pitillo> And let everyone using, developing, maintaining the project speak about
<SiFuh_>
That isn't logica
<SiFuh_>
ptillo, my personal opinion is what you are asking for?
<cruxbridge>
<pitillo> The more eyes and hands checking things, the stronger the thing gets
<cruxbridge>
<pitillo> Your opinion for me is really important, like others opinions, the ones who were using CRUX for 20 years
<SiFuh_>
ptillo When that code was manipulated in such an important piece of software. Yes, I think jaeger and jue and other users should have reviewed it.
<cruxbridge>
<pitillo> And I miss CRUX meetings, when averything was discussed by those who maintained the project
<SiFuh_>
But no one did. And it was released to the mainstream.
<SiFuh_>
It is an rc script. Every man and his dog should be inspecting and tearing it to pieces. Making sure it was perfect
erik_ has quit [Ping timeout: 276 seconds]
<cruxbridge>
<pitillo> So where is the problem? I can see only one problem, opening a discussion/talk on IRC. Change IRC by mailing list where everything is kept and can be followed in an easier way than IRC
<SiFuh_>
Technically I am the problem because I was so mad at beerman saying "enjoy"
jaeger has quit [Quit: leaving]
<SiFuh_>
When it was clearly a shit script that mpark had no idea about.
<cruxbridge>
<pitillo> You don’t like gitea, but IMHO the mailing list could be a better place where we can check any issue
<SiFuh_>
No, pitillo. I am an ancient. I ain't fscking with that shit. Here is fine. It still works. So I stay
<cruxbridge>
<pitillo> But it isn’t optimal SiFuh_
<SiFuh_>
Wanna bet?
<cruxbridge>
<pitillo> You didn’t get the response you expect from this way, why not keep it in the other old school way?
<cruxbridge>
<pitillo> I don’t like bets :(
<SiFuh_>
jaeger is quite attentive and reasonable and he jumps on it very fast. It works still
<SiFuh_>
changing the source pitillo isn't going to change the outcome
<SiFuh_>
Beerman was adamant in his ways. Shit code. Sure! But he is not the problem. It should have been reviewed
<SiFuh_>
No one reviewed it and let it out
<SiFuh_>
But he defends it so arogantly
<farkuhar>
jaeger and erik_ both left already, but it was an interesting suggestion by jaeger, that erik_ somehow managed to boot his VM in BIOS mode rather than EFI mode. I couldn't tell from the screenshot, what erik_ was using for his VM.
<SiFuh_>
Should have said something like "Yeah I fucked up, can you fix it for me?"
<cruxbridge>
<pitillo> Because may be we didn’t use the right place/way/method to keep track of it and the issues
<SiFuh_>
piillo, I have nothing at all against his modifications to introduce colour. I just think it was written wrong and even shellcheck complains how shit it is.
<cruxbridge>
<tim> SiFuh: I defend nothing but the fact that all you ever do is bad mouth stuff in the worst way possible, and it's simply off putting. Tickets, pull requests, and whatever not are exactly the tool to work on better quality. Just throwing accusations around never got anything fixed. It's also not broken, it's just a warning. You are ignorant in the fact that I fixed other warnings and problems reported by people that listen to you and didn't open a...
<cruxbridge>
... ticket.
<cruxbridge>
<tim> You might be too old school for a mobile phone but you can't read a git history? Right.
<cruxbridge>
<tim> I get the feeling that all you want to do is being angry with somebody, and you are proud of that.
<SiFuh_>
But now it has become some huge fscking deal. And beerman says 'well provide a patch'. Huh? Am I his fucking maid? You broke it, you fix it. Ask for help there are many of us here will glad to help him. He isn't alone. We are here.
<SiFuh_>
tim That's what you think. Stop it!
<cruxbridge>
<tim> Me stop it? How about you stop it?
<cruxbridge>
<pitillo> Well, I think the mailing list could be a better place for this kind of issues, but like I said, it’s a noob opinion
<SiFuh_>
tim well don't write bad code. Ask us for help.
<SiFuh_>
tim we are not your enemy.
<cruxbridge>
<tim> You might have been so long in this channel, yet everybody has done a thousand things more for CRUX than you have ever done. It's sickening that we have somebody like you linger around and make everything feel bad. Name calling, bad mouthing, with nothing to show to make it better. Bravo to you
<cruxbridge>
<tim> We? What is the collective you talk about?
