jaeger changed the topic of #crux to: CRUX 3.6 | Homepage: https://crux.nu/ | Ports: https://crux.nu/portdb/ https://crux.ninja/portdb/ | Logs: https://libera.irclog.whitequark.org/crux/
ocb has quit [Remote host closed the connection]
ocb has joined #crux
SiFuh has quit [Ping timeout: 244 seconds]
SiFuh has joined #crux
darfo has joined #crux
<elderK> Nice work updating ncpamixer and babl, guys.
<elderK> Unfortunately, babl will fail to build if you have git installed.
<elderK> If you remove git, then install babl, everything appears happy.
<SiFuh> elderK: Yeah it was already in the CRUX Digest
<SiFuh> To quote the line from the email "My workaround was to uninstall git, update babl, reinstall git."
Guest23 has joined #crux
Guest23 has quit [Client Quit]
ocb has quit [Ping timeout: 268 seconds]
ocb has joined #crux
guido_rokepo has joined #crux
farkuhar has quit [Remote host closed the connection]
ardo has quit [Read error: Connection reset by peer]
ardo has joined #crux
elderK has quit [Quit: Connection closed for inactivity]
groovy2shoes has quit [Ping timeout: 255 seconds]
groovy2shoes has joined #crux
farkuhar has joined #crux
<farkuhar> SiFuh: I find it interesting that they called it "subproject improvements" to add such an imprecise test for the usability of git, https://gitlab.gnome.org/GNOME/babl/-/blob/4383ee8693c6cfc1c11214cd294c9648fb361daa/meson.build
<farkuhar> Earlier versions of meson.build did check for the presence of a git executable, but they never followed up by running the command that was found. The Feb 2021 blob was the first one to have run_command(git_bin,...
<SiFuh> Yeah I find that offensive. It should have been called the 'slaveproject'
<SiFuh> Anyway, I do not care about gegl or babl.
<SiFuh> It is out of my league
<farkuhar> Merely finding a git executable in $PATH is not enough to ensure that the source tree was actually downloaded by git. If I cared about babl as much as elderK does, I would file a bug report and ask for a sharper test before run_command(git_bin,...
<SiFuh> I don't think elderK cares either. I just think he was bringing the issue to our attention ;-)
<SiFuh> I've never filed a bug report myself. No point, last serious issue took more than 2 decades to repair
<SiFuh> 16 years sorry
<farkuhar> You just have to build up enough of a reputation for the devs to take you seriously. Look how quickly beerman's pull request got merged for the useradd man page.
<SiFuh> I used CRUX before anyone here
<SiFuh> Reputation my ass, just based on biases due to animoisty by members who don't like being told off for insulting other members.
<frinnst> <SiFuh> I used CRUX before anyone here
<frinnst> Did you know Per?
<SiFuh> Yes
<SiFuh> He is the one who refused to fix the serious issue
<farkuhar> anyway, kudos to elderK for bringing it to our attention, but the babl build failure is really a problem with meson.build that should be fixed upstream. CRUX maintainers should not be burdened with hacking babl's broken meson setup.
<SiFuh> As I said, it has already been mentioned in the CRUX mailing list
<farkuhar> And the workaround on the CRUX mailing list was to uninstall git first?! Just because it works doesn't mean it's the best solution.
<SiFuh> Yes, I quoted it above
<SiFuh> Tell that to ffmpeg
<farkuhar> just curious, what was the serious issue that Per refused to fix?
<SiFuh> So many ports that rely on ffmpeg are so slow to update to ffmpeg3 or ffmpeg4
<SiFuh> Drop down to init 1 and then back to init 3 and the system crashes.
<SiFuh> Every distro since 0.91 had the issue and I brought it up right up until 1.2 then gave up until around 2018 :-P
<SiFuh> Any good Rally Games equivelent too or better than Colin Mcrae Rally that are available for linux but are NOT steam related
SiFuh has quit [Ping timeout: 256 seconds]
SiFuh has joined #crux
guido_rokepo has quit [Quit: guido_rokepo]
<frinnst> SiFuh: what was the issue?
