furrywolf | to be a thing, they first have to not suck. :) | 00:03 |
---|---|---|
_abc_ | fsmithred: you need the whole kernel tree to run config on, then you can build just the modules I think | 00:14 |
fsmithred | I guess I'm thinking of modules you build with third-party code | 00:14 |
buZz | fsmithred: like https://github.com/lwfinger/rtl8723bu ? :P | 00:15 |
buZz | still supports dkms ;) | 00:15 |
buZz | but also works without so w/e | 00:16 |
Centurion_Dan | anybody give a quick opinion on using gnome-keyring or something of similar ilk... in jessie it made a mess of gpg key support... | 05:44 |
djph | Centurion_Dan: I try to avoid it. | 11:22 |
Centurion_Dan | djph: so do I but unlocking my ssh key 10 times when wanting to access a vm console in virt-manger is a bit tedious... | 11:58 |
KatolaZ | Centurion_Dan: ssh-add | 12:03 |
KatolaZ | :) | 12:03 |
djph | Centurion_Dan: ^ | 12:07 |
djph | ssh-agent >>> gnome-keyring | 12:07 |
KatolaZ | never used gnome-keyring, but ssh-agent is more than enough | 12:08 |
man_in_shack | https://www.youtube.com/watch?v=dc8O6Sg-zzU | 12:32 |
slphil_n6 | I recently got Devuan running on my Samsung Chromebook ARM, using a 4.18.13 kernel. It's a cheap and useful device. Is there a place for me to document the process for anyone else? | 13:19 |
djph | there is the mailing list, or perhaps the gitlab pages ... there was a forum at one time, but it seems to have never really taken off -- mailing list is considerably more active. | 13:21 |
Centurion_Dan | slphil_n6: dev1galaxy.org | 13:33 |
fsmithred | Centurion_Dan, someone with ryzen issues - maybe you can help? https://dev1galaxy.org/viewtopic.php?id=2418 | 13:34 |
fsmithred | djph, there is an active (and official) forum ^^^ | 13:35 |
djph | yay | 13:53 |
Diziet | Hi folks. pm215 and I have been looking at improving the situation with non-systemd in Debian buster. | 14:25 |
Diziet | Currently it's in poor shape for desktop stuff. We've been looking at using elogind and have been making some progress, based on the elogind package from devuan. | 14:25 |
Diziet | We don't have something tidy or even fully working yet but ISTM that it would be good to chat about how to organise things so that we (Devuan on the one hand; and Debian non-systemd folks on the other) have the best cooperation. | 14:26 |
Diziet | Right now we have been treating the elogind package from devuan (234.4-1+devuan1.1) as an upstream. | 14:28 |
fsmithred | Diziet, I see 234.4-1+devuan1.4 in experimental and newer version, 234.4-2 in ascii. | 14:39 |
Diziet | I think we just haven't rebased since what I picked up a week or two ago. | 14:40 |
* Diziet does git fetch to see whats new | 14:40 | |
fsmithred | you probably want to talk to the maintainer, amesser (or close to that name) | 14:40 |
fsmithred | some stuff is out of order in the repos, because we were scrambling to get ascii out | 14:41 |
fsmithred | and bypassed the normal exp->unstable->testing route | 14:41 |
Diziet | Oh, I hadn't spotted the experimental branch. I think I am being confused by the git layout, but I have found the +devuan1.4 tag. | 14:42 |
fsmithred | I'm using elogind in the latest refracta xfce release | 14:42 |
fsmithred | ah, ok | 14:42 |
fsmithred | check ascii branch | 14:42 |
Diziet | I think we probably want 3eb67a27103b4c4e534c467519a243bcee64b5e2 Disable automatic user process killing by default | 14:44 |
fsmithred | if that means what I think it means, then yeah, you probably do | 14:44 |
Diziet | Right now our most obvious bug is that in our wip configuration with lightdm, the shutdown/reboot options are greyed out. | 14:45 |
pm215 | plus gdm doesn't work at all... | 14:45 |
fsmithred | yeah, in refracta I dropped lightdm and replaced it with lxdm | 14:45 |
Diziet | fsmithred: Should we be talking to Andreas Messer by email do you reckon ? | 14:46 |
fsmithred | power buttons work, language switcher works | 14:46 |
fsmithred | yeah, or post on devuan-dev mailing list | 14:46 |
fsmithred | might get more people involved with the latter | 14:46 |
Diziet | And do you think it is sensible for us to try to treat devuan as upstream for elogind ? It looks like it probably is. | 14:47 |
MinceR | elogind :( | 14:47 |
fsmithred | afaik, we're planning on keeping it | 14:47 |
Diziet | MinceR: apologies if this is reopening some flamewar, but ... what would your preferred approach be ? | 14:49 |
MinceR | dunno | 14:49 |
MinceR | personally i just use startx | 14:49 |
MinceR | and look at the developments in despair | 14:49 |
MinceR | i suppose consolekit is not an option anymore | 14:49 |
fsmithred | see ascii release notes section on session management and policykit backends: https://files.devuan.org/devuan_ascii/Release_notes.txt | 14:49 |
fsmithred | and yeah, the desktop landscape is rough in non-systemd-land these days | 14:50 |
MinceR | it's rough in systemd-land as well | 14:50 |
fsmithred | MinceR, consolekit is still default with most of the dekstops | 14:51 |
MinceR | because of systemd | 14:51 |
fsmithred | I hear mixed reports about that - some don't notice it because everything works, and some can't proceed because something got screwed up | 14:51 |
MinceR | eventually they'll notice | 14:52 |