<cruxbridge>
<tim> Talk for yourself.
<cruxbridge>
<pitillo> Just focus in what is important, the project. Don’t get mad
<SiFuh_>
tim oh really? That same old story where you maintain ports and do lots of work but judge me? It doesn't work. I wrote the original init system for CRUX!
<cruxbridge>
<tim> Bravo
<cruxbridge>
<tim> And you also followed that 18 year old standing bug, I forgot. Tell the story again to make it better
<SiFuh_>
I am the one from day one was there with Per (even though he doesn't like me) to create the pkgadd system
<SiFuh_>
Get over it. CRUX existed before you.
<cruxbridge>
<tim> I am over it, trust me, you are pulling the same stories over and over and over again
<SiFuh_>
Because they are true
<SiFuh_>
And you hate it don't you?
<cruxbridge>
<tim> Sure.
<cruxbridge>
<tim> Nah, I don't. I try not to pester myself with negativity too much, the world is already full of it.
<SiFuh_>
That's a good thing
<cruxbridge>
<tim> You make it harder than it needs to be though.
<cruxbridge>
<tim> Since you are not running ANYTHING you are in no place to demand that IRC is the bug tracker "since we are all here"
<cruxbridge>
<tim> It has nothing to do with "taking a knee" like you put (again, very negatively loaded) - but just how we can organize and get to things.
<cruxbridge>
<tim> I already agreed with you yesterday that it needs another look, sure, but I can't make a fcking warning my priority now just because you "feel mad" about it.
<SiFuh_>
Not really. Your arrogance does. I have done more for crux than you have in what 6 years? Than I have done in 25 years? You seem to think what you know in 6 years if the be it all end it all of actual work.
<cruxbridge>
<tim> Get your emotions under control because nobody here wants them. Thanks!
<SiFuh_>
You haven't done that much.
<SiFuh_>
You just think you have.
<cruxbridge>
<tim> And now I'll do something else, and I will not continue this discussion with you. It's tiresome, I have better things to do.
<SiFuh_>
You really need to evaluate your stance tim. No offence but you are a beginner.
<cruxbridge>
<tim> Sure, I am. Have a day ;)
<SiFuh_>
tim we have a work in progress with a successul repair of your broken script for rc.
<SiFuh_>
tim No we are working on repairing the 'possible' lvm2 problem.
<SiFuh_>
Just shut up and wait. It is coming.
<SiFuh_>
pitillo Sorry about that. He always seems to do that
<SiFuh_>
pittilo, I am not joining some gitea thing or message board because everything works here. Why change?
<cruxbridge>
<pitillo> It’s fine SiFuh_ I try to don’t get out the train and learn a bit more about new ways of project management, which should work too :)
dlcusa has quit [Ping timeout: 252 seconds]
<SiFuh_>
pitillo I actually don't know what just happened. It was very fast and I was accused of being 'emotional'. It is kind of weird.
ppetrov^^ has joined #crux
ppetrov^ has quit [Read error: Connection reset by peer]
ppetrov^ has joined #crux
ppetrov^^ has quit [Ping timeout: 248 seconds]
<mpark>
farkuhar, SiFuh I was just making a remark about bash not being entirely POSIX - I had not yet delved into crux's RC scripts, so I had no further comment or insight to give, only that I would look into it once I'm done dealing with the aspects I'm currently working on to see if there were anything I could contribute later. If I do involve myself further, it would be writing them to be posix compliant and follow a sane coding standard.
<SiFuh_>
mpark: I would prefer CRUX to be more POSIX to be honest
<SiFuh_>
But this is the modern world
<mpark>
You're right when you say I have no idea about the issues to which you are specifically complaining about, because obviously I'm not even at that part of my tinkering yet. I was merely pointing out it's an Open Source project, so if you know how to fix it, maybe you should fix it and submit changes since I doubt anyone is being paid very much if at all to work on this as it's probably mostly a volunteer effort given the size of the community surrounding it.
<SiFuh_>
Sure, I agree
<mpark>
In any case, the rc scripts are very important to me, so when I do get to that part of my project, I will rewrite anything I feel like needs rewriting ( with pure POSIX compliant scripts of course ) and submit my changes to the community for review.
