<andi->
Yeah, there were nixpkgs releases until 0.14 and then in 2015 the first "proper" release war 15.09
<samueldr>
was the move to github the same as the move to git?
<andi->
I can't tell from the history alone. I scrolled through it for a while.. Apparently eelco was already using git locally but importing from SVN?
<samueldr>
abathur: you saw (I guess) that github will destroy most usefulness of its search
<abathur>
is *deduplicate identical or N%-identical entries*
<samueldr>
yeah
<samueldr>
that would be useful
<samueldr>
put them all under the same "meta-result"
<abathur>
yeah
<samueldr>
with a way to expand upon it
<samueldr>
you *may* want to search for the same file!
<abathur>
nod
<samueldr>
that's the reason I skipped PREBUILT here
<abathur>
sometimes that's *exactly* what you want
<samueldr>
and it'd help having them grouped
<abathur>
but when you want to survey usage of something
<abathur>
and get 50k results
<samueldr>
yep
<abathur>
and 49k of them are the same file from the same popular project
<abathur>
well, more like the same 6 test suite files :P
<pie_>
OH GOD PLEASE YES <abathur> is *deduplicate identical or N%-identical entries*
<abathur>
the other thing I'd wish for is just better path matching, but it's possible there are search syntax affordances I just don't grok
<pie_>
their operators are named weird anyway
<abathur>
path search is nice in a single repo, and occasionally helpful outside of that, but sometimes useful to be able to express that I want something that is a child of a directory named test(s) that uses this function
<pie_>
syntax isnt that hard to deal with cmon
<pie_>
so i think it should be a matter of wanting to do it
<samueldr>
> syntax isnt that hard to deal with
<{^_^}>
error: syntax error, unexpected WITH, expecting ')', at (string):452:31
<lovesegfault>
but they are like, idk, alaskan explorer pants
<lovesegfault>
it's like wearing an oven
<lovesegfault>
it's not cold enough here 🤣
<eyJhb>
lovesegfault: You got what you wanted :p
<eyJhb>
They are WARM :D
<infinisil>
eyJhb: Ah no, i think the white ones are actually called "british short hair" xD
<eyJhb>
OHHHH!
<eyJhb>
That is the none cuddly one right?
<infinisil>
All of them are pretty cuddly :D
<eyJhb>
British Shorthair are usually less cuddly :p Does it like to be picked up?
<eyJhb>
Or, they show affection quite different
<infinisil>
They all love to be pet, can be picked up, mostly love cuddling, pretty normal cat stuff!
<eyJhb>
Hmmm.... Still love British Shorthair, most are just more "leave me the hell alone", especially from strangers :D "I" (parents) have had British Shorthair until I was like 16.. And then my brother got one
__monty__ has joined #nixos-chat
thibm_ is now known as thibm
<infinisil>
Are there any games that balance different items based on the win percentage of the people using them?
<infinisil>
That seems like an interesting idea, but I haven't seen it implemented
<sphalerite>
pretty sure that's part of how balancing is done generally?
<infinisil>
sphalerite: Oh yeah, but I mean like in an automatic way
<infinisil>
So that like, every game gets sent to a server, which then calculates the effectiveness of each item, then nerfs/buffs them accordingly
<infinisil>
This could have interesting effects, like "If I win a lot with this item, and it's not getting nerfed, I'm just better than most other players with this item, and it's not just OP"
<__monty__>
I'd expect such continuous changes to just be frustrating for 80-90% of the player base though.
<__monty__>
Why bother saving up for that sweet epic gear if it's just gonna get nerfed?
<infinisil>
__monty__: I'm not sure, should games become easier the longer you play them?
<infinisil>
This auto-adjusting would probably encourage people to really become better with certain items
<infinisil>
Actually every item has a chance to become your "main" one, because just by using it very often, more than other players, you get better at using it
<infinisil>
Though I guess that depends on the game, and what the items do
<__monty__>
Dunno, sounds like making gaming a chore tbh.
<__monty__>
I don't mind grinding to get some good gear.
<__monty__>
But I don't want to keep doing that.
cosimone has joined #nixos-chat
<infinisil>
Hm yeah it kind of seems like an entirely different goal
<__monty__>
Maybe I'm biased because I'm a casual player but I think most playerbases are more casual?