MinceR | just as they noticed it in windows | 14:52 |
fsmithred | yup | 14:52 |
MinceR | GNU/Linux will probably be gone by then, though | 14:52 |
fsmithred | I got the five-minute boot wait a few times and abandoned it | 14:52 |
pm215 | startx needs a logind in debina buster, incidentally -- otherwise it doesn't get read access to the /dev/input/* for the mouse (there are ways to configure X or your device perms to avoid that, of course) | 14:53 |
* man_in_shack replaces MinceR's kernel with hurd | 14:54 | |
MinceR | ew | 14:54 |
MinceR | hurd is dead | 14:54 |
man_in_shack | well you seem to enjoy living in a fantasy world | 14:54 |
Diziet | MinceR: I deprecate all statements of the form "foo is dead" (unless you're the maintainer of foo who is killing it, I guess) | 14:54 |
man_in_shack | <MinceR> GNU/Linux will probably be gone by then, though | 14:54 |
fsmithred | pm215, alternative is xserver-xorg-legacy | 14:54 |
man_in_shack | gonna be replaced by sco xenix i guess | 14:54 |
fsmithred | and run X as root | 14:55 |
fsmithred | works in beowulf (buster) | 14:55 |
MinceR | more likely to be replaced by BSD-s | 14:55 |
man_in_shack | well fine | 14:55 |
Diziet | "foo is dead" is the pattern everyone uses to get people to use their own thing, or for dropping foo compatibility, or whatever. | 14:55 |
filipdevuan_ | hey is intel management engine a real issue?? or its just made up by FSF?? | 14:55 |
man_in_shack | install sco xenix userland with a netbsd kernel then | 14:55 |
MinceR | ok, then, i'll use a different pattern, to make it easier to understand | 14:55 |
MinceR | https://www.gnu.org/software/hurd/open_issues/systemd.html | 14:56 |
MinceR | also note that prior to this step, GNU/HURD used to be a multiserver microkernel system | 14:56 |
slphil_n6 | Systemd truly tries to worm its way into everything | 14:59 |
MinceR | yet there's no shortage of defenders for it | 15:00 |
MinceR | i ran into them today as well | 15:00 |
filipdevuan_ | hey can u see me posting?? | 15:00 |
sixwheeledbeast | it is an issue yes | 15:01 |
man_in_shack | there are exactly 0 people in here who like systemd, MinceR | 15:02 |
man_in_shack | what's your point? | 15:02 |
MinceR | not here | 15:02 |
man_in_shack | so? | 15:02 |
man_in_shack | you're the one who brought it up | 15:02 |
filipdevuan_ | hey is intel management engine a real issue? or its just made up by FSF??? | 15:03 |
MinceR | it's a real issue | 15:03 |
MinceR | and i'm surprised that red hat/nsa hasn't paid them enough to silence them about it | 15:04 |
man_in_shack | it's got an entire minix kernel in it | 15:04 |
filipdevuan_ | so i should feel quilty that i spend like 500 dollars for useless acer laptop with optimus nvidia graphics card instead of buying fsf endorsed laptop ;/ | 15:04 |
filipdevuan_ | spent* | 15:05 |
man_in_shack | obviously | 15:06 |
filipdevuan_ | im going on devianfork channel, have a nice day :) | 15:06 |
man_in_shack | https://www.quiltingcompany.com/store/quilty-january-february-2018-digital-edition << feel VERY quilty | 15:07 |
man_in_shack | filipdevuan_: honestly the worst thing about your choice of laptop is the acer part | 15:07 |
filipdevuan_ | yeh. it's what happens when you go to supermarket and know nthing about laptops and they advice you to buy acer laptop ;/ | 15:08 |
man_in_shack | to paraphrase summer from rick & morty: nvidia has binary blob drivers and intel runs a secret kernel in their cpu. is there a tech company that ISN'T evil? | 15:08 |
man_in_shack | asus is about the only laptop brand i trust | 15:09 |
man_in_shack | toshiba, acer, hp, msi, all junk | 15:10 |
filipdevuan_ | i paid like 600 pounds for this laptop just because of 1tb hdd and some 940mx geforce i5 i7 and now i cant sell it for more than 150 pounds | 15:10 |
filipdevuan_ | to buy something different | 15:10 |
man_in_shack | samsung were nice back when chromebooks were a thing | 15:10 |
filipdevuan_ | whats so different in asus that is different in hp or acer?? | 15:11 |
man_in_shack | build quality | 15:13 |
man_in_shack | and cooling | 15:13 |
man_in_shack | acers snap in half if you look at them the wrong way | 15:14 |
man_in_shack | hps just bend when you type on them | 15:14 |
filipdevuan_ | oh yeah acer cooling sucks :P. however i dont think asus would be any different if it still has intel management engine | 15:14 |
man_in_shack | filipdevuan_: it's more about the cooling hardware design | 15:14 |
man_in_shack | heatpipe layout, fans | 15:14 |
man_in_shack | asus knows more about that shit | 15:14 |
man_in_shack | oh, and people keep telling me lenovo are fine but i never believe them | 15:15 |
man_in_shack | give me a modern system retrofitted into a proper ibm thinkpad case any day | 15:15 |
man_in_shack | just without the goddamn mouse nipple | 15:15 |
MinceR | dHells break themselves apart from the inside | 15:15 |
MinceR | out of lenovo, only thinkpads are fine | 15:16 |
filipdevuan_ | well im sad because i spent too much money for this laptop i wanted it to play some modern games but in the meantime i stopped using windows 10 and stopped playing modern games as well and i feel like my laptop is just junk now ;/ | 15:16 |
man_in_shack | i know them feels | 15:16 |