<SiFuh_>
mpark: I just understand why some people need to prove they are better than others and not willing to fix their mistakes but rather leave on on others and the complain they do more than anyone else. So naturally their shit smells nicer.
<SiFuh_>
mpark: They will be fixed within the moth
<SiFuh_>
month*
<SiFuh_>
We are just trying to figure out the lvm2 issue that may or may not exist.
<mpark>
I am makinhg no claims about the quality of my code - only that I'll fix what I see as problems on my system, and let the community decide.
<mpark>
Oh, yeah, I was looking into LVM, and decided I had enough to deal with this early on and will tackle that later
<mpark>
hopefully the issues, if any, will be resolved by then and I can just work on other things.
<SiFuh_>
mpark: Personally, I never saw any issues. I think jue did a great job. But others said it isn't doing what it should.
<mpark>
yeah, I only got as far as thinking about LVM, I didn't actually go through the code yet
<SiFuh_>
mpark: Once we get to the bottom of it. We will release the patch.
<mpark>
so, I have no comment
<mpark>
Yup
<mpark>
as one would expect
<SiFuh_>
mpark: There are some patches. But I need to review them personally so I can be the one you'all blame. Hahaha
<mpark>
Hahaha, I'm not one to pass blame, if I see a problem, I'll submit a patch and an explaination as to what I did and why
<mpark>
I've maid more than my fair share of dumb mistakes, SiFuh_ - I'd be a hippocrate if I called out everyone elses mistakes without submitting a fix.
<SiFuh_>
mpark: Good idea but I am kind of against that. I think the one who fucked it up should be the one to fix it. Then if cannot, be humbile enough to ask for help
<SiFuh_>
We have a great team. Willing to help anyone
<mpark>
I've been working Linux, FreeBSD, Solaris, and HP-UX for the last 25 years -- I hate hate hate HP-UX - thankfully I don't have to deal with old legacy systems anymore.
<mpark>
I think the fix is more important than anyones feelings, and rather have a fix quickly - because if they couldn't do it right the first time, why would I expect their fix to be better, and that way at least I know it's correct.
<SiFuh_>
I agree
<mpark>
It'll be a while before I get more involved than playing with it in a VM, so by the time I do get around submitting corrections, you'll be looking at them shaking your head and thinking their trash too, lol
<mpark>
But I won't get mad if you tell me their trash, so long as you tell me where I screwed up so I can fix them.
Guest51 has joined #crux
<mpark>
in any case, farkuhar you don't need to protect me, though I do appreciate the effort - I've been around long enough ( not this community obviously ) that I know how to handle myself - my skin is thicker than that.
<SiFuh_>
mpark: Well our channels exist for that reason. No judging. You fuck up we will tell you.
Guest51 has quit [Client Quit]
<SiFuh_>
And if you ask mpark we will help and assist you all the way.
<mpark>
SiFuh_: the community has been far more welcoming than I would expect for a Linux project of this nature.
<SiFuh_>
We are actually good guys
<mpark>
I mean, I'm not expecting LKML levels of flame wars, but I was expecting some level of superiority complex and gatekeeping
<SiFuh_>
We have disputes and spats and shit but in general we are all good guys
<mpark>
it's all too common in the linux community, which is a shame
<SiFuh_>
Expected
<SiFuh_>
We have them here as you saw above
<mpark>
yeah, but they're tame compared to other communities I've seen, so far.
<mpark>
which is a good thing\
<SiFuh_>
mpark: Beerman taught me how to make awesome bread. How cool is that? But his god given authority how CRUX exists because of himself is a fucking joke.
<SiFuh_>
I still bake his bread recipe and it is really good.
<SiFuh_>
So we clash but in the end CRUX is here.
<mpark>
anyway, I'm going to formulate an email later, but the only issue I've run into so far is that qrencode's Pkgfile links to a non-existant archive, so I went to the website listed in the Pkgbuild, and it referred me to their github, which if you download the release from there, it doesn't include a .configure script, so you have to execut the autgen.sh and the directory name and archive file name is different which breaks the build further, ( they added lib to the na
<mpark>
just the version number )
<mpark>
I'm working on the issue right up now, and will send it to the opt team later.