<infinisil>
With this model your goal would be to find the items you're good at
<__monty__>
*the majority of most playerbases
<__monty__>
Hmm, you're suggesting making *all* items equally good?
<__monty__>
Are we talking RPG or something like TF2?
<infinisil>
__monty__: I was thinking of some generic shooter
<infinisil>
I guess yeah, all items would be equally good over time, but only for the average player
<infinisil>
Oh, bad consequence: If you're just bad at the game in general, all items would be bad for you, and your win percentage would be very low
<infinisil>
Though, I guess that kind of makes sense, and is probably the case without this model as well
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos-chat
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos-chat
cosimone has quit [Remote host closed the connection]
<abathur>
I only noticed because I want to fish for feedback on the resholve Nix API, so I was looking to see if anyone I wasn't already aware of was using it in public--and literally got 0 code results for resholve (no lang filter)
<infinisil>
"I send you this message with heavy tears but in good faith."
<infinisil>
Oh boy here we go again
<infinisil>
"My name is Mrs. Monika Jocelyn Greg, battling with cancer. I have a legacy of$ 2.8 million from my late husband."
<infinisil>
Of course
<infinisil>
"For your information, My late husband also deposited 37kg of (GOLD)in a neighboring country"
<infinisil>
Gold as well!
<andi->
I take the cash, you keep the gold.
<infinisil>
It's actually pretty close
<bbigras>
shipping gold must cost a lot.
<infinisil>
> 37*60492 # <- gold price per kg
<{^_^}>
2238204
<matthewcroughan>
LinuxHackerman: hey
<matthewcroughan>
So yeah, LinuxHackerman gchristensen, how can I provision a nixos system.
<matthewcroughan>
Disk partitioning and all. No monitor/keyboard/mouse
<sphalerite>
matthewcroughan: where does it live? What access do you have? Is there an existing OS on it?
<sphalerite>
(I'm LinuxHackerman also)
<sphalerite>
(and actually this could be more suitable for #nixos :p )
cosimone has quit [Quit: cosimone]
endformationage has joined #nixos-chat
<eyJhb>
infinisil: That white cat, did there ever come some images? :D
<__monty__>
infinisil: Heh, do us nixers all get the same spam?
<eyJhb>
sphalerite with the many names!
<adisbladis>
Multiple personalities :3
<eyJhb>
(and thereby the conclusion is, that many is defined as... one, many.. -> 1++ -> many)
<eyJhb>
I don't blame them adisbladis ! :D
<eyJhb>
Fuck I want dinner soon
<lovesegfault>
bbigras: no, but I've been having other issues
<lovesegfault>
it seems like the magic rollback bin keeps running forever
<lovesegfault>
eating CPU
<infinisil>
eyJhb: Um, will make a pic soon!
<infinisil>
__monty__: We're gonna be rich!
cole-h has quit [Quit: Goodbye]
cosimone has joined #nixos-chat
<sphalerite>
Mic92: do you maintain mumble.nix-community.org?
<Mic92>
sphalerite: no, but I know who does
<Mic92>
ehmry
<sphalerite>
it would be cool if it had a LE cert, and also if it were also on mumble.nixos.community :D
<eyJhb>
infinisil: Sounds great :D
<makefu>
last time we tried mumble.nix-community.org we has issues with the performance of the instance and we had to switch back to the instance of lassulus
<infinisil>
eyJhb: ^ this is the one that snuck under my cover :)
cosimone has quit [Read error: Connection reset by peer]
cosimone has joined #nixos-chat
thibm has quit [Quit: WeeChat 2.6]
cosimone_ has joined #nixos-chat
cosimone has quit [Ping timeout: 240 seconds]
cosimone_ is now known as cosimone
<makefu>
sphalerite: as i said, we tried but encountered some issues which we couldn't fix ourselves - and had to move
<sphalerite>
yeah fair enough
<sphalerite>
tazjin: currently living through some frustration with the state of nix. Is tvix in its current state a working all-purpose replacement for nix stable?
<__monty__>
Cool, sounds like a long way off though.
__monty__ has quit [Quit: leaving]
<colemickens>
Why did no one tell me that you can get 164ft Thunderbolt 3 cables for <$500?
<colemickens>
This is giving me ideas.