MinceR | IT is increasingly garbage | 15:17 |
MinceR | everything is going to hell | 15:17 |
man_in_shack | turn it into a htpc? | 15:17 |
djph | MinceR: indeed | 15:17 |
filipdevuan_ | are there any models of laptops you would recommend just for devuan that would be quite safe too?? | 15:18 |
man_in_shack | short answer is they all have the same stuff on the inside these days | 15:19 |
filipdevuan_ | even all asus laptops have intel management engine installed?? | 15:19 |
man_in_shack | yah, it's unavoidable | 15:19 |
man_in_shack | only way out of it is to use an amd cpu | 15:19 |
man_in_shack | which had an embedded cortex a7 in it instead | 15:20 |
man_in_shack | (: | 15:20 |
filipdevuan_ | but iv read that amd has management engine as well but just under different name | 15:20 |
man_in_shack | yup | 15:20 |
man_in_shack | they do | 15:20 |
man_in_shack | and it hasn't been leaked like the intel one was | 15:20 |
man_in_shack | basically it's a sideeffect of microsoft's bullshit "secure boot" uefi shit | 15:20 |
filipdevuan_ | i basically feel like selling this acer for i dont know how much to buy something cheaper. im not happy with the acer. | 15:21 |
filipdevuan_ | and i dont really like intel as well... | 15:21 |
man_in_shack | no one likes intel | 15:21 |
filipdevuan_ | id get something without windows 10 preinstalled | 15:21 |
filipdevuan_ | i wonder if somebody would paid me 300 dollars for this laptop ;/ | 15:21 |
man_in_shack | donate it to a good cause | 15:22 |
man_in_shack | post it to computerbank.org.au :P | 15:22 |
filipdevuan_ | yeah but then im gonan be without a laptop i havent got any particular income... i wanna use devuan but my laptop is useless because iv read its very insecure, its acer, it has nvidia geforce 940mx that i dont need on devuan, im fed up with non free stuff and setting bumblebee up for optimus | 15:22 |
man_in_shack | filipdevuan_: it's no more insecure than any other laptop | 15:23 |
filipdevuan_ | so there arent any choices in terms of security apart from FSF endorsed laptops?? | 15:23 |
MinceR | the best i can recommend is thinkpad | 15:24 |
man_in_shack | i mean you can read all this scary stuff about spectre and meltdown bugs affecting fuckloads of cpus from amd, intel and arm | 15:24 |
man_in_shack | but they still require root access software to initiate them | 15:24 |
man_in_shack | so don't be an idiot with your software and your hardware will remain safe | 15:24 |
filipdevuan_ | oh okay so its not like i have IEM and devuan and my laptop is insecure like i mean very insecure?? is IEM just about remote access or its about mining data about users as well?? | 15:25 |
filipdevuan_ | IME* | 15:26 |
man_in_shack | spectre and meltdown were proofs of concepts | 15:26 |
MinceR | with IME, you're insecure if you use the ethernet port on a public network | 15:26 |
MinceR | since it will just listen to commands from the network, regardless of what you do | 15:26 |
man_in_shack | i'm pretty sure they're actually very difficult to exploit in any usable way | 15:27 |
man_in_shack | ^ | 15:27 |
filipdevuan_ | so do you think i should sell this acer for cheaper than i bought and buy something more specific or remain there with acer | 15:28 |
filipdevuan_ | cuz i feel like my laptop is pointless now to have cuz i dont use windows 10 i dont play modern games nor i need this geforce 940mx | 15:29 |
man_in_shack | well unless it's cooking your desk/lap when you turn it on, you can still do things with it | 15:29 |
man_in_shack | define "modern games"? | 15:29 |
man_in_shack | there's a whole bunch of games even on steam that'd work find on a 940mx on a linux system | 15:30 |
MinceR | modern, like systemd! :> | 15:30 |
filipdevuan_ | uhh this is the thing i dont like steam as well i dont trust them i think theyre not less evil than google and microsoft | 15:30 |
man_in_shack | hehe fair enough | 15:31 |
filipdevuan_ | but anyways thanks for advice :) | 15:31 |
filipdevuan_ | uhh this 940mx is ok for devuan too but bumblebee on devuan is quite weird to set up | 15:31 |
man_in_shack | so best advice for something to do with it would be htpc | 15:31 |
filipdevuan_ | okay htpc is home theather pc?? | 15:32 |
man_in_shack | chuck kodi on it, put a bunch of movies on that 1TB hdd | 15:32 |
man_in_shack | plug it into your 85" tv | 15:32 |
filipdevuan_ | uh its ok my mate has lots of dvd movies and bluerays and 5.1 home theatre hehe :P cant get any better is kodi free software btw | 15:33 |
man_in_shack | it's in debian main repo i think | 15:33 |
filipdevuan_ | oh okay i can see its open source, okay cheers then | 15:33 |
man_in_shack | yah | 15:34 |
man_in_shack | i got it running on my rasbpi | 15:34 |
man_in_shack | which is still running systemd-infected debian | 15:34 |
telst4r | devuan works on raspi 1 at least :) | 15:34 |
man_in_shack | but it's also pulled in deb-multimedia.org repo so i dunno (: | 15:34 |
man_in_shack | yah | 15:35 |
man_in_shack | this is rasbpi v3 | 15:35 |
man_in_shack | my rasbpi v2 to destroyed by floofkitty pissing on it | 15:35 |
man_in_shack | *got destroyed | 15:35 |
telst4r | aren't there rpi 3 images too? | 15:35 |
filipdevuan_ | yeah i know gnu linux users are hyped about steam and games but i dont really like steam :( | 15:36 |