<mpark>
well, I'll update ports and make sure it's not fixed already before I submit.
<mpark>
s/.configure/./configure
<mpark>
so, I added one build instruction and modified which variables are used in order to come up with the expected result.
<mpark>
all the other issues where the build system complaining because there were new files that were not in the .footprint.
<SiFuh_>
mpark: I am guessing you don't bake bread
<mpark>
LOL, I'm not sure what that means, I'll bite.
<SiFuh_>
mpark: I have a slightly tweaked version of beermans recipe for hot climate.
<SiFuh_>
Everyone says it is the best loaf they have ever eaten
<SiFuh_>
beerman is tim by the way.
<mpark>
ah
<mpark>
still no idea what you're talking about
<mpark>
lol
<SiFuh_>
Yeah, that crux guy that yelled at me before. :-P
<mpark>
there's an inside joke there somewhere
<SiFuh_>
No, his bakery skills are fucking aweome.
<mpark>
in any event, by the end of this I'll be rewriting quite a few packages to fit my customizations because the package system is dead simple... but I'm going to be looking for a tool to manage a personal binary repository so I don't have to recompile for EVERY machine or vm in my lab.
<mpark>
if I have to, I'll write my own, but I doubt I'll need to.
<SiFuh_>
mpark: Very easy. Most of us stripped out shit like linux-pam and dumb_runtime_dir and other crap.
<SiFuh_>
sudo even moved to opt
<mpark>
I actually need linux-pam because my systems use central management for credentials and permissions.
<SiFuh_>
Then be it
<cruxbridge>
<pitillo> Check pkg-get mpark , may be that could help you managing packages, or take it as a base/template to adapt to your needs
<mpark>
and pam just makes that aspect easy, so I won't be removing it.
<SiFuh_>
But most of us see it as uneeded junk.
<SiFuh_>
Yeah, I don't doubt you mpark. Just many of have ne need for it
<mpark>
I'll look into it ptillo - i misread that as prt-get for a sec and thought you were telling me to use the tool that's not doing what I described
<mpark>
fair enough
<SiFuh_>
mpark: pitillo is not a newbie. He's been around for a long time.
<mpark>
There's a lot of useless garbage in most linux distros - because somebody somewhere needs and uses it and they try to be everything to everyone.
<SiFuh_>
Also a good guy to be honest
<SiFuh_>
mpark: Yep and creeping into CRUX slowly and someone needs to stop it.
<mpark>
I'm probably not going to be using crux long term honestly, I'm probably going to use it as a base to spin my own distro - but I'm not going to dedicate myself to it enough that I would ever redistribute it
<mpark>
I'm just not ready for that level of commitment
dlcusa has joined #crux
<mpark>
when all is said and done, it will have one tool for any given task, except in the case of shells, because I want to be able to run shell scripts I didn't write myself and don't want to write every little detail from scratch.
<SiFuh_>
mpark: CRUX is great though.
<mpark>
I'm going to have one desktop environment, one cli text editor, etc etc...
<SiFuh_>
mpark: I've been using fluxbox since it was first released.
<mpark>
oh yeah, which is why it'll be the base for whatever I do at least that's what it's looking like, and I'll contribute back for sure
<mpark>
fluxbox is great, I prefer openbox personally, but with wayland and development no longer happening, it's got no futer.
<mpark>
future*
<SiFuh_>
mpark: Actually there is some work in the motion of stripping crux down even more. Then provide a Per Liden fork of CRUX since modern CRUX has just gone off targer.
<mpark>
that said, I'm doing plasma because I also need wifey approval, and I don't want to have to maintain more than one environment.
<SiFuh_>
Well we are still in the works. But chances are high that it will come about.
<SiFuh_>
And MUSL is going to be a twin of it
<mpark>
I'll keep an eye out, I'm always playing with stuff in a VM just for fun.
<SiFuh_>
Stick with us, even help would be awesome
<mpark>
I'll consider it. I'm not making any commitments 2 days into a community :D
<SiFuh_>
Well if you join you have a way to to guide it to what you want
<mpark>
I understand that, and first impressions are mostly positive, but I like to know more about the people I'm working with when we're talking about volunteering my time.
