gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
Guanin has joined #nixos-chat
zimbatm has quit [Ping timeout: 252 seconds]
colemickens has quit [Ping timeout: 248 seconds]
Ralith has quit [Ping timeout: 252 seconds]
jasongrossman[m] has quit [Ping timeout: 264 seconds]
popcx has quit [Ping timeout: 246 seconds]
drakonis has quit [Ping timeout: 264 seconds]
pasukon has quit [Ping timeout: 264 seconds]
drakonis has joined #nixos-chat
pasukon has joined #nixos-chat
wirew0rm has quit [*.net *.split]
monsieurp has quit [*.net *.split]
WilliButz has quit [*.net *.split]
Willi_Butz has joined #nixos-chat
popcx has joined #nixos-chat
zimbatm has joined #nixos-chat
<drakonis> https://i.imgur.com/i043EOx.gif what a majestic dog
colemickens has joined #nixos-chat
Ralith has joined #nixos-chat
Guanin has quit [Quit: Leaving]
wirew0rm has joined #nixos-chat
Irenes has quit [Ping timeout: 258 seconds]
Irenes has joined #nixos-chat
drakonis has quit [Quit: WeeChat 2.4]
<ivan> https://youtrack.jetbrains.com/issue/IDEA-212423 please like and subscribe if you see this too
<tilpner> > let __findFile = _: _: "foo"; in <nixpkgs>
<{^_^}> "foo"
<tilpner> Huh, there's probably some way I could use this to make my config less maintainable
monsieurp has joined #nixos-chat
<yorick> EXPORT_SYMBOL_GPL is the worst, IMO
<yorick> what are people gonna do, blackbox reverse engineer zfs in python?
<etu> colemickens: Wow, people go mental about that patch
<etu> I don't think it's any difference at all from before. So what, our 5.x kernel is patched to be a bit more the kernels <5.x and ZFS happens to be a lot faster.
<etu> There's no new violation in any way (I think).
<yorick> it's legally dubious
<manveru> oracle is always legally dubious :|
<eyJhb> WHy does it hurt so bad, being forced to switch to Gitlab :(
<manveru> you don't like gitlab?
<yorick> manveru: oracle has nothing to do with this, this would be a GPL violation
<manveru> doesn't oracle also hold copyrights in the kernel?
<MichaelRaskin> Arguably, it is both CDDL and GPL violation to distribute
<yorick> to distribute ZOL and Linux kernel as a single work, yes
<yorick> but that's not what we're doing here
<yorick> we're changing EXPORT_SYMBOL_GPL to EXPORT_SYMBOL
<MichaelRaskin> It is not clear if compiled ZOL module is already bad to distribute
<yorick> stackoverflow says its fine and ubuntu's doing it
<yorick> you cannot merge ZOL into the linux kernel and compile and distribute that
<yorick> but ZFS is obviously not a derivative work from the linux kernel
endformationage has quit [Ping timeout: 258 seconds]
hedning_ has joined #nixos-chat
<MichaelRaskin> ZFS module _binary_ that uses things not clearly documented as externally supported interfaces might be argued to be a derivative work of ZFS with a mix of derivative work of kernel headers
tilpner has quit [Quit: WeeChat 2.4]
__monty__ has joined #nixos-chat
<eyJhb> manveru: I feel like Gitlab is very "closed" in the way, that they have designed the UI, and very very cluttered with useless sh**.. :/
<eyJhb> And also, it feels dead compared to Github
<MichaelRaskin> I would say a specific annoyance of GitLab is that it requires JS in more places than GitHub
<MichaelRaskin> Of course both systems lack actual issue tracking, but that's not a point of comparison…
<manveru> i really wonder how the github package repo will work out...
tilpner has joined #nixos-chat
<manveru> eyjhb: well, i think by now i'm just used to github ui, i understand that gitlab didn't want to 1:1 copy it or people would be mad about it too, but yeah :|
<manveru> i still plan to publish my new projects via gitlab just because i think we need diversity
<eyJhb> Yeah.. And the only reason for switching, is that Github won't allow my previous annoyance, that I pretty much need to have a organization, which requires private repos and a lot of contributers. But they will be like, one off contributers......... And they want me to pay, 5$, every month, pr. user. Which quickly becomes 20+ users, which will not actively work on the private repos
<eyJhb> So basically, fuck that
<manveru> jup
<eyJhb> Hmm... Currently I just have this - https://github.com/deviosec , but we just finished our event this weekend, and need to put challenges some place
<eyJhb> Some public, some private....
