gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
srk has quit [Quit: ZNC - http://znc.in]
srk has joined #nixos-chat
hedning has quit [Quit: hedning]
pie__ has joined #nixos-chat
pie_ has quit [Ping timeout: 240 seconds]
endformationage has quit [Quit: WeeChat 2.3]
drakonis has quit [Quit: WeeChat 2.3]
tom39291 has joined #nixos-chat
<etu> Oh wow, this is a cool website: https://explainshell.com/
<etu> (Someone was on "can I use" in bash or something here before, this is not that but still cool)
<joepie91> etu: not quite caniuse, but http://tldp.org/LDP/Bash-Beginners-Guide/html/x7369.html may be useful to them nevertheless
<manveru> Is there really not a single talk at FOSDEM about nix?
hedning has joined #nixos-chat
__monty__ has joined #nixos-chat
MichaelRaskin has joined #nixos-chat
iqubic has quit [Ping timeout: 268 seconds]
Guanin has joined #nixos-chat
ninjin has quit [Remote host closed the connection]
ninjin has joined #nixos-chat
tilpner has quit [Quit: WeeChat 2.3]
tilpner has joined #nixos-chat
tilpner has quit [Quit: WeeChat 2.3]
tilpner has joined #nixos-chat
endformationage has joined #nixos-chat
<MichaelRaskin> A few days of rebuilds (shrinking the scope of the rebuilds with each step…), and I now know that my VGA output problems after update were because of Xorg and not the kernel!
<MichaelRaskin> (just filed an issue about the findings)
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos-chat
aszlig has quit [Quit: Kerneling down for reboot NOW.]
aszlig has joined #nixos-chat
drakonis has joined #nixos-chat
<gchristensen> nice!
<MichaelRaskin> I guesss I need to GC now that locating the problem is over…
<joepie91> MichaelRaskin: 532165 store paths removed, 99284MB freed
<joepie91> :P
<MichaelRaskin> Actually, quite a likely outcome
<MichaelRaskin> Let's see…
<MichaelRaskin> I am considering opening a third issue, but maybe as a completely disinterested bystander I shouldn't…
<MichaelRaskin> (The issue would be about NixOS tests with /boot full)
<gchristensen> hrm neat
<MichaelRaskin> (But I don't use NixOS mainline or NixOS boot menu generatation code)
<samueldr> MichaelRaskin: care to share the high level things you had in mind if you don't follow through?
<MichaelRaskin> Well, once in a while we have complaints «ouch my boot is full what do how help help help»
<samueldr> yeah
<MichaelRaskin> (which are reasonable complaints because our approach does lead to more /boot use)
<MichaelRaskin> There are two questions: 1) if you GC on a single-OS machine you should be enable some /boot-cleaning magic
<MichaelRaskin> So that everything succeeds afterwards
<MichaelRaskin> 2) It should be hard to get to unbootable state just via letting /boot become too full
<MichaelRaskin> 48444 store paths deleted, 102992.09 MiB freed
<MichaelRaskin> joepie91: exceeding expectations!
<joepie91> MichaelRaskin: ha! I was surprisingly not far off with my size estimation..
<MichaelRaskin> OK, now resizing the store back from 210GiB to 100GiB…
<MichaelRaskin> (helps to prevent «I should have remembered to GC a week ago, before that huge build»)
iqubic has joined #nixos-chat
<iqubic> Is it possible to play Factorio on a laptop with a trackpad?
<iqubic> Or will that just bring hell upon me?
<joepie91> welp, looks like I'm done with Rust for the day
<joepie91> tooling's broken
<gchristensen> oops
<{^_^}> rust-lang/rls-vscode#500 (by joepie91, 7 minutes ago, open): Working on `no_std` crates with a custom target specification
<{^_^}> rust-lang/rls-vscode#501 (by joepie91, 4 minutes ago, open): No way to specify custom target specifications relative to project
<{^_^}> rust-lang/rls-vscode#448 (by joepie91, 12 weeks ago, open): RLS runs with the home directory as the CWD, breaking RLS wrappers
<joepie91> iqubic: yes.
<joepie91> :P
<joepie91> iqubic: more seriously, yes, it's possible, but like with anything that requires intensive mouse usage with precision, a trackpad isn't going to be great for it...
<MichaelRaskin> samueldr: basically, the two test cases are all I thought about re: /boot, if you wish I can write the issue and mention you
<gchristensen> rewriting my r13y script in to a rust program it its already 2x as long and w/ 1/2 the functionality
<MichaelRaskin> What was the original language?
<gchristensen> bash
<MichaelRaskin> ahahaha
<gchristensen> =)
<MichaelRaskin> You are talking about the task which is close to bash's reason to exist
<MichaelRaskin> And typical Rust code is low-level
<gchristensen> I know :(
<gchristensen> it was a good pick for the first revision!