filipdevuan_ | i think they spy on you | 15:36 |
MinceR | i don't like steam either, as it's a DRM platform | 15:36 |
man_in_shack | they do to a certain extent | 15:36 |
man_in_shack | ^ | 15:36 |
MinceR | and the best games i've ever played are free software, at least for the code | 15:36 |
man_in_shack | they certainly keep close track on what games you play, when, and use that for direct marketing | 15:36 |
man_in_shack | which so far, they've got right like ONCE for me in the 2-3 years i've been using it | 15:37 |
man_in_shack | MinceR: apt install crack-attack | 15:37 |
filipdevuan_ | this is what steam does https://voat.co/v/technology/2475543 | 15:37 |
man_in_shack | so-called because it's as addictive as crack cocaine | 15:37 |
MinceR | i was thinking more of xonotic and various id game engine source ports | 15:37 |
man_in_shack | "Steam could be monitoring any files across the entirety of users hard drives" umm | 15:38 |
man_in_shack | i'm not interested in "could" | 15:38 |
MinceR | :> | 15:38 |
man_in_shack | "Visitors need only view your profile page" << which can be locked down | 15:39 |
telst4r | replace it with 'will' | 15:39 |
man_in_shack | although admitedly, not as much as i like | 15:39 |
man_in_shack | telst4r: will, one day, maybe, but currently doesn't | 15:39 |
telst4r | :) | 15:40 |
man_in_shack | "steampowered also employs common www.google-analytics.com as well as steamcdn-a.akamaihd.net and steamcommunity-a.akamaihd.net" very true | 15:40 |
man_in_shack | and the fucking in-client browser is a goddamn monstrosity | 15:40 |
man_in_shack | "Valve’s Steam phone app has been found to utilize four trackers and at least two sensitive application permissions" << and they FORCED you to use it if you wanted to do anything on the community market | 15:41 |
man_in_shack | i'm not sure if that policy has changed or my profile has changed, but i can now sell random shit i get without needing to confirm on my phone | 15:41 |
man_in_shack | "Support for payments of Steam software with Bitcoin was dropped, forcing users to now provide their address in order to make purchases." << that's not actually a bad thing in and of itself | 15:42 |
man_in_shack | every other goddamn company in the world asks for a billing address for online transactions | 15:42 |
man_in_shack | even going to a brick-and-mortar store, your bank still has your address | 15:42 |
MinceR | i could pay with cash in a brick-and-mortar store | 15:43 |
man_in_shack | "Steam users have been deceived into opening malacious screenshots containing executable SCR files which target the Steam client" <<< HAHAHAHAHAHAHAHAHHAH IDIOTS | 15:43 |
man_in_shack | MinceR: which is LESS secure these days than using a visa | 15:44 |
man_in_shack | mastercard can go fuck themselves | 15:44 |
man_in_shack | but visa actuall has decent fraud prevention stuff | 15:44 |
MinceR | what's the difference between mastercard and visa? | 15:44 |
man_in_shack | "Valve can, and has in the past, removed games from Steam as well as retroactively revoked them from user’s libraries." << i still have a number of titles that no longer have store pages. it's fun :P | 15:45 |
man_in_shack | MinceR: mastercard are assholes if shit happens to your card | 15:45 |
man_in_shack | visa, less so | 15:45 |
MinceR | how less so? | 15:45 |
man_in_shack | i've had my details stolen a few times. bank be like "sure, here's your money back. we'll send this shit to visa" | 15:46 |
man_in_shack | the visa fraud insurance deals banks have protect them (and bank customers) very well | 15:47 |
man_in_shack | i don't know the details, but i've heard horror stories from mastercard holders about chasing up fraud on their cards | 15:47 |
man_in_shack | and never heard one from visa | 15:47 |
man_in_shack | "Steam chat operates a topic & language filter which cannot be deactivated and whose arbitrary target terms are not disclosed" << well that's your stupid fault for using fucking steam chat | 15:48 |
MinceR | ic | 15:48 |
MinceR | one should not rely on a corporation for a platform | 15:49 |
man_in_shack | "Steam suppresses reviews from users who did not purchase directly through Steam under the guise of improving review scoring" << that's actually valid. if you didn't get your product through steam, then your review might not reflect the product steam is selling | 15:49 |
man_in_shack | partly because publishers are assholes | 15:49 |
man_in_shack | partly because steam worms its way into games with acheivements and shit | 15:49 |
man_in_shack | "Under threat of permanently deactivating user accounts without refund, Steam forces new subscriber agreement upon gamers." << what would the alternative be? it's a content distribution service | 15:50 |
man_in_shack | "Steam prevents the purchase of certain software from within restricted regions." << not actually steam's fault. they're actually pretty good on this point. it's just they have to deal with every country's different classifications boards | 15:52 |
man_in_shack | "Steam “Offline Mode” first requires authentication and an internet connection in order to function. Even still, Offline Mode has proven less than reliable." << very true. "offline mode" is a complete lie | 15:53 |
