gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
<aanderse> yawn
<aanderse> hey y'all
pie___ has quit [Ping timeout: 252 seconds]
Synthetica has quit [Quit: Connection closed for inactivity]
<buckley310> hello
drakonis has joined #nixos-chat
<drakonis> who was it that didn't like steam workshop?
<drakonis> https://mod.io/
<drakonis> this seems like a cool replacement for it today.
<Ralith> I don't trust something that appears to describe a CDN as a "mod API"
<Ralith> that is not what those words mean
jasongrossman has quit [Read error: Connection reset by peer]
endformationage has quit [Quit: WeeChat 2.4]
Myhlamaeus has quit [Ping timeout: 248 seconds]
jasongrossman has joined #nixos-chat
drakonis has quit [Read error: Connection reset by peer]
<manveru> what's wrong with that site... it doesn't let me scroll
Jackneill has quit [Remote host closed the connection]
Jackneill has joined #nixos-chat
n_db has quit [Remote host closed the connection]
<joepie91> huh. 0ad uses it?
<joepie91> anyway, aside from the horrible site, that looks good
<joepie91> I'm missing the business model though
Synthetica has joined #nixos-chat
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.2)]
jasongrossman has joined #nixos-chat
pie___ has joined #nixos-chat
srk has quit [Ping timeout: 258 seconds]
srk has joined #nixos-chat
Twey has joined #nixos-chat
<pie___> fetchgit seems to manually set some ssl certificate variable, but cacert has a setup hook, should the latter be relied on instead? https://github.com/NixOS/nixpkgs/issues/3382#issuecomment-490475790
endformationage has joined #nixos-chat
<samueldr> 7 hours from now?
<gchristensen> how did you see the future? :o
<samueldr> the future already happened 18 minutes ago
<sphalerite> guess: timezone fails or pre-dated commits
<samueldr> no commits there, so some kind of internal issue I guess, a comment + close event
<ar> /44
<gchristensen> ar: I see you often sending " /xx" -- is that switching windows or something?
<ar> gchristensen: yup
<gchristensen> right on
<Synthetica> samueldr: I'm also seeing that, looks like github is bugged somehow
<gchristensen> I'm seeing other reports from "around" that show github time traveling
<ar> … did they move github backend to azure?
<gchristensen> wow that is quite the azure superpower
<Synthetica> Maybe they moved the servers to the planet from interstellar
<ar> system time drifting on multi-vcpu VMs is a solved problem just about anywhere except for hyperv and azure
<ar> (from my experience)
<pie___> so....im trying to use go2nix save on a repo acquired via fetchgit
<pie___> and i need the .git to exist
<pie___> but fetchgit seems to return a weird looking .git?
<pie___> or rather, does fetchgit do something that will make the git repo all weird
<pie___> i set leaveDotGit = true, and git status looks weird
<gchristensen> amazing
<pie___> this is for a freshly fetchgit-ed repo https://bpaste.net/show/8f7086e2ca41
<gchristensen> https://screenshotscdn.firefoxusercontent.com/images/c96490ce-d63b-4283-a0d5-af5b6261bb3d.png -- the bottom review was left a few hours *ago*
<Synthetica> I think they messed up the timezones
<gchristensen> yeah
<gchristensen> this is going to be nasty
<samueldr> heh, good to know what I'll be doing https://stuff.samueldr.com/screenshots/2019/05/20190508105437.png
endformationage has quit [Quit: WeeChat 2.4]
<joepie91> Lidl's customer service phone number has a very... Spongebob-y tune as their hold music
<pie___> we really should get some kind of warning when rebuilding something that depends on a changed but not reevaluated fixed output derivation definition...
<pie___> s/not reevaluated/not succeeded/
<Taneb> pie___: I'm not sure I understand what circumstance you mean
<pie___> Taneb, foregetting to change the output sha256 when you change the definition
<Taneb> Oh, yeah, I've ran into that before
<pie___> oh crap....go2nix uses nix-prefetch-git which uses nix-store
<pie___> recursive nix problems
<Synthetica> pie___: Can't fixed-output derivations automatically be re-checked if the input changes?