<manveru> plus you can always host gitlab yourself without paying insane amounts of money :)
<eyJhb> Surprised we endeed up having 43 challenges for people to complete! And I think 24 "server based challenges", so each team had a droplet from DO, with 24 containers in :p - Poor thing never saw it coming
<MichaelRaskin> I currently use multiple non-Gitlab.com GitLab instances
<eyJhb> non-gitlab.com ?
<eyJhb> Ohh
<eyJhb> Privately? Or work?
<eyJhb> We use it at AAU too
<eyJhb> (Aalborg University of Denmark)
<MichaelRaskin> Well, once
<MichaelRaskin> I have used research travel funding to present stuff I host on Common-Lisp.net GitLabb I can say both cases are work
<MichaelRaskin> (I also use TU Munich instance)
<eyJhb> Only ever used Github and Bitbucket.. And 3 commits to a Gitlab instance.. But I guess I either have to magically increase my budget from 0,- DKK to 500+,- DKK pr. month. (and want to waste money on that)...
tilpner has quit [Ping timeout: 258 seconds]
tilpner has joined #nixos-chat
<__monty__> Source hut looks like a refreshing take on this. What I don't like about gitlab is how they try to do everything github does but slightly differently. Either be a clone or put some real thought into UX.
<eyJhb> __monty__ `Achtung! Sourcehut is still under heavy development, and while many of the services are usable, expect to find lots of missing polish, broken links, incomplete docs, and so on. Here be dragons!`
<eyJhb> Not really what you want to hear
<__monty__> Didn't claim it was production ready or anything. Just that it's an interesting take on source code hosting : )
<__monty__> I'm not sure about the everything's a mailing list approach but maybe it's not as terrible as it sounds with a proper setup, notmuch or something.
<eyJhb> I really really really hate maling lists for some reason
<__monty__> Probably because you don't have a good email setup : p
<gchristensen> there is a good mail setup?
<__monty__> Just because of the amount of times I end up on some mailing list archive when debugging hard issues I have to admit that maybe I'm missing out on high quality information.
<MichaelRaskin> gchristensen: better than GitHub UI? Absolutely
<eyJhb> Tsk, I just come to gchristensen when I have a hard problem!
<gchristensen> MichaelRaskin: that is true, I can delete emails much faster than I can delete anything in the github UI
ottidmes has joined #nixos-chat
<andi-> Having used github for some time I started to miss mailingslists.. mostly because i have issues with websites..
<__monty__> gchristensen: Better yet, you can have filters do it automatically for you.
<__monty__> eyjhb: I admit that's a valid approach. But that bus factor though!
<MichaelRaskin> _That_ (at least heading/status-based) I can have on GitHub too…
<gchristensen> if email is from: «*» move to: «delete forever»
<MichaelRaskin> If it is only visible via GitHub UI, claim it was never shown
<MichaelRaskin> Right now if I ask for open PRs where I was mentioned, top-10 contains an already-merged PR
tilpner has quit [Quit: WeeChat 2.4]
tilpner has joined #nixos-chat
Willi_Butz is now known as WilliButz
Synthetica has joined #nixos-chat
popcx has quit [Ping timeout: 258 seconds]
<eyJhb> Someone also got annoyed this weekend with my challenge, so he just did "rm -rf /" on the container, because he didn't get any output. Needless to say, he deleted it quite well
popcx has joined #nixos-chat
<MichaelRaskin> Without --no-preserve-root ?
<eyJhb> MichaelRaskin: of course, he didn't care, he just wanted to get a reaction from the server, and he did :p
<eyJhb> The server worked, but the binary behind it got nuked.. Including the flag.txt
<MichaelRaskin> Wait, where did you get rm that deletes / without --no-preserve-root ?
<eyJhb> MichaelRaskin: alpine linux, busybox. Busybox don't care
<pie__> eyjhb, challenge?
<pie__> are we running ctfs on nixos now because thats interesting to me
<pie__> sidenote, given that people are generally ok with this what do i need to do to get it merged https://github.com/NixOS/nixpkgs/pull/61052
<{^_^}> #61052 (by deliciouslytyped, 5 days ago, open): trivial-builders: add writeShellScript and minor cleaning
<lassulus> pie__: add the suggestion from alyssais
<pie__> lassulus, and then i have to do the force push thing right?
<lassulus> yip
<lassulus> amend & force-push
<qyliss> Hi, I'm alyssais, just fyi
<pie__> qyliss, ohai. i kind of guessed but wasnt sure :p
<qyliss> happy to merge once addressed btw. I've wanted this for ages but never got around to it.