<frinnst> that per refused to fix, that is
<SiFuh> Ahh, something in the init scripts
<SiFuh> don't worry, jaeger actually brought it to the teams attention and had one of the core team fix it.
<frinnst> uh, when Per ran things there was no "team" :-)
<SiFuh> Per didn't like me much
<frinnst> well, i suppose there was
<SiFuh> So he cared not what I said
<frinnst> well you do have a certain way about you :D
<SiFuh> I'm harmless but I don't like bullies
<frinnst> my crux work never overlapped with Per's so I never really spoke to him
<frinnst> i got involved in 2006-2007 or something
<frinnst> dec 27 2008 crux-2.5-x86_64-3.iso
<SiFuh> Everything before 1.0 was considered RC
<SiFuh> https://crux.nu/gitweb/?p=ports/core.git;a=blob;f=rc/rc.single;h=07c2f51c02a2db6ff40143167963e6374491c2dc;hb=9e5123a0de9912315c581b9194e68fb7d3e7610a
<SiFuh> Here is the fix, says jaeger did it
<SiFuh> Jun 2019 HAHAHA not 2018 I was cclose
<SiFuh> Oopspatch
<SiFuh> https://crux.nu/gitweb/?p=ports/core.git;a=blobdiff;f=rc/rc.single;h=07c2f51c02a2db6ff40143167963e6374491c2dc;hp=4291eb0b73b90694a1ac5442722fd50105e550ca;hb=9e5123a0de9912315c581b9194e68fb7d3e7610a;hpb=0d0033642eb20eafa3cae93ef92bad5105f8c48d
<SiFuh> That is the patch
<farkuhar> SiFuh: the Jun 2019 fix inserts the same tests that stenur and ivandi have been pushing for in /etc/rc (test whether /proc and /sys are already mountpoints before running mount).
<SiFuh> Wow, they were pushing it in what century?
<farkuhar> 2022-07-30 ivandi wrote: it would be nice to have this in /etc/rc: /bin/mountpoint -q /proc || /bin/mount -t proc none /proc, same goes for sys. w/o it there is an error message when using dracut
<SiFuh> So the 21st century?
<SiFuh> HAHA
<SiFuh> In fact I remember the fix had a '||' because
<SiFuh> Guess it was changed
<SiFuh> This is reminding me of login.defs that jue suggested should be the default back in February 2012 :-) That you are all concerned with during the past week
<SiFuh> To be 100% frank and I don't care if it offends anyone. The core is priority not the bleeding edge updates of ports. The focus on CRUX is in the wrong place. Seriously.
<SiFuh> It takes a few days to fix a broken port but decades to fix the main CRUX system/base/core
<farkuhar> Heh, at least you got your fix in rc.single. I wonder if stenur or ivandi have the patience to wait 16 years for /etc/rc to catch up.
<SiFuh> This should be priority and there should be a priority base based on the actual core and CRUX
<SiFuh> By the way farkuhar I didn't tell you but I have been using the default login.defs for awhile.
<farkuhar> You mean the default that comes from upstream, or the default that we inherited from jw and gradually modified by adding SYS_UID_MIN and SYS_GID_MIN?
<SiFuh> I use a slightly modifed one
<farkuhar> So the one that comes from the shadow project, not the one in our ports tree?
<SiFuh> UID_MIN 1000 GID_MIN 100 <-- don't touch this and the yes one let me check
<SiFuh> USERGROUPS_ENAB yes
<SiFuh> Correct
<farkuhar> By the way, SiFuh, how did you dig up the Feb 2012 quote from jue? Did you grep through your IRC logs for login.defs?
<SiFuh> Since 3.5
<SiFuh> I have logs for every conversation I ever had or online
<SiFuh> I grepped it
<SiFuh> Not a bot, so many chunks will be missing when I am offline
<SiFuh> I use to have a server running for myself back in Australia and a highly agressive brother that plays games online that would pull the plug to my server because it must be me that runs IRC that causes his lag
<farkuhar> The NetBSD thread that you and stenur were analyzing last week ... kre was worried about attracting the wrong kind of users (ones who would need support because they couldn't figure out PATH by themselves), and a small project just doesn't have the resources to provide all that support.