MinceR | well, it would be nice if games "purchased" via steam would actually be "purchased" and not just "subscribed" for an unspecified amount of time | 15:56 |
man_in_shack | yes | 15:56 |
* man_in_shack throws ea's sim city remake nightmare at MinceR | 15:56 | |
man_in_shack | companies are afraid of customers owning products | 15:57 |
* MinceR parries it with a BFS | 15:57 | |
man_in_shack | in fact, they always have | 15:57 |
MinceR | customers can't own software products, only a license to them | 15:57 |
man_in_shack | "YOU'RE NOT ALLOWED TO LOOK AT HOW THIS SOFTWARE WORKS" has been in every eula since the invention of closed-source | 15:57 |
MinceR | yet another reason to stick to free software games | 15:57 |
* man_in_shack sells MinceR a brand new fresh copy of tuxracer | 15:58 | |
* telst4r changes the way man_in_shack's chess engine predicts | 16:00 | |
man_in_shack | pawn to king 4 | 16:01 |
MinceR | thermonuclear strike to d4 | 16:02 |
telst4r | GNU to A6 | 16:02 |
man_in_shack | THE ANGEL OF DEATH to e8 CHECKMATE | 16:08 |
telst4r | fffffuuuuuu---! | 16:15 |
* booyah waves at telst4r | 16:37 | |
booyah | all hail Bitcoin | 16:37 |
telst4r | such booyah | 16:49 |
telst4r | :) | 16:49 |
nemo | https://popcon.debian.org/stat/vendors.png devuan seems to drop off as "wazo" climbs | 18:25 |
nemo | are they both non-systemd? | 18:25 |
nemo | strange. how the heck is wazo climbing so fast. it seems so specific | 18:26 |
nemo | maybe opts in to popcon by default | 18:26 |
nemo | https://wiki.debian.org/Derivatives/Census/Wazo anyway. is systemd. so drop in devuan is perhaps simply due to it being a smaller percentage of the pie as something grows rapidly | 18:27 |
cehteh | lol about doing that on a logarithmic scale :) | 18:33 |
Ryushin | Are there any numbers regarding how many Devuan users there are? Just hoping the Devuan stays viable as I rely on it. | 18:34 |
nemo | Ryushin: funny that's exactly why I was looking at popcon | 18:35 |
nemo | well, also that and trying to figure out how many debian hedgewars users there were since Wuzzy saw fit to remove our local physfs copy fallback *AND* require physfs3 | 18:35 |
nemo | debian stable that is | 18:35 |
nemo | but at same time was like. man it sure would be nice if devuan was popular enough to ensure influence on upstream debian/kernel decisions | 18:35 |
Ryushin | nemo: I agree on all of those counts. | 18:38 |
Ryushin | Someone posted that the wind is coming out of the sails on Devuan. http://maemo.cloud-7.de/irclogs/freenode/_devuan-dev/_devuan-dev.2018-09-05.log.html | 18:40 |
Ryushin | Devuan has accomplished so much, it is staggering. | 18:41 |
cehteh | the best thing what could happen would be when devuan becomes obsoltete .. by a) technically: debian supports multiple init systems again and b) socially: people find a friendly/peaceful solution to end these rage and differences | 18:43 |
cehteh | or maybe eventually 'systemd' just works .. i mean really works in a way that everyone is happy, no bugs and no fuckups, no glitches | 18:44 |
MinceR | the current design and attitude is highly unlikely to lead to such a conclusion | 18:45 |
Diziet | ... ported to bsd. | 18:45 |
Diziet | Will be a few years. | 18:45 |
cehteh | yeah, i think all about that, very unlikely | 18:45 |
cehteh | but that would be the best outcome | 18:45 |
Diziet | I think (a) is reachable. I'm running sysvinit on my laptop with Debian stretch. buster definitely needs fixing. | 18:45 |
Diziet | I'm hoping Debian can borrow some of Devuan's answers there. | 18:46 |
cehteh | i am running some debian with systemd .. where it works and doesnt slap me in the face its ok | 18:46 |
cehteh | *but* often enough there are some subtle bugs which led me here | 18:46 |
_abc_ | Hi. I tried to install kernel-package to rebuild a kernel the debian way on ascii and it tried to pull in half the world, among others tex, texinfo and other monsters. Is there really no sanity left in debian? Does devuan intend to provide any simplified debian alternatives for key things like kernel rebuilds and so on? fsmithred ? | 18:47 |
cehteh | kernel-package is obsolete, modern kernel makefiles have some debian targets try make bindeb-pkg or make deb-pkg | 18:48 |
cehteh | kernel-package wants to build the documentation packages as well, which needs a shitload of tex infrastructure | 18:51 |
fsmithred | also, try installing with --no-install-recommends | 18:52 |
fsmithred | I'm not aware of any discussions of changing kernels or kernel build procedures | 18:52 |
_abc_ | Okay. I was sure there are workarounds. | 18:53 |
_abc_ | I will need TeX in the future but not right now. | 18:53 |
cehteh | i usually do the bindeb-pkg thing | 18:53 |
_abc_ | Ok. Right now I am building just modules using the running system's config. | 18:55 |
_abc_ | This is my real goal. | 18:55 |
fsmithred | ok, this may sound like a stupid question, but how do I install opengl? | 18:55 |
fsmithred | someone mentioned tux-racer, and I remembered planet penguin racer, which is no longer in the repo. | 18:56 |
fsmithred | so I'm trying to compile it from source, and it's complaining that it can't find opengl | 18:56 |