<eyJhb>
infinisil: Cute kitty! :D <3
<eyJhb>
I feel like `One goal is to slowly bring the codebase in line with the Google C++ style guide` could as well be to slowly move away from C++, and maybe use something like Rust
<eyJhb>
That would be cool
<samueldr>
colemickens: you never asked
<samueldr>
I wouldn't be surprised that part of the fork's concerns are about the fact that nix is starting to use rust
<samueldr>
the concerns are all so hand-wavey that they might as well just have said "because we feel like it"
<samueldr>
not all, but some
<sphalerite>
I don't think so, my impression is that it's a lot more about flakes.
<samueldr>
hopefully they won't diverge too much in compatibility with Nix so that users of tvix end up in a situation where their results cannot be trusted
<samueldr>
but having diversity in implementations is generally good
cosimone has joined #nixos-chat
<drakonis>
ooo
<drakonis>
tvix looks interesting...
<colemickens>
"In addition, the upstream Nix project is diverging from the opinions of some community members via the introduction of concepts such as Nix flakes.".
<colemickens>
Can someone please give me a single coherent argument against flakes?
<colemickens>
I literally deal with multiple NIX_PATH issues every single week while supporting users in #nixos.
<bbigras>
those damn sjw snowflakes
<bbigras>
^ /s
<drakonis>
lol
<samueldr>
not cool
<drakonis>
nix_path is hell
<samueldr>
colemickens: NIX_PATH or misconceptions about which NIX_PATH and when, coupled with nix-channel? :)
<samueldr>
but yeah, I definitely can see that getting away from NIX_PATH can be a good goal
<colemickens>
It is.
<samueldr>
though there's been so much FUD and so much hype thrown around for flakes
<colemickens>
That exact hellacious workflow plays out in #nixos multiple times a week and is a nightmare that simply can't occur if someone shares their flake.nix/flake.lock file.
<samueldr>
and, imo, there's a rush towards "flaking everything" while itself, flake is not even done?
<samueldr>
colemickens: yeah, your list points towards my interpretation of the main problem :) NIX_PATH is hard
<samueldr>
there's too much nuance in there
<samueldr>
the defaults can do so much
<samueldr>
and the defaults can definitely be bad in the reproducibility sense
<samueldr>
added with the differences coming from changes outside Nix
<samueldr>
e.g. sudo and environment variables
<samueldr>
this was a recent change that changed how it acts
<samueldr>
what I would have liked to see is a "nix path companion", first
<sphalerite>
I'm less in the "gaaah flakes is terrible" camp but all-in on the "nix's code quality is bad" aspect.
<sphalerite>
samueldr: you mean nix-channel? :p
<samueldr>
maybe that's what niv is, I don't know
<samueldr>
sphalerite: for a companion? no
<samueldr>
something that can relate a specific nix path to a project
<samueldr>
that nix knows about
<drakonis>
i wonder when tvix started
<samueldr>
so <nixpkgs/...> uses that
<sphalerite>
oh right so direnv
<colemickens>
If the solution doesn't involve pure eval, I'm uninterested.
<sphalerite>
:p
<colemickens>
Not to be brash, but.
<samueldr>
sphalerite: not exactly
<samueldr>
colemickens: make <> pure
<samueldr>
that's what I mean here
<sphalerite>
drakonis: May this year according to the git log.
<colemickens>
But my point is that you can't unless you force pure eval.
<samueldr>
flake is doing a bunch more than only dealing with paths
<drakonis>
hm
<drakonis>
interesting.
<samueldr>
(AFAIUI, I don't really know much details about flakes)
<samueldr>
colemickens: that's thinking with current nix
<samueldr>
colemickens: if you're changing nix, you can change <>
<colemickens>
okay. I guess I'd be curious to understand why you want that over specifying it explicitly next to the nix code
<samueldr>
hard to describe what is basically an unformed thought
<samueldr>
but in such a "companion" you would have the paths next to the nix code
<samueldr>
but strictly dealing with that, no other features
<colemickens>
fair enough, I think `niv` is pretty close to what you're describing.
<samueldr>
[18:14:11] <samueldr> maybe that's what niv is, I don't know
<samueldr>
:)
<colemickens>
yeah
<samueldr>
but *also* something that's easy to break-out from
<samueldr>
I don't know if flakes is hard to break-out from
<samueldr>
I still haven't invested any time in flakes
<samueldr>
default to hermetic builds, but trivial to lift the lid up
<colemickens>
I'd put it this way, if someone can eval my flake config and get a different hash output, I'll eat a shoe :P
<colemickens>
like, a small candy shoe.