<pie__> would have been done last week if i wasnt so clumsy with git
<qyliss> Have you seen https://git-rebase.io
<qyliss> I just came across it the other day. Could have used something like that when I was getting started!
<pie__> ill take a loo
<pie__> k
<sphalerite> teehee, pie__ said loo
<pie__> sphalerite, i was waiting for that :P
<pie__> damn git clone --single-branch --branch patch-2 https://github.com/deliciouslytyped/nixpkgs.git still results in a huge clone
<pie__> not sure that even helped at all
<pie__> on that note. what are we going to do when git history gets impossibly huge
<qyliss> that's far enough off it's not worth worrying about
<sphalerite> I suspect linux will have that problem first :)
<sphalerite> (have and solve)
<qyliss> but you don't necessarily need a copy of the whole history
* samueldr thinks about his kernel looking-arounds and how painful it already is
<samueldr> I hope they look at it :/
<samueldr> maybe though it is exacerbated by how chromeos has a ton of branches due to their release process
<qyliss> Less of an issue for Git as well, since you're sending patches rather than PRs
<qyliss> You don't even necessarily need git
<qyliss> For Linux, I mena
<samueldr> (chromeos makes one branch per LTS version they support, per release, on which they rebase their patches, so each chromeos releases creates 6 branches; add to that 6 previous branches for the WIP tree for that release)
<samueldr> makes for a ton of branches, but also makes it easy to look around
<pie__> ok this page isnt helping
<pie__> how do i squash two commits that already exist on my current branch
<qyliss> git rebase -i
<qyliss> And then the commit or branch you want to base your changes on
<qyliss> So, like, `master', probably
<__monty__> Or just the parent of the oldest commit you want to change if you don't want to rebase on a branch.
<pie__> so...i rebase on to the commit before my changes and "pick" the changes? or am i off by one
<qyliss> Yeah, commit before your changes
<qyliss> And then you can change "pick" to "squash" or "fixup" -- read the comments at the bottom of the file it opens
<pie__> qyliss, can you check it
<pie__> i did the force push
<qyliss> sure
<pie__> Writing objects: 100% (5/5), 779 bytes | 779.00 KiB/s, done.
<pie__> Total 5 (delta 4), reused 0 (delta 0)
<pie__> seems like maybe more than it should have been though
<pie__> so idk if the history is totally clena
<qyliss> nah, that makes sense
<qyliss> objects aren't just commits
<pie__> i know
<pie__> i just dont know offhand how to count them
<qyliss> Don't worry about it
<qyliss> anyway, you did the rebase right :)
<qyliss> there's some trailing whitespace though. would you mind removing that?
<pie__> how do i check for it
<samueldr> with vim: /\s\+$
<pie__> if you mean the stuff between the functions and the next comment, that was on purpose
<qyliss> If you do `git show` it should show you the trailin gwhitespace
<qyliss> It'll be red blocks
<qyliss> By trailing whitespace, I mean spaces at the end of a line
<pie__> oh looks like theres some indentation yeah
<pie__> also i use nano :P
<qyliss> yeah
<pie__> (but i think that also has a mode. anyway.)
<pie__> is there something where i can just edit the diff thing of the last commit and get a new commit
<qyliss> not really
<pie__> wow theres all this random stuff and they dont have that? :P
<qyliss> at least not to my knowledge
<qyliss> there's something similar in git add -p, but that wouldn't help in this case
<qyliss> Git generally doesn't have "random" stuff.
<qyliss> There's always a good reason for it to have thigs, it's just not always relevant to every project.
<pie__> so i have to do a reset where i get the contents of the patch in my tree
<pie__> or i can make the change and squash again and not have to redo the commit message
<qyliss> do the second one
<qyliss> Just clean up the whitespace, at rebase again
<infinisil> I occasionally use `git reset --soft HEAD~` to undo the last commit
<pie__> qyliss, pushed again
<qyliss> You missed one after writeText
<qyliss> :)
<pie__> sheesh :P
<qyliss> You might want to configure your editor to automatically strip trailing whitespace
<qyliss> It's generally frowned upon.
<pie__> at least im getting used to basic rebase usage
<qyliss> Yeah! Practice is good
<qyliss> it gets easier
<pie__> im going to switch to emacs...one of these days
<eyJhb> pie__: no no, just doing very much off-topic :) Ran it on Debian with Docker
<pie__> qyliss, pushed again btw
<qyliss> Looking good!