<MichaelRaskin> Bash is just a good pick for that task
<MichaelRaskin> A Julia or Lisp rewrite in a terse style might not be bloated. But Rust… at least you won't have buffer overflows
<MichaelRaskin> And I guess you don't have the type of parallelism where data race safety would pay off
<gchristensen> I'd leave it in bash, but I'm adding this whole thing around submitting results to a remote server
* joepie91 mumble mumble javascript mumble
<MichaelRaskin> curl
<gchristensen> samueldr: MichaelRaskin is being a bad influence
<joepie91> lol
<qyliss> I've never found a comfy language for this kind of thing
<samueldr> you think so? because I thought those things before knowing of nix and him :)
<qyliss> I wish I could do things as nicely as in bash, but without bash's weirdness
<qyliss> I should learn Perl I suppose :P
<MichaelRaskin> NOPE
<MichaelRaskin> Perl is optimised for unreadability
<MichaelRaskin> Bash is readable. J is readable. Perl is not.
<qyliss> I made a two line change to a Perl script in nixpkgs yesterday and it took me aaaaaages
<samueldr> you might be forced to when developing in some NixOS portions :/
<samueldr> (or hydra)
<joepie91> actually, gchristensen, have you considered writing it in Nix :D
<gchristensen> anyway, I do actually have a thread pool
<MichaelRaskin> qyliss: if you want a personal comfy environment for writing stuff that should be write in bash, but in a faster language, there is a path there.
<gchristensen> MichaelRaskin: do go on
<MichaelRaskin> But probably no chance of a one-size-fits-all ready-made one
<MichaelRaskin> Step one. Learn _basic_ Julia stuff.
<qyliss> gchristensen: what do you mean?
<MichaelRaskin> Step two. Learn Julia macros and AST.
<MichaelRaskin> Step three. Take a bash-oriented task and write dream code that doesn't feel as completely not Julia after step one.
<MichaelRaskin> Step four. Implement the macros that would allow that dream code to actually run…
<gchristensen> ah I see
<MichaelRaskin> We-ell…
<gchristensen> basically, make the bash styleI want to see in the world
<MichaelRaskin> Approximately
<gchristensen> and then make the world to support it
<MichaelRaskin> Maybe we should start a collaborative collection of dream code — if multiple people converge on a compatible set of styles (maybe different, bt obviously feasible in a single consistent logic) we could even write an implementation of a transpiler to something, without the limitations of being a well-behaved macro in language X
<gchristensen> I do regularly feel grumpy about how bash is great at stuff and then you make your script 5 lines too long and it turns in to regret
<samueldr> gchristensen: use "." to hide the regret away :)
<MichaelRaskin> Dunno, 3000-line bash program I wrote had only the problem of performance
<gchristensen> lol
<MichaelRaskin> Well, munching thousands of data entries in a way that calls for bash loops does indeed take time
iqubic has quit [Ping timeout: 264 seconds]
<MichaelRaskin> For some of bash-like processing there is this problem that making structure processing right is that 20% of work that takes the second 80% of time
<MichaelRaskin> Especially together with preexisting command-line tools
<cransom> the performance of your bash also greatly depends on how bashy your script really is. if you shell out with greps and stuff and not use it's own parameter substitution, etc, it's real bad, real fast.
<MichaelRaskin> Julia is a language that tries to pick up everything from language development history that they have a remote hope of trying to sell to mainstream
<MichaelRaskin> That title looks like «OK what else from Common Lisp haven't we tried picking up yet?»
<gchristensen> lol
<MichaelRaskin> They often do a good job there, too
<{^_^}> JuliaLang/julia#16927 (by tpapp, 2 years ago, open): Added noteworthy differences from Common Lisp.
<MichaelRaskin> Think of the fact that adding such a document _makes sense_ in a modern language
endformationage has quit [Ping timeout: 272 seconds]
<joepie91> complaint: no code sample on the homepage
<joepie91> :p
<MichaelRaskin> Equally applicable to Julia and Common Lisp! Actually, Nix, too
<gchristensen> I do tend to gravitate to projects which suck at that :P
<samueldr> no code sample here either
<samueldr> proof!
<MichaelRaskin> Even Rust
<MichaelRaskin> Maybe it is because some languages are defined by goals, not by aesthetics?
<gchristensen> :D
<joepie91> MichaelRaskin: sure, but would you buy a hammer in a plain box without any imagery?
<joepie91> probably not, you'd want to know its shape, to estimate its ergonomics, etc.
<MichaelRaskin> I have some questions about places that sell hammers in sealed boxes
<gchristensen> haha
iqubic has joined #nixos-chat
<joepie91> MichaelRaskin: exactly!
<joepie91> thus my complaint :)
<MichaelRaskin> joepie91: a natural way to obtain a hammer is as a part of a plain unmarked box, which you need to approach and unfasten and open
<MichaelRaskin> (it contains more than a single hammer usually)
<iqubic> This sounds like a metaphor for something.