<SiFuh_>
mpark: You'll be fine. It is only in the mind and not on paper yet
<mpark>
basically just want to learn your personalities and what not.
<SiFuh_>
mpark: You already know the people. CRUX purists.
<mpark>
not your ssn and home addresses hahaha
<mpark>
I don't think I'm a purist.
<mpark>
I'm more practical, I go with what works before I worry about someones definition of purity
<SiFuh_>
mpark: The are sick of the way crux is moving from Per Liden's crux. They want to go back to the old ways.
<mpark>
yeah, I wasn't here for the old days
<SiFuh_>
Well we have many in the team
<SiFuh_>
Mostly CRUX users
<mpark>
My philosophy is that core should only be what is absolutely needed to boot a system and connect to the internet, and everything else should be in categories outside of that. I would prefer non-core items be installed into /usr/local rather than /usr because those are system specific packages.
<SiFuh_>
And users will have an opinion and not the devs only.
<SiFuh_>
mpark: Mine too
<mpark>
I also think that a project needs strong leadership, and while users opinions should be considered, they should not be the only deciding factor, because a retired sysadmin, I know most users don't know jack
<mpark>
and make a lot of unreasonable demands
<SiFuh_>
mpark: That is why we will write a constition. That cannot be broken
<mpark>
I always try to get as close to their requests as is feasable without breaking security, and what not
<mpark>
Like, I want to give them what they want, but it may be a compromise because maybe the way they want to do it is just stupid
<SiFuh_>
Any dev who breaks it will stripped of dev status
<mpark>
fair enough
<mpark>
I also use flatpak for a lot of userspace apps, not because I think it's a good idea, but because a lot of distros just don't package things correctly so if the official project has an official flatpak, I'll use that instead because at least then if it's messed up, I can get support.
<SiFuh_>
mpark: I think so because Per's way is no longer being followed and many excuses like hardware differences and I like to eye candy is not implemented
<SiFuh_>
not/now*
<SiFuh_>
Our plan is back to basics. Get CRUX what is was suppose to be
<mpark>
The only thing a distro should be customizing about user interfaces like GNOME, or KDE, or whatever is the backdrop and maybe the application menu icon.
<SiFuh_>
No I disagreee it should be focused purely on core. You buil off of it
<mpark>
if they wrote the app themselves, that's one thing, but I want my software to be as vanilla as possible by default
<mpark>
see, I'm not a purist.
<mpark>
I do believe core should be the minimum required, and the extras on top of that should only be added if core is solid and well maintained
<mpark>
but they should not be added to core
<SiFuh_>
like sudo. linux-pam amd dumb_runtime_dir ? All not needed but now esists in core?
<mpark>
yeah, no, they need to be in a repo for system administration extras or something along those lines, they're not required for core.
<SiFuh_>
opt
<mpark>
I'd probably have more variety in repo names, but that's just how my brain works, repos dedicated to a purpose.
<mpark>
opt would be just extra libs for example.
<mpark>
I don't know, I'm okay with opt, extra, contrib I just wouldn't do it that way personally
<farkuhar>
mpark: I had this debate with SiFuh_ a few days ago, whether Per's vision for core was "the minimum needed to boot the system", or "the minimum needed for everyday use". For a source-based distro, it makes sense to have gcc and binutils in core, but by the first definition, they would be considered extra (since you can get away with binary packages compiled on a separate machine).
<mpark>
I would have gcc in core for a source based distro where every package is being compiled locally from source, but if you're a binary distro, I'd move those into a dev repo.
<mpark>
because how are you to install anything if the system doesn't come with the tools to build packages on a source based distro. They're required for functioning be the nature of the distro.
<farkuhar>
SiFuh_ wanted to have an in-between repo called "comp" so that users who don't need a compiler suite can deselect all those packages at the repo level, not by unchecking individual package names during the install.
<mpark>
farkuhar: I'd more lean toward your side of the argument, unless there is also a binary repo people can install from, in which case, I'd be more aligned with SiFuh
<mpark>
because if there's a binary repository, at that point, compilers become optional
<farkuhar>
I seem to recall that the earliest CRUX releases didn't steer users toward compiling from source, they were all based on binary packages for i686. Only later did pkgmk and the ports collections become part of the default install.