<pie__> yay
<pie__> lassulus, ^
<qyliss> One thing before I merge
<pie__> oh nvm lassulus unnecessary highlight :D
<qyliss> If you do `git show --pretty=fuller', you'll see that your Author and Commit fields are different
<qyliss> The Commit field doesn't have a valid email
<qyliss> So GitHub won't recognize it as being you
<pie__> qyliss, so what do i put there as opposed to an actual email
<pie__> oh. probably the one in the author field
<qyliss> Well, you can either put an email address, or you can copy the noreply address GitHub gives you
<qyliss> Depends what you want. Real emails can be useful in case people want to contact you about your patches.
<pie__> i dont have an email suitable for actual public use yet
<pie__> or something
<qyliss> In that case use the GitHub noreply one
<pie__> yea i was going to do that
<qyliss> Once you've got that configured, do `git commit --amend --no-edit` and it'll make a new commit that's exactly the same
<qyliss> And you can force push then
<__monty__> Wait that exists? People without a usable email address in this day and age? I thought babies were born with them nowadays?
<qyliss> Since you're only dealing with one commit here, btw, `git commit --amend` is probably easier than git rebase. It just modifies your last commit.
<qyliss> But you get the same result in the end, so up to you :)
<pie__> i have like...several email addresses but i dont want to post them
<pie__> qyliss, n+1-th attempt pushed
<qyliss> Merged!
<qyliss> Thanks pie__!
<pie__> qyliss, how did you notice all this stuff
<qyliss> tools!
<pie__> do you have some super verbose view you use
<pie__> aha
<qyliss> Git shows me trailing whitespace in diffs
<qyliss> I can't remember if it does that by default or not
<qyliss> And GitHub showed me an "unknown user" avatar, so I knew to check the author and committer.
<pie__> aha.
<qyliss> I know to check these things because I've made the same mistakes myself
<qyliss> And then I kick myself later
<pie__> :p
<pie__> ok ive gotta run off for a bit
<qyliss> Thanks for sticking with it through all that :)
<qyliss> hope it's been educational
<pie__> yeah
<pie__> well i gotta learn this stuff eventually because i have bigger changes pending
<pie__> :p
<{^_^}> #60664 (by deliciouslytyped, 1 week ago, open): ghidra: refactor
ajs124 has quit [Quit: Gateway shutdown]
drakonis has joined #nixos-chat
ajs124 has joined #nixos-chat
endformationage has joined #nixos-chat
<manveru> zimbatm: i think we should start adding more projects to builtwithnix... so much awesome stuff around now :)
<sphalerite> manveru: I'm sure zimbatm will appreciate your PRs ;)
<manveru> :D
<manveru> well, i think we also want the projects to link back
<manveru> not sure how to start that process best without annoying everyone
<joepie91> manveru: start a webring!
<joepie91> :P
<pie__> manveru, link?
<drakonis> manveru: make a badge
<drakonis> oh wait, the badge exists and it does link back
<drakonis> or perhaps i'm missing the idea here
<pie__> wow fancy background you have there
<pie__> theres two empty spots in the "trusted by industry leaders" scroller thingy
<manveru> pie__: yeah, somehow nobody responded the last time we asked companies for testimonials :|
<pie__> gotta keep your competitors from finding better tech
<__monty__> Oh, you should get target to admit they use nix : >
<gchristensen> admit?
<pie__> alternatively, if you really think nix has a super high entry threshold tell all your competitors to get nix and watch them go bankrupt
<pie__> kekeke
<pie__> __monty__, that would be...interesting
<__monty__> gchristensen: Yes, because apparently companies aren't eager to disclose they use nix : )
<gchristensen> lorri is sponsored by Target, and so was the graph-based layered docker image work
<gchristensen> (and flakes) I guess the thing missing is the statement?
<drakonis> they sponsored flakes?
<drakonis> ah, require.nix was it?
<gchristensen> require.nix was a prototype, flakes are in development
<__monty__> gchristensen: Nah, just a joke in reference to manveru's observation that companies weren't eager to share testimonials. Combined with Target probably being one of the bigger companies using nix.
<gchristensen> aye :)
<pie__> oh huh
<gchristensen> there are some other big companies that, indeed, aren't eager be associated
<qyliss> huh
<qyliss> that's surprising to me
<gchristensen> Regulations
<qyliss> regulations that don't prevent using nix, but do prevent talking about it?
<__monty__> Required by law to use inferior technology? >: )
<gchristensen> if I understand correctly, any disclosures about their business and operations are tightly regulated
<drakonis> but why?
<qyliss> huh
<gchristensen> although one of these businesses did sponser NixCon2017 ... http://nixcon2017.org/#sponsors
<drakonis> that'd be innoq?
buckley310 has quit [Remote host closed the connection]
<gchristensen> I was thinking Bridgewater
<drakonis> oh
<drakonis> woooow.