<pie___> Synthetica, i'unno
<pie___> i think the whole point of a fixed output derivation is to avoid the rebuild
<pie___> (and to have a well defined input to its dependents)
<pie___> looks like it only uses it as a roundabout call to nix-hash though? https://github.com/kamilchm/go2nix/blob/52270bd92cd689729097fb4326f5b53eea431cd6/hash.go#L16
<Synthetica> pie___: But how often does the input "magically" change? Honestly, the only case I can think of is the case where it would be useful to be sure
<pie___> Synthetica, idk, my head is somewhere else, but this is a problem im annoyed by constantly
<pie___> maybe there's the case where you have *a* result but dont want to rebuild again because its nondeterministic
<eyJhb> Is it just me, or when my system starts using swap it is absolutely unuseable? 16gb of ram, 8 gb swap, samsung something something pro ssd ...
<cransom> even if swap is on ssd, swap is still amazingly slow if you are relying on it.
waleee has joined #nixos-chat
<samueldr> it also depends on whether it's evicting things it doesn't need, or if the memory is not enough to cover for everything going on
<eyJhb> Well. I guess it isn't fair either, that the ram from 2GB to 16GB got used in under a minute, and kept growing
<cransom> yeah. if the working set got purged to swap, thats the problem.
<samueldr> swap can't substitute for ram when needed, but help cushion extremes when _whatever_ working thing(s) fits in ram
__monty__ has joined #nixos-chat
Myhlamaeus has joined #nixos-chat
<pie___> joepie91, as if the situation with nix-shell wasnt confusing enough, it would be nice if we could have a flag thats basically callpackage with <nixpkgs>
<pie___> so i didnt have to type out -E ...
<pie___> or is that getting too kitchen sink
<pie___> ah crap...how do I get my system overlays to show up with -E?
<pie___> ...my self from 5 days ago answered the question in the default.nix
pie___ has quit [Remote host closed the connection]
pie__ has joined #nixos-chat
pie__ has quit [Ping timeout: 252 seconds]
<andi-> working on systemd242... systemd now expects commands to be ran after installation on the target FS modifying the setup of the /etc/systemd folder... *tableflip* https://github.com/systemd/systemd/blob/master/NEWS#L252-L260
<samueldr> I might be missing something, but this seems to be an improvement? unless we were relying on those files being symlinked at install time in the systemd nix store path?
<andi-> I think it is overall a good improvement just have to adopt our stuff to create the now missing symlinks and figure out how to do that with changing (upstream) units. The amount of files might change, the dependencies etc..
<andi-> It is already a bit of work to port our patches and test it. It just gets worse very systemd release. We should try figure out a way to upstream some things but nobody has invest that amount of time yet.
<ar> it looks like something that probably *every* distribution will have to patch out
pie__ has joined #nixos-chat
<tilpner> sphalerite: Where'd you get that time machine? https://tx0.co/mco
<sphalerite> tilpner: yeah, github is bugging up
<tilpner> :/
<sphalerite> been happening all over the place all day
<joepie91> lol
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos-chat
ivan has quit [Ping timeout: 250 seconds]
<sphalerite> So I bought a PlayStation (1) game, planning to rip it and play it using pcsxr… And I can't seem to rip it in a way that pcsxr likes. >_<
lassulus has quit [Ping timeout: 250 seconds]
LnL has quit [Ping timeout: 250 seconds]
<joepie91> sphalerite: aren't downloadable ISOs pretty easy to come by now, for PS1 games?
cransom has quit [Ping timeout: 250 seconds]
avn has quit [Ping timeout: 250 seconds]
ivan has joined #nixos-chat
<sphalerite> joepie91: probably? I was hoping _not_ to need to download it probably-illegally
avn has joined #nixos-chat
<simpson> sphalerite: Are you using ripping tools which understand subchannel metadata? PS1 discs have screwy subchannel data that has to be preserved.
<sphalerite> yeah, I used cdrdao as described in https://pandorawiki.org/Creating_images_of_PSX_games_using_Linux
<sphalerite> would trying a different CD drive make any sense?
<simpson> And your hardware supports it too?
<simpson> If your emulator supports playing directly from disc, you could try playing from disc and seeing whether your hardware is to blame. I don't know how common it is for hardware to not be able to read PS1 discs though.
<sphalerite> I don't have a CD drive in my laptop
<joepie91> sphalerite: ripping the disc is also illegal
<joepie91> so if you were going for legality... :P
<__monty__> How is it illegal
<joepie91> you're making a reproduction
<__monty__> You can't even violate copyright if there's no redistribution.