<SiFuh> It is okay now, he a prophet of god and his dreams are prophecies and our dreams are contracts with demons. The Earth is flat, a dome is over it and the stars in the sky are governments shining torch lights on it.
<SiFuh> I didn't analyse NetBSD thread. I checked the shadow file of the NetBSD source to see what the group file had 'users' set as
<SiFuh> I just read the thread
<farkuhar> https://marc.info/?t=166042424200006&r=1&w=2 for reference. Anyway, it reminded me of that thread when you wrote that "The focus on CRUX is in the wrong place."
<SiFuh> I am aware of the thread
<SiFuh> But I had mentioned something to Romster :-(
<SiFuh> And he agreed as well
<SiFuh> Why are we updating ports every second it changes a version number?
<SiFuh> It should be changed only when you either get around to it, or it is a serious security issue
<SiFuh> CRUX seems to be focused on daily updates of everything but the main system itself
<SiFuh> stenur's English is interesting.
<SiFuh> 'Yeah, do not make turtle soup, let the turtles do smalltalk.'
<farkuhar> https://crux.nu/Wiki/FaqPortsSystem says right there in point 7, it's considered a bug when a port in core, opt, or xorg is out of date. Some maintainers go even further and try to keep up with all the ports in contrib!
<SiFuh> I agree with that, but I think core should be 100% priority
<SiFuh> opt and xorg second
<SiFuh> Contrib should be on the bottom of the list
<SiFuh> And an 'out of date' port is NOT a bug
<farkuhar> participants in that NetBSD thread were worried about overwhelming their small team with support requests, if the userbase grows too rapidly. Meanwhile our FS#1410 shows concern about a *shrinking* userbase, because the bleeding-edge software that people want is not available.
<SiFuh> If I had a threadripper or a decent PC, I would compile over and over again, everything core, opt, xorg, contrib to make sure nothing broke
<SiFuh> I beleive that those 4 repos are needed to be perfect
<SiFuh> I would also provide binary packages of all
<farkuhar> so I would caution the folks who chimed in to FS#1410, be careful what you wish for! Making a distro with too broad an appeal might overwhelm the team with support requests from an influx of new users.
<SiFuh> Okay
<SiFuh> Who are the two guys that did that unto themselves ?
<farkuhar> https://crux.nu/Main/Mantra ... it says right there: We don't want to build the Linux distribution for the masses.
<SiFuh> But CRUX isn't
<SiFuh> But two particular CRUX uses exploded the scene and then got way to busy to handle the load
<SiFuh> SHow you something that pisses me off I can get this info
<SiFuh> That
<SiFuh> That shits me so much, 15 freaking ports for the same port (Exaggerating)
<farkuhar> beerman already flagged that problem in FS#1410: "There are a few people willing to whip together a repository, often one timers, they then abandon those ports and feel like they did something good."
<SiFuh> And those who die?
<SiFuh> I like the idea jaeger has to tar ball them all
<SiFuh> But I think there should be a repository where you can search the tar balls
<SiFuh> How about this
<farkuhar> if a repo maintainer meets the reaper, and they stop paying their web hosting bills, eventually the site will stop responding to httpup requests and the repo will be removed from the portdb, right?
<SiFuh> There is a user contributed repo that can be reviewed but a specific core team developer
<SiFuh> You know jaeger's unoffical port db does a pretty good job
<farkuhar> there was talk of making that the official portdb, in fact.
<SiFuh> So basically a repo is made over the user repos. Anything green keep everything else annihilate.
<SiFuh> If duplicates appear then a human can review and choose
<SiFuh> I am a huge believer though that all ports MUST be built on a core only system
<SiFuh> I am very sorry but this is a perfect example
<SiFuh> That repo will be instantly rejected because nothing green
<farkuhar> https://lists.crux.nu/pipermail/crux/2020-June/006617.html Here's where beerman proposed making the crux.ninja portdb the official one.