fsmithred | I don't know which package to install - it's not obvious from the names | 18:56 |
cehteh | you need the proper graphics drivers *AND* the respective -dev packages installed when compiling from source | 18:56 |
cehteh | that are tons of libs | 18:57 |
cehteh | mostly mesa and its dependencies | 18:57 |
fsmithred | yeah, I have mesa and mesa-dev stuff installed already | 18:57 |
cehteh | and i recommend the backports stuff, lots of bugs fixed | 18:57 |
_abc_ | trying to install qemu with aptitude is terrible. Is there a way to tell aptitude to do the equivalent of --no-install-recommends ? | 18:57 |
fsmithred | ok, but I'm trying to compile something from 2005 | 18:57 |
cehteh | then check the build/configure log what it exactl tests for | 18:57 |
fsmithred | _abc_, yes: -R | 18:58 |
_abc_ | fsmithred: -R where? aptitude cli? | 18:58 |
fsmithred | aptitude -R install stuff | 18:58 |
debdog | _abc_: you can turn that off permanently inside aptitude's ncurses ui | 18:58 |
_abc_ | I'm using the curses ui | 18:58 |
_abc_ | Oh, where? | 18:58 |
debdog | _abc_: Options -> Prefernces | 18:59 |
_abc_ | The option was off I think. | 18:59 |
_abc_ | Hm trying again. | 18:59 |
_abc_ | This is beyond stupid. qemu-system claims it depends on qemu-system-sparc etc | 19:02 |
fsmithred | that must be a metapackage | 19:03 |
_abc_ | What a clusterf*. I don't believe it. There is no way to force partial package install. I've run qemu for like 20 years I think, sometimes compiled from source. | 19:04 |
fsmithred | I have 5 qemu packages installed | 19:04 |
_abc_ | qemu depends on qemu-system and qemu-system depends on ALL systems. sparc, ppc all | 19:04 |
_abc_ | fsmithred: manually? Cli? | 19:04 |
fsmithred | yes | 19:04 |
fsmithred | I didn't install any graphical front-end for it. I just use a script I wrote, but all I use it for is to boot live-isos | 19:05 |
_abc_ | Of course. | 19:06 |
fsmithred | anyway: qemu-system-common qemu-system-x86 qemu-kvm qemu-utils ipxe-qemu | 19:06 |
fsmithred | and ipxe-qemu (not sure how that one got in) | 19:06 |
_abc_ | It wants to install all of the previous ones using cli and --no-install-recommends too | 19:07 |
_abc_ | Exact same thing | 19:07 |
_abc_ | Are these dependencies coming from upstream? | 19:07 |
fsmithred | yes | 19:07 |
fsmithred | well, upstream at least as far as debian | 19:07 |
fsmithred | I don't know beyond that | 19:08 |
_abc_ | reading man apt-get install says there is no direct way to force installation of the selected package only | 19:09 |
_abc_ | Looks like I need to fetch the package and manually install it with deb | 19:09 |
_abc_ | Unbelievable. | 19:09 |
fsmithred | download package and install with dpkg -i | 19:10 |
nemo | _abc_: well you already fetched it | 19:10 |
fsmithred | but if deps are missing, it may not work | 19:10 |
nemo | _abc_: just force install from your cache folder ☺ | 19:10 |
nemo | that's what I did last time I ran into a weird set of deps on debian | 19:10 |
nemo | and maybe back it up in /root or something for future fixing | 19:10 |
fsmithred | _abc_, are you in ascii? I just tried, and I can install those packages without other qemu stuff. Doesn't matter if I exclude recommends - same packages either way. | 19:13 |
_abc_ | I am in ascii yes. But my lists file probably looks different to yours | 19:14 |
fsmithred | oops, I'm using pkgmaster.devuan.org. Should have changed that to deb.devuan.org | 19:15 |
_abc_ | What's the bios package name again? | 19:15 |
_abc_ | For qemu | 19:16 |
fsmithred | oh, gimme a minute | 19:16 |
fsmithred | you mean for uefi? | 19:16 |
fsmithred | ovmf | 19:16 |
_abc_ | No, I don't need uefi yet | 19:17 |
_abc_ | ipxi ? Something | 19:17 |
_abc_ | I don't need that either in fact. | 19:17 |
_abc_ | There's only one qemu-system-arm ? Others have 2, one for 32 one for 64? | 19:18 |
_abc_ | qemu-system-i386 has no manpage | 19:19 |
_abc_ | What the | 19:19 |
fsmithred | I don't know. All I have is what I listed. | 19:20 |
fsmithred | and the ipxe package came automatically with the others | 19:20 |
fsmithred | qemu-system-x86 handles both 32 and 64. The old amd64 package is just a virtual package now. | 19:20 |
fsmithred | Probably same for arm | 19:21 |
fsmithred | the i386 is also a virtual package | 19:21 |
_abc_ | the bios is seabios | 19:22 |
_abc_ | In my case it broke subsequent runs of aptitude/apt-get had to do apt --fix-broken install and it installed all the deps as they were in the unwanted list | 19:23 |
_abc_ | fsmithred: can we discuss why in your case it did not do this mess? | 19:24 |
_abc_ | I have 28 (!) qemu-system-xxx on the machine now. But they work. | 19:25 |
_abc_ | What the blazes is going on? fsmithred do you have an explanation? | 19:25 |
fsmithred | no, I don't | 19:26 |
_abc_ | specifically there is qemu-system-i386 qemu-system-x86_64 | 19:26 |
_abc_ | re: "the same one runs both" | 19:26 |
_abc_ | Looking at startup scripts and stuff a bit | 19:26 |
fsmithred | oh, I can explain why I got no difference with or without recommends, but that doesn't answer the real question | 19:26 |
fsmithred | those are fake packages | 19:26 |
_abc_ | They are not fake, each is a distinct binary. I expected a startup script but no | 19:27 |