<joepie91> (some countries have very specific exemptioins for this, but they usually don't apply for software)
<joepie91> yes, you can
<gchristensen> license violation
<joepie91> this is one of the absolutely ass-backwards parts of copyright law
<sphalerite> ah, what a bummer
<sphalerite> but playing from the CD directly is ok?
<joepie91> yes
<joepie91> so long as you're not breaking copy protection
<joepie91> or 'non-trivial copy protection', depending on country
<joepie91> because that is also illegal :)
<joepie91> (the exact text is usually something like 'circumventing technical measures that prevent the blah blah blah' but basically it's about breaking DRM)
<samueldr> ask your lawyer whether you can or not play your playstation game
<gchristensen> "no"
<joepie91> I mean, my personal opinion is that all of this is BS and you should basically disregard legality and just do your thing
<joepie91> because the laws make no fucking sense whatsoever
ottidmes has joined #nixos-chat
<ottidmes> how relevant is the difference between 700:1 and 1400:1 contrast on a IPS display?
cransom has joined #nixos-chat
Guest99428 has joined #nixos-chat
lassulus has joined #nixos-chat
waleee has quit [Quit: WeeChat 2.4]
<colemickens> LOL, copy/paste works after a VT switch for me, AND fixes alt keys in Firefox.
<colemickens> It's stupid, but I'm pretty happy about this. :D
<infinisil> Software works in mysterious ways..
__monty__ has quit [Quit: leaving]
<andi-> the moment freenode tells you "you are in too many channels" -,-
<clever> ive got nearly 100 windows open in irssi, yet i somehow havent hit that in ages
<andi-> 120 channels in freenode alone
hedning_ has quit [Remote host closed the connection]
<andi-> uargh, touching systemd now also rebuilds the kernel every time.. slow feedback cycle :/
<infinisil> I periodically clean out channels that aren't worth being in anymore
<gchristensen> andi-: :o
<samueldr> "CHANLIMIT=#:120" part of the initial replies the ircd send on connection :)
<andi-> samueldr: yeah.. I easily forget about that since most of the other networks I am on I am either an IrcOp with out limits or below 10 channels :/
<samueldr> "why #:120 and not just 120?" because historically IRCDs could have different kinds of channels, public and private, with different namespaces
<andi-> "local" :D
<andi-> It is a shame much of that stuff is irrelevant/forgotten these days..
<samueldr> in a past life of mine, I wanted to make an irc torture suite :3
<samueldr> ircd that tests compliance and how non-compliant clients are, and conversely, client to test servers
<andi-> cccccckfckdgdrribnrhrfleicfehntnbkkcdtvbgflc
<andi-> arghs, stupid yubikey
<andi-> While I wait for the systemd, gtk, mesa, kernel, ... rebuilds: Is there any good TV series worth watching?
<samueldr> I've been watching babylon 5 lately, it's sad they lost the CGI sources
<andi-> ouch
<samueldr> it was filmed in 16:9 even though TV at the time was always filmed 4:3
<samueldr> so the 16:9 dvd release pan and scan the 4:3 parts with CGI, it shows
<samueldr> but otherwise, while being the kind of sci-fi series from its time, I find it really engrossing and with a good set of stories + overarching plots, hope the end holds up
<andi-> I watched it when it aired on television. Might be worth rewatching.
<samueldr> to add some details; if the CGI files were available, they would have been technically able to remaster to an HD 1080p resolution without issue, due to the film used
<lopsided98> I just recently watched Battlestar Galactica, and its probably one of my favority TV shows now
<andi-> lesson learned: backups!
<samueldr> (and while it would have riled up fans, they could have tweaked the early CGI a bit to make the camera and viewpoints more organic in those space shots :))
<samueldr> well, someone apparently has the files, http://www.b5scrolls.com/
<infinisil> Anime anybody? :D
<samueldr> (or they remade it, didn't dig too much in that since I don't want to be spoilered too much)
<ottidmes> infinisil: any tips?
<infinisil> ottidmes: Haibane Renmei is my recent favorite :D
<ottidmes> infinisil: thanks! I did not know that one yet :)
<elvishjerricco> Is there a sorting algorithm that can just ask a real human to do each comparison in their head (potentially letting them say "i don't care"), and deal the imperfections in the ordering and maybe even split the groups of "i don't care"s into tiers?