<MichaelRaskin> No, a literal description of a Jonnesway tool case
<joepie91> MichaelRaskin: I was talking about the purchasing process though, not obtaining it from your toolbox
<joepie91> :P
<MichaelRaskin> joepie91: purchasing process includes going to store, using the category labels, finding the approximately correct set of boxes and asking to show the contents (or opening them yourself, depends on the store details)
<joepie91> ???
<joepie91> they sell hammers in boxes there?
<MichaelRaskin> As I said, if hammer has a box in the first place, it is a part of a set
ar1a has left #nixos-chat ["WeeChat 2.3"]
jackdk has joined #nixos-chat
<gchristensen> what a discussion :')
<joepie91> lol
endformationage has joined #nixos-chat
<samueldr> I hate technology
<samueldr> I hate how disposable it is
<samueldr> well, how it is made to be disposable
<samueldr> realised a laptop, not really old, 4 years, has a battery issue, its battery is now blown up (bigger, not exploded)
<samueldr> and the dang thing needs to be disassembled to change
<samueldr> and obviously it's a batteyr model specific to some models for that brand, and impossible to get
<samueldr> who wants to bet on whether the laptop works with the battery removed?
<andi-> ha technology.. Yesterday I booked a car on the local car sharing platform.. It tried to infer my timezone from my language settings.. great clusterfuck.. Ended up booking things a few hours in the wrong direction…
<andi-> samueldr: yeah, that stuff needs to be fixed... Can't have non-replaceable (non-servicable) hardware... It is a waste of resources
<samueldr> andi-: do you feel I got printing times right in the hydra patch I made? look anywhere there is either an elapsed time or a date https://hydra.nixos.org/jobset/nixos/release-18.09
<samueldr> also, the issue is three-pronged imho: standardize consumables (e.g. batteries) give access to parts to the end-users, and make it user-replaceable for consumables, and even some non-consumables
<andi-> samueldr: It shows the wrong local time for me.. Off by one hour
<andi-> but I blame that on firefox right now
<samueldr> you have js turned on, right?
<andi-> (different issue then yesterday)
<andi-> I have it in a very restricted namespace, without access to timezone data
<samueldr> ah
<samueldr> it should print, in a tooltip, the reasoning behind the time it prints
<andi-> running it in chrome (without additional sandboxing) fixes it
<andi-> Yeah, saw it
<joepie91> samueldr: caution, don't continue using anything with a blown-up battery
<joepie91> nor try to charge it
<samueldr> joepie91: guess why I'm removing the battery
<joepie91> okay, just wanted to make sure it wasn't an "I'll get around to it eventually..." thing :)
<joepie91> I've seen a few too many people try that lol
<samueldr> a peeve about that: WHY is this allowed to happen without the device going all "no, not booting" :/
<joepie91> because the battery still provides power :P
<samueldr> I only noticed because the laptop's screen felt curved
<joepie91> I'm not sure that it's electrically detectable
<samueldr> until I realised it was the palm rest
<joepie91> oof
<jackdk> My thinkpad is physically falling to pieces and while it's electrically decent I'm saving up to replace it with something else... shit sucks
<samueldr> putting the battery deep into the internals is a recipe for disaster
<samueldr> especially given how much energy there is in these pouches
<joepie91> jackdk: I find it interesting how it's almost always thinkpads that physically fall apart (sometimes outright have holes in the casing!) but still technically work
<joepie91> even the newer ones
<joepie91> (which use much weaker case plastic)
<joepie91> most other laptops I see mechanical components (like fans or HDDs) failing, or electrical components, while the case still looks fine...
<samueldr> holy crap, I don't believe it it boots without battery
* samueldr hates phone mfgs for that reason
<__monty__> My 2008 T400 is still going strong : )
<samueldr> oh lol, the top case integrity of that laptop was partially due to the battery :/
<samueldr> now that there's a void it's... mushy
<andi-> I am currently in the process of rescuing a bunch of x220 and x230 from being demolished.. Will have them as test and backup hardware, They are still great for terminals and text editing..
<__monty__> The loss of modular storage is so sad with modern phones and tablets. (as compared to desktops and laptops, not older phones necessarily)
<joepie91> samueldr: gee that doesn't sound like it might be the cause of the battery failure at all <.<
<joepie91> >.>
<samueldr> I can't see anything going wrong with that design
<samueldr> though I've been blinded by a random fire on my laptop
<samueldr> brb
<samueldr> (not really)
<samueldr> (joking about seeing)
<samueldr> I even think one of the screws holding the parts together (thankfully no glue) was forced out of its threads, as I had one fewer screw than should have
<samueldr> so the laptop literally had a screw loose
<__monty__> Sometimes that's just "optimizing manufacturing though" just because there's threads doesn't necessarily mean there was a screw.
<samueldr> there was, I already took it to bits before
__monty__ has quit [Quit: leaving]