_abc_ | The one above is 9MB | 19:27 |
pm215 | qemu-system-i386 is the 32-bit cpu one, qemu-system-x86_64 is 64-bit | 19:27 |
fsmithred | yes, they are virtual packages: distinct binaries that don't contain anything other than a dependency | 19:27 |
_abc_ | 10 | 19:27 |
_abc_ | But is it normal to get 28 binaries installed just to run ONE of them? | 19:27 |
pm215 | qemu-system-foo should be in the qemu-system-foo package, which you should in general not have to install unless you care about architecture foo | 19:28 |
fsmithred | no, it's not normal. | 19:28 |
fsmithred | Package: qemu-system-x86-64 | 19:28 |
fsmithred | State: not a real package | 19:28 |
fsmithred | Provided by: qemu-system-x86 (1:2.8+dfsg-6+deb9u4) | 19:28 |
_abc_ | Maybe I should NOT have installed qemu itself? Only the system package for what I need? | 19:29 |
fsmithred | CORRECT | 19:29 |
_abc_ | There is no standalone binaru called qemu at all anywhere. | 19:29 |
pm215 | x86-64 and i386 are a slightly special case as they share a package | 19:29 |
fsmithred | it's a metapackage | 19:29 |
_abc_ | Ah I am getting it now. Wow. | 19:29 |
pm215 | yes, there's no 'qemu' binary, because each target arch has its own executable | 19:29 |
fsmithred | that means it's just there to pull in a bunch of real packages | 19:29 |
_abc_ | I get it now. | 19:29 |
_abc_ | I think. | 19:29 |
fsmithred | yes, I think so too | 19:29 |
_abc_ | Ok, let's reduce the damage a bit. Roll back install qemu, install just the systems I need. | 19:30 |
fsmithred | remove qemu and just install | 19:30 |
fsmithred | yup | 19:30 |
pm215 | the 'qemu' metapackage is for if you want everything that the upstream gives you for a full-fat configure & build | 19:30 |
_abc_ | apt-get purge quemu did NOT remove the things it pulled in | 19:32 |
fsmithred | right | 19:33 |
fsmithred | apt-get doesn't remove the other packages automaticall. aptitude does. not sure about apt. | 19:33 |
fsmithred | do apt-get autoremove | 19:33 |
pm215 | is this different since stretch? in stretch the qemu pkg Depends on qemu-system which Depends on qemu-system-arm &c, so removing qemu should prompt to remove the other stuff | 19:34 |
KatolaZ | pm215: it would be the other way round... | 19:35 |
KatolaZ | if A Depends on B and you remove A you don't get B removed | 19:35 |
pm215 | doh, yes | 19:35 |
KatolaZ | if you remove B, A gets removed | 19:35 |
KatolaZ | it has always been like that in Debian-based distros | 19:35 |
KatolaZ | :) | 19:36 |
_abc_ | what is vde2? | 19:36 |
pm215 | "apt show vde2" should answer that question... | 19:37 |
_abc_ | I purged the offending packages, the package manager claims they are gone, the files are still there. | 19:38 |
_abc_ | What the blazes | 19:38 |
_abc_ | dlocate qemu-system-misc for example comes up empty, but I can see the files in /usr/bin/qemu-system-cris | 19:39 |
_abc_ | which is one of the system images installed by -misc | 19:39 |
_abc_ | Something is very broken | 19:39 |
pm215 | try dpkg -S qemu-system-cris | 19:39 |
KatolaZ | again, _abc_ | 19:40 |
KatolaZ | qemu-system-misc is a meta-package | 19:40 |
KatolaZ | if you remove it, you won't remove all the packages it depends on... | 19:40 |
KatolaZ | (and, if you just "remove" without "purge", config file will remain there) | 19:41 |
_abc_ | I used purge and the app is there. Probably only removed configs. | 19:41 |
_abc_ | purge alone without remove | 19:42 |
pm215 | also, dlocate uses a cached set of data which is only updated overnight, so won't reflect package nistalls/uninstalls you just did five minutes ago | 19:42 |
_abc_ | I know, I bumped dlocate | 19:42 |
pm215 | use dpkg -S the-binary-you-still-have-installed to find which package it's in | 19:42 |
pm215 | then uninstall that... | 19:43 |
_abc_ | I use dlocate for that. | 19:44 |
_abc_ | That's not the problem. apt-get remove $pkg then apt-get purge $pkg ; second one claims package is no longer installed. | 19:45 |
pm215 | KatolaZ: qemu-system-misc isn't a metapackage, it's the one with qemu-system-cris/alpha/other minor archs in it | 19:45 |
_abc_ | I tried in the reverse order, purge 1st, then the package is not removed for some reason. | 19:45 |
_abc_ | I know what it is. | 19:45 |
_abc_ | I used it as a test for the others | 19:46 |
_abc_ | and the 64 bit arm is aarch64 ok | 19:46 |
pm215 | purge is supposed to be a superset of remove (right the way through down to how dpkg works) so if that is not the way it's behaving then something is very weird | 19:46 |
_abc_ | Right, I won. They are gone. I need to practice much more with debian package tool herding. I'm used to slackware more than anything. I pulled an etch based system using slack packages installed manually for like 10 years. | 19:47 |
_abc_ | pm215: yes, it was very weird | 19:47 |
_abc_ | Also all qemu manpages are installed, even those for the systems which were removed. | 19:49 |
_abc_ | I have started not liking the way things work. Is there a way to create "tagfiles" for packages for oneself, to override the mess coming from upstream? | 19:50 |
KatolaZ | _abc_: what is your problem, exactly? | 19:51 |
_abc_ | I would like more control over dependencies. And clearer labelling of virtual packages. | 20:05 |