<mpark>
but so long as they are required for installing of software, they are required in the core IMHO
<mpark>
the instructions make no mention of binary repos that I saw.
<mpark>
in any case, I wasn't around for the early releases of crux, so I have no idea of how things used to be.
darfo_ has quit [Ping timeout: 276 seconds]
<farkuhar>
No binary repos nowadays, but in the early 2000s there must have been. Not everybody had the luxury of hardware fast enough to compile every package from source.
<farkuhar>
Publicly-listed binary repos, that is. You're certainly welcome to spin up a binary package server for your local network, using pkg-get from the opt collection.
<mpark>
for me, binary repositories are a necessity, at least a local binary repository, so I can compile on one machine and then the rest can just install.
<mpark>
If I were better off financially, I'd spin up a public binary repository people could install from, but alas, I don't want to pay for all that bandwidth
<mpark>
maybe one day when I'm rich and famous :D
<mpark>
I'll have my own foundation with big company execs on the board to pay me to speak in public, LOL
<farkuhar>
The packages might be crippled or otherwise lacking in functionality, if you only built them in a clean container with the minimum of dependencies. Ports these days are written very flexibly, able to inspect what's available on the system and activate those optional features accordingly. So the binary packages built on one machine might be useless on somebody else's system.
<mpark>
My normal package build setup for my arch linux installs is a base arch linux container with a build script that only installs that packages dependencies, and builds it, then copies the package to a repo folder and shuts down destroying the container so every build is clean.
<mpark>
and that way I'm sure what dependencies the package actually has not the ones that were missing when the original packager built the package on their host.
<farkuhar>
For a binary package repo on your local network, that's not an issue; you have greater control over the systems where the binary packages will be deployed, and you can ensure uniformity. But hosting a public binary repo for the wider world is more prone to breakage.
<mpark>
oh yeah, which is why I'd only do it if I were a lot richer than I am now.
<mpark>
me being rich and famous was a joke to say that I'm not seriously considering it.
<mpark>
no, I don't forsee any public binary repos from me in the near future.
<mpark>
but I may start a blog and do a write up on how to create your own.
<mpark>
when I do eventually get to that point in my exploration of crux
<farkuhar>
Blog about pkg-get? Already done by ppetrov^ (slackalaxy.com)
<mpark>
farkuhar: yeah, but when it comes to blogging, there's no such thing as too much duplicated effort HAHAHAHAHQA
<mpark>
and also, his blog won't cover my experiences :D
darfo has joined #crux
lavaball has quit [Remote host closed the connection]
<ppetrov^>
mpark, go ahead, i'd be interested to read
<mpark>
I just spun up a quick blog at wordpress, because it's free and I'm not paying money unless I really get into it. In which case, I'll self host.
<mpark>
ppetrov^: LOL - I will probably do that when I go down the packaging rabbit hole.
<ppetrov^>
i am not promoting my own humble blog, but it might be useful
<cruxbridge>
<pitillo> and that stuff for old ARM devices is something really interesting too... taking note here to probably add a link in the documentation
<bleb>
not seeing anything for april, march, feb, jan, or dec '24
<bleb>
usually there's a way to see all the threads across multiple months, no?
<bleb>
is this a usability innovation brought to us by python 2 deprecation?
<cruxbot>
[contrib/3.7]: libvirt-python: updated to version 11.2.0
<cruxbot>
[contrib/3.7]: libvirt: updated to version 11.2.0, added new dependency libnl
<cruxbot>
[contrib/3.8]: Merge branch '3.7' into 3.8
<cruxbot>
[contrib/3.8]: libvirt-python: updated to version 11.2.0
<cruxbot>
[contrib/3.8]: libvirt: updated to version 11.2.0, added new dependency libnl
Burna has joined #crux
tilman has quit [Ping timeout: 248 seconds]
tilman has joined #crux
<Burna>
Gave up trying to get 32-bit iris driver to compile on 3.8. Reinstalled from 3.7: llvm-32#15.0.3-1.pkg.tar.gz mesa-32#22.2.3-1.pkg.tar.gz. This seems to be working with wine.
<Burna>
Sometimes the quickest path is through the mud
<Burna>
32-bit Slackware current doesn't have it, either.