<SiFuh> I am sure a few of us guys could do the test builds
<SiFuh> It isn't difficult to have scripts that build it
<SiFuh> Hans :-)
<farkuhar> mistrbzr is the nick I remember
<SiFuh> WTF am I reading?
<SiFuh> My personal thoughts on how to make CRUX look more attractive: <-- everything after this line, is to the point
<SiFuh> Yeah I agree
<SiFuh> Except the Advertising part
<SiFuh> If it is good, it will be found
<farkuhar> you'd think if he wanted prt-get.aliases advertised more prominently, he would have kept rust-bin. I was editing that section of the handbook to include *two* examples, and when rust-bin got dropped I retreated to just one example.
<SiFuh> For them *-bin is precompiled binary
<SiFuh> For me *-bin belongs in /opt because it is binary :-P
<SiFuh> Anyway, I don't care. I have full faith in TimB
<SiFuh> Pissed off my port was stolen added to core and no one asked me advice about though ;-)
<SiFuh> I even had an updated version that I had not uploaded yet, but the core version was my old port.
<farkuhar> you mean libuv?
<SiFuh> Wrong channel?
<farkuhar> Which port of yours was stolen and added to core?
<SiFuh> nftables
<SiFuh> I was pissed off because not only did I try to push it back in 3.5 and everyone was like "wow", "ooh", "ahh" but they didn't even bother to ask me or even check if an update existed
<SiFuh> I was using nftables back in Slackware, and the Void before I ended up making a port for CRUX and aparantly no one had heard or if they had no care for it, now it is in core.
<SiFuh> And then it is now in core, but the person who took it and moved it to core was so lazy they never bothered to check if an update existed. Which I was working on and hadn't pushed yet
<farkuhar> I'm surprised you got such a unanimous reaction. Surely there must have been some holdouts who were wary about abandoning the firewall that was well-established at the time (iptables?)
<SiFuh> They were pulling patches out of the arse trying to make it work. Why? If they bothered to check the updated version of nftables you'd not need to pull patches out of your arse
<SiFuh> To be honest I have always hated Iptables.
<SiFuh> It is, hmm.. It is a gordian knot
<farkuhar> It must have been a common sentiment, judging by the reaction you got when you spread the word about nftables.
<SiFuh> It is a huge twisted mess of bullshit. Nftables made it much better, and is probably the closest Linux has to OpenBSD's PF
<SiFuh> It was in 2019
<SiFuh> Don't get me wrong though. I like nftables and I am glad it is in core now. I just wish whomever poached it actually did the research rather then stealing a port
<SiFuh> I believe it should be in core
<SiFuh> This brings me back to our very first discussion. CORE should be perfect
<SiFuh> CRUX must be so solid slackware and debian turn to mush
<SiFuh> I beleive there should be a few guys who strictly work on CRUX core and only CRUX core.
<SiFuh> farkuhar: I reckon the semi-unanimous was by those who are too old to change
<SiFuh> Bet they all use GRUB :-)
<SiFuh> HAHAHAHAHA farkuhar I want to buy a 60 litre boiler, so I put in an order at 400MYR. They cancel and say out of stock. Same company have the same boiler for 700 MYR so I order and they cancel saying out of stock. Same company have boiler at 800 and I order and they say out of stock. I call my friend and he order from same company at 900 MYR and they have plenty. He cancelled ordered
<SiFuh> This is common in Asia they fish you in. Try to direct you to higher and higher price.
<farkuhar> Definitely a culture shock if you're coming from a place where such negotiating tactics are frowned upon.
<SiFuh> Culture shock?
<SiFuh> Sorry I have lived in Asia more than half my life
<SiFuh> Cultrue shock is a fart to me ;-)
<SiFuh> I probably get more of a culture shock going home to visit my parents :-P
<SiFuh> Woah, too much text
pehaef has joined #crux
pehaef has quit [Client Quit]
<stenur> a0/QUIT
<stenur> sorry
tilman has quit [Ping timeout: 248 seconds]
tilman has joined #crux