Centurion_Dan | infobot+ | 21:00 |
Centurion_Dan | fsmithred: re the ryzen issue... he has to update the bios... many bugs in ryzen early firmware/microcode... | 21:01 |
buZz | Centurion_Dan: cant you update microcode from OS? | 21:04 |
buZz | you can on intel, not on AMD? | 21:04 |
Leander | bios fixes are more than just CPU microcode, there's also improved RAM support and other things that can't be changed after booting | 21:24 |
buZz | yeah totally, i ment the cpu microcode specifically | 21:35 |
Centurion_Dan | buZz: yup, the package is amd64-microcode | 21:37 |
buZz | ah, alright | 21:37 |
Centurion_Dan | but in this case I found when I built a ryzen system with an asrock board that a firmware update was indeed required to avoid that die while sleeping bug.... and the system has been running 9 months with no issues... | 21:39 |
Centurion_Dan | s/die while sleeping bug/die while idle bug/ | 21:41 |
buZz | could have been a regulator or something | 21:41 |
buZz | too low/high voltage for the cpu | 21:41 |
Centurion_Dan | I think from what I read back then it's most likely a too low voltage at min clockspeed or clockspeed transition... | 21:42 |
Centurion_Dan | infobot | 22:00 |
_abc_ | How much disk does one need to build a linux kernel in the 4.9 series ? I've used up like 3GB and it's not ending | 22:12 |
buZz | lets see | 22:13 |
buZz | buzz@h81m:~/code/linux-4.17.2$ make defconfig ; time make -j4 | 22:13 |
buZz | i assume newer is fine for your test? :P | 22:13 |
_abc_ | yes | 22:14 |
buZz | should be ~20 mins | 22:14 |
buZz | i'll post it here | 22:14 |
_abc_ | heh | 22:15 |
_abc_ | make modules is actually more interesting since that's what I did | 22:15 |
_abc_ | I wonder if I can force build just one module. Trying. | 22:15 |
_abc_ | The answer seems to be no | 22:16 |
buZz | you'd need to build most of kernel to be able to build modules , afaik | 22:17 |
_abc_ | you can build just modules. | 22:17 |
buZz | either way, you asked about building kernel | 22:17 |
buZz | so thats the info you get | 22:17 |
buZz | i can do make modules afterwards | 22:17 |
_abc_ | yeah, the total size would be nice to know | 22:17 |
buZz | oh, already done | 22:18 |
buZz | real 4m12.417s | 22:18 |
buZz | 1.3GB now | 22:18 |
buZz | buzz@h81m:~/code/linux-4.17.2$ time make modules; du -sh . | 22:18 |
buZz | after make modules, its still 1.3GB | 22:18 |
_abc_ | that is amazing | 22:19 |
buZz | i used defconfig though | 22:19 |
_abc_ | I used oldconfig with the conf of the running system. | 22:19 |
_abc_ | I have 4GB free after make config then it runs out of space upon make modules | 22:19 |
_abc_ | What the hell. | 22:19 |
_abc_ | linux does not love me today | 22:20 |
fsmithred | I deleted an old kernel source tree today and freed up 8.6GB | 22:20 |
_abc_ | fsmithred: ah. | 22:21 |
_abc_ | That may be it, then. Source uncompressed is not that big. | 22:28 |
_abc_ | I tried to cheat, enable only the module I want in a custom config. make oldconfig breaks on that, it deletes whole sections of the config | 22:29 |
_abc_ | make menuconfig complains about no curses, but I have ncurses-bin installed. | 22:35 |
_abc_ | Ah I need the -dev package too | 22:36 |
_abc_ | I edited .config with regexp in vim, was too entusiastic in removing =m's. Then edited it with make menuconfig (tedious), pared it down to 27 modules, do not expect the kernel to work like this! After that building ONE module builds all of them, and it does so in 10 seconds. | 23:08 |
_abc_ | So success. I can build a single module stand alone without the whole kernel. | 23:09 |
_abc_ | Trying it out. | 23:09 |
_abc_ | Retracting what I said. The module is built but not usable, wrong format. Needs more work. | 23:10 |
_abc_ | Bye for now | 23:10 |
buZz | farewell | 23:19 |
Digit | wrestling my nvidia/X/slim back after they wouldnt start after a powercut. EE no screens, n so on. after it asks if i want to attempt restoration of the original X config, no matter if i say yes or no, it seems to hang. got a progress bar this time using apt rather than apt-get, n it's still 0%. something else i should do? uninstall first maybe? | 23:29 |
Digit | searching for nvidia in htop, and there's a lot of them, seemingly doing nothing | 23:34 |
man_in_shack | force an fsck or 5 | 23:41 |
buZz | which nvidia driver do you run? | 23:44 |
buZz | i've never seen apt getting stuck at 0% | 23:45 |
buZz | usually apt upgrade breaks my nvidia driver here :P so i tend to cherry pick stuff to upgrade to avoid touching nvidia driver | 23:46 |
buZz | then just apt install -t ascii-backports the right nvidia debs | 23:46 |
buZz | wish i understood how to force apt upgrade to not upgrade -backports packages with ascii packages | 23:46 |
Digit | 396.24, according to nvidia-installer --version (if that gets me the right version number). | 23:51 |
Digit | current plan... procrastinate for a bit, then cancl it if still at 0%, then uninstall all the things, then reinstall following the instrutions found online, hoping they're correct. | 23:52 |
gnarface | if apt is hanging at 0% i would expect that to be a network problem | 23:53 |
mtnman | i mean helo | 23:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!