<elvishjerricco> The application I currently have in mind is just to sort the MCU movies into tiers of my order of preference :P But I imagine it'd be useful in other contexts.
<infinisil> elvishjerricco: I did think about something similar for a while regarding rating songs
<elvishjerricco> that'd be another good application
<infinisil> The problem being that it's hard to rate songs on an absolute scale (how good is it from 1-10?), but it's much easier to tell which song of two you like better
<elvishjerricco> yea
<infinisil> So I'd like to have an application play songs in a row, and at each transition I could say "Oh this new song is better/worse than the last one"
<infinisil> And then it could sort all sorts as to how good they are
<elvishjerricco> Plus I don't trust my ability to consistently rate movies / songs the same way. I.e. I might accidentally give one movie a higher score than another, even if I'd say I favor the other between the two of them
<infinisil> Yes exactly!
<infinisil> The current rating schemes everywhere are really flawed
<infinisil> imo
<elvishjerricco> of course a comparison based sort could have similar accidents; i.e. I might say `a > b`, `b > c`, and `c > a`
<elvishjerricco> but I'd hope I'm smart enough to just say each of those comparisons are equal, not greater than
<infinisil> elvishjerricco: Another slight problem is that movies/songs might change in how good they are over their span. So it might be great at the beginning but bad towards the end. So the rating depends on when you rate it or whether you take the full length in consideration
<infinisil> elvishjerricco: I figured it would have to be a probabilistic comparison. So when you say a > b, it's really "a probably > b, but maybe not"
<elvishjerricco> that's a good way to think of it
<infinisil> It would be really cool to study such a rating method and come up with something very practical
<elvishjerricco> a plain ole stable sort might be close enough with an impure comparison operator
<elvishjerricco> obviously not very rigorous, but probably somewhat accurate, if you make sure to use EQ for things you're not sure on
<infinisil> I'm not sure how well that would work
<infinisil> Regarding the probabilistic approach, it would make sense to have a rating from -5 to 5, -5 means "I prefer choice 1 by far" and 5 means "I prefer choice 2 by far"
<infinisil> Instead of just LT/EQ/GT
<infinisil> Can get more info out of the user this way and it makes more sense too imo
<elvishjerricco> yea that'd be good
<elvishjerricco> dunno of any probabilistic sorting algorithms though
<elvishjerricco> Here's a 26 page paper on the subject, off the top of google :P http://www2.math.uu.se/~svante/papers/sj153_QSerror.pdf
<infinisil> Neat
<jasongrossman> ottidmes: From what I gather from review sites, you should ignore contrast ratios. I don't remember why - maybe vendors lie about them?
<ottidmes> jasongrossman: yeah I read that about dynamic contrast ratios, when they say things like 30000:1 which is pure marketing, but I was not sure about the non-dynamic (static?) ones. I just bought the laptop with the higher contrast display (not for that reason though, chose it a 14" over a 15.6" due to the lower weight / keyboard layout)
<elvishjerricco> infinisil: just using `import Data.List (sort)` with an `unsafePerformIO`-based comparison function, LT/EQ/GT actually worked reasonably well for the MCU movies for me.
<infinisil> elvishjerricco: Ew, unsafePerformIO!
<infinisil> But neat :)
<elvishjerricco> I didn't wanna have to go copy the `sortBy` source :P
<infinisil> There must be a monadic sort out there somewhere
<elvishjerricco> Possibly but this was easier :P
<jasongrossman> ottidmes: Well, actual contrast is important of course, so probably you did the right thing in terms of contrast, plus your other reasons.
<jasongrossman> ottidmes: Personally I like small laptops, not for the low weight but for the lack of bulk.
<ottidmes> jasongrossman: it was hard to decide which one to take, because the 15.6 was only 3cm wider (compared to the 14"), not that much bulk, but I loved the size of my old 13.3" and this new 14" is only 1cm wider, so size wise it was probably also for the best, but hard to tell those things without actually using them for a few days in practice
<gchristensen> I feel GitHub is lying to me about the status of my svn import
<gchristensen> it has been "Optimizing repository and pushing commits to GitHub..." for about an hour
<simpson> How gnarled was your svn history? It can definitely take a long time sometimes.
<gchristensen> no idea, it is virtualbox
<infinisil> Can confirm, it can take a very long time to convert svn to git..
ottidmes has quit [Ping timeout: 244 seconds]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.2)]
jasongrossman has joined #nixos-chat