<samueldr>
aw
<drakonis>
this guy knows his shit
<drakonis>
at least wasnt the guy that promised to eat his leather slippers
<sphalerite>
colemickens: I'll eat a small candy shoe if you give me one, I don't even need anyone to evaluate anything
<colemickens>
just a bit of an admitted flake-fanboy, but worried about being too much so, so I'll back off :P
<samueldr>
I think both of the FUD and ultra-hype around it have probably been causing more problems in image in the end
<samueldr>
especially at a time it was far less "done"
<samueldr>
(I don't know if it is close to done)
<samueldr>
(I don't know of a roadmap)
<samueldr>
which maybe is also part of the problem
<samueldr>
I don't know what is yet to be done, if anything, what is complete
<colemickens>
My biggest frustration with nix right now is that there isn't a roadmap (for flakes)
<drakonis>
there isnt a roadmap for anything right now
<colemickens>
I do feel guilty talking about them when they're not stabilized.
<drakonis>
eelco was going to create one in 2018
<drakonis>
there's a lot of closed doors development going on right now
<drakonis>
the only things i'm reasonably aware that are going on right now with nix are flakes and CAS
<drakonis>
ah and nickel
<drakonis>
but that one's still far from being usable
<drakonis>
i think nix really needs some sort of clean slate
<sphalerite>
rewrite in rust /s
<drakonis>
hahaha, wasnt that going to happen at some point?
<ekleog>
just saw that nixpkgs on repology is top (by far) by number of non-unique packaged projects… I seemed to remember we were tricking the system with our autogenerated package sets, how come?
<ekleog>
do we actually have that much more packages than other distros? :D
<drakonis>
we had a discussion on repology in the past
<drakonis>
should be in the logs
<ekleog>
also AFAIR rust in nix is slowly happening, so… :D
<drakonis>
it got axed
<ekleog>
oh :(
<drakonis>
the rust bits got replaced with c++ bits
<ekleog>
well in a way that makes sense, having just a few rust bits doesn't make much sense if it's not in an actual rewrite project
* zgrep
wonders how much "it made sense at the time, but a different approach would be better nowadays" has been enshrined in the way nixpkgs works, and not just the nix utility itself.
<drakonis>
a lot
<zgrep>
I guess new or better tooling for querying and manipulating large swaths of nix-language code would probably be helpful in mitigating the effects of that.
<drakonis>
yes
<drakonis>
the design of nixos modules is one of those things
<zgrep>
How so?
<drakonis>
services arent functional
<drakonis>
i think that was one of the problems
<drakonis>
for the longest, extending modules was basically extraconfig
<ekleog>
there's also “and switching to the better approach would cost a big lot of time and effort”
<colemickens>
Random question, but if I got my RMA'd RAM back, and put it in, and I'm immediately getting BSODs, what do you think you think the chances are that 1) the new ram is bad, or 2) my mobo is bad? My backup RAM that worked fine was only 2 sticks, versus the new kit being all 4 slots?
<ekleog>
try plugging only 2 of the 4 sticks maybe?
Baughn has quit [Ping timeout: 246 seconds]
Baughn_ has joined #nixos-chat
Baughn_ is now known as Baughn
<colemickens>
Yeah, I'm just going to have to try 2 bad in both sets of slots, and then the good ram in the other two slots, that should at least tell me which is the culprit
<samueldr>
unless the good ram makes the bus go slower and it doesn't exhibit that issue?
<colemickens>
I hadn't considered anything like that, ... hm. I wonder, is a mobo memory slot going bad a thing that actually happens IRL?
<drakonis>
hmm
<drakonis>
the stick is quad channel?
<ekleog>
I'd say the answer to “is $hardware going bad a thing that actually happens IRL” is probably “yes” regardless of the value of $hardware
<drakonis>
is your motherboard also quadchannel?
<drakonis>
this is important
cosimone has quit [Quit: cosimone]
<colemickens>
I am finding out that I know less about memory than I even realized.
<colemickens>
The RAM is sold as a 64GB kit of 4 sticks, so I assume the ram is meant to be used this way.
<colemickens>
The mobo is Asus Z170-A, trying to see if it is quad channel or not.
<samueldr>
there's something called XMP I don't really know much about
<samueldr>
not sure if it can be relevant, but AFAIK it is for speeds