<__monty__> gchristensen: The amount of nix-related NDAs you are bound by, is it over 50? : >
<drakonis> or is it OVER NINE THOUSAND
<gchristensen> heh
<samueldr> gchristensen can neither confirm or deny the existence of NDAs ;)
<gchristensen> .....probably...... but most of them are transitive (through Tweag)
<drakonis> isn't that usually a implicit confirmation of NDAS?
<drakonis> NDAs
<samueldr> it's the glomar response
<__monty__> Do we need NDA canaries?
<drakonis> we need canaries for everything
<samueldr> __monty__: do we?
<qyliss> that's very interesting
<samueldr> and how useful are they? I would technically be under NDAs unrelated to anything nixy, but the canary would be... canaried because it's an NDA; and not sure saying "nix related NDAs" has that much more value
<qyliss> my employer is a subsidiary of one of the largest banking groups in the world. I'm always terrified of somebody coming along and imposing that sort of requirement on us.
<drakonis> gchristensen: you've used the nix flakes tree, is it in any state to replace using vanilla nix?
<qyliss> they generally try to keep us as far away as possible from that sort of thing though AAUI
<gchristensen> drakonis: mmmm it is probably pretty early
<gchristensen> drakonis: but maybe a couple more weeks? it definitely should get a good kick of the tires in not too long
<drakonis> ah, nice.
<gchristensen> I've used flakes early because it makes it feasible to use ofborg for projects which aren't nixpkgs
<gchristensen> so, I'm a bit excited :)
<drakonis> aye
<drakonis> if i understand this correctly, i can also use flakes to pull from language registres?
<gchristensen> I'm not sure I understand
<qyliss> Is {^_^} broken?
<qyliss> It's not responding to my PMs
<gchristensen> ,help
<{^_^}> Use `,` to list all commands, `,foo = Foo!` to define foo as "Foo!", `,foo =` to undefine it, `,foo` to output "Foo!", `,foo somebody` to send "Foo!" to the nick somebody
<qyliss> ,locate bin ruby
<{^_^}> Found in packages: ruby, jruby, ruby_2_3, ruby_2_4, ruby_2_6, logstash5, logstash6-oss
<qyliss> ...
<qyliss> wonder why it isn't replying to me
<{^_^}> tweag/nix#6 (by edolstra, 12 weeks ago, open): Extensible flake references
<gchristensen> qyliss: I can't PM it either
<drakonis> like this.
<drakonis> i can fetch dependencies from a proglang package registry
<drakonis> i think i'm missing the point of that issue.
<drakonis> what that issue proposes is a way to replace fetchers, isn't it? it is still necessary to write package definitions to install them?
Synthetica has quit [Quit: Connection closed for inactivity]
<pie__> qyliss, whats aaui
<qyliss> As I understand it, except I got it wrong
<qyliss> lol
<pie__> infinisil, people cant pm the bot (see above)
<pie__> qyliss, ah
<infinisil> Whoops
<infinisil> Should work again in a couple minutes
<pie__> :D
<pie__> "<infinisil> oh no. OH NOOOO"
<infinisil> It *might* be time to add some tests to my nixbot lol
<infinisil> (Although, what happened now couldn't have been prevented by tests)
<qyliss> what happened?
<infinisil> Yesterday I implemented being able to specify plugins for channels in the config file, all plugins being disabled by default
<infinisil> I turned on the plugins for channels again, but I forgot for PMs
<infinisil> So I rebuilt it with PMs enabled
<infinisil> But in parallel, I also made a change to the unregistered plugin so it wouldn't be channel-conditional anymore
<qyliss> ahh
<infinisil> But because I turned on the plugins for channels, all of them, the unregistered plugin was in there too
<infinisil> But now it wasn't conditional on nixos-unregistered anymore
<infinisil> > 1
<{^_^}> 1
<pie__> so which one of you is the one working for bridgewater
<infinisil> Okay now everything should be good again
<__monty__> pie__: I can't say because of NDAs.
__monty__ has quit [Quit: leaving]
arianvp has joined #nixos-chat
ottidmes has quit [Quit: WeeChat 2.4]
Guanin has joined #nixos-chat
pie___ has joined #nixos-chat
<drakonis> this is beautiful.
<drakonis> so for stadia games to exist, they need to be ported to a linux stack
<samueldr> though, this doesn't mean that there will be more games for Linux; it's only used as a means for the service to exist
<drakonis> they could potentially release more linux versions
<drakonis> because they have to develop the game on linux before stadia comes in
<drakonis> or just develop and not release at all
<drakonis> who knows