<savanni>
happy dance!!! With secure boot turned off, my machine installed just fine. With kernel 5.3.7 and no extra kernel options, ACPI works! I can suspend and resume the machine!!!
<savanni>
I'll have to post my configuration file in a gist as soon as I have X working.
<infinisil>
ar: Aw yeah I got a converter working in Nix :P
<infinisil>
For all unicode
<infinisil>
Well except the null byte, but that's not a valid SSID anyways
<clever>
joepie91: hmmm, yeah, its pinned to channel 5 right now...
<joepie91>
clever: does your router show you an interference/congestion graph for channels?
<clever>
i have an android app that will
<joepie91>
that works too :)
<clever>
i dont see much on channel 1
<joepie91>
also I should sleep...
<clever>
The device is restarting...
<joepie91>
huh. usually channel 1 is congested out the wazoo
<clever>
silly router, restarting is for windows, lol
<joepie91>
because a ton of shitty cheap routers default-pin to 1
<clever>
i only see 4 networks in the 2.4ghz band
<clever>
and the instant i say that, a 5th appears on 1, lol
<joepie91>
lol
<joepie91>
oh also
<clever>
the rest are on 7? 9 and 11?
<joepie91>
an SSID I ran across while travelling today
<joepie91>
"Bill Wi the Science Fi"
<clever>
the one on 9 also appears to be rather fat
<clever>
and covers the same range that the 7 and 11 cover seperately
<clever>
no change with it set to channel 1
<joepie91>
hm dunno then
<joepie91>
I'm going to bed though :P
<joepie91>
falling asleep
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 250 seconds]
<jackdk>
I have also seen "Pretty fly for a Wi-Fi"
xd1le has joined #nixos-chat
drakonis1 has joined #nixos-chat
<ashkitten>
augh, questionable content's rss feed did the thing again where it doesnt have any comics for a while and then it just lumps a bunch into one post
<savanni>
@ashkitten: Oh, I thought I was the only one that was happening to...
<ashkitten>
yup
endformationage has quit [Ping timeout: 276 seconds]
drakonis1 has quit [Quit: WeeChat 2.6]
<ar>
clever: your device doesn't have 5GHz wifi?
<ar>
infinisil: i'm confused. for some networks iwd needs this funky configuration file name, for some it needs the plan-ssid.psk
<ar>
hm. iwd manpages are even less useful than some of the gnu manpages saying that "useful manual is in the info pages"
<ar>
how did they achieve that? their man pages don't even have links to a useful documentation
<ar>
ah, the logic for whether to use hexstring or not is: for (c = ssid; *c; c++) if (!isalnum(*c) && !strchr("-_ ", *c)) break;
chreekat has joined #nixos-chat
<ar>
also interesting, despite setting dns_resolve_method=resolvconf, iwd still tries to call resolved over dbus
<ar>
src/resolve.c:resolve_resolvconf_init() Trying to find resolvconf in $PATH
<ar>
src/resolve.c:resolve_resolvconf_init() resolvconf found as: /run/current-system/sw/bin/resolvconf
<ar>
oh, great, it works if i start the thing manually with --debug, but not when started as a systemd unit
adamt has quit [Quit: Connection closed for inactivity]
<ashkitten>
i was asked to implement something like windows' libraries on linux using fuse, and i'm a bit torn
<ashkitten>
i wish i could wait a few months and use the upcoming version of the fuse crate with more idiomatic stuff like futures and such
<ashkitten>
i guess i'll just suck it up and rewrite it later whenever that happens
<ashkitten>
(basically this would work similarly to an overlay filesystem with a key difference: edited files are always modified in their original location)
<ashkitten>
wonder how windows deals with name collisions
<ashkitten>
hmmmmm
<ashkitten>
really, i should just implement this as symlinks to the original files
<ashkitten>
that would be closer to the windows implementation
<ashkitten>
the thing is, afaik windows doesn't actually implement it at the filesystem level, rather at the ui level
<ashkitten>
so then libraries aren't actually addressable outside of explorer, and they don't need to do any name collision resolution because they can differentiate within the ui
<ar>
>windows' libraries
<ar>
you mean like Music, Pictures, Documents, and so?
<ashkitten>
yeah
<ashkitten>
they're collections of the contents of multiple directories
<ar>
hm. systemd.services.iwd.path = [ pkgs.openresolv ]; didn't fix the above; now i get "iwd[1658]: Cannot write to /run/resolvconf/lock" "iwd[1658]: resolve: /nix/store/0sh5b0jf27znkpy9wrzna3yyq4yd78x7-openresolv-3.9.2/bin/resolvconf exited with status (256)."
<ar>
which is interesting, because iwd is ran as root
<ar>
ah, because /run is ro for iwd process
__monty__ has joined #nixos-chat
<sphalerite>
hm, I want to configure samba to only allow access from my LAN. For ipv4 this is simple enough — hosts allow = 192.168.2.0/24 — but I want to allow IPv6 as well, and I have a dynamically allocated (=> changing) IPv6 prefix over time… Any ideas?
<__monty__>
Isn't that what the fixed by mac-address link local IPv6's are for?
<sphalerite>
I don't have a way to ensure that it's only ever accessed via link-local addresses though
<sphalerite>
clients may use their global addresses too
<eyJhb>
Yeah, but I am considering if you could make some "NixOS" like, but just Android. E.g. declaring settings, apps, etc. adn using a recovery menu to switch between versions
<clever>
ar: every single time i tried to upgrade to a N band router, bad things have happened :P
<clever>
ar: the first one, did ipv6 route advertisement, even with ipv6 disabled, so it null-routed all v6 traffic
<ar>
clever: i don't remember how long ago i upgraded *from* N-band router
<clever>
ar: the latest one, wont allow traffic from wifi->wired, so the wifi is basically useless
<eyJhb>
infinisil: turned togames done quick :p
<ar>
but i stopped using my accesspoints as routers sometime before that
<clever>
and of course, no option exists to disable that security
<clever>
ar: i need to just buy a proper access point, and stop abusing routers as AP's
<clever>
i'm just puzzled by the sudden and extreme dip in performance
<ar>
that reminds me i should replace my edgerouter-lite with a nixos box
<clever>
i also started testing hostapd on my nas, as an ap, lol
<clever>
i set it up, confirmed it works, then never tested it
drakonis1 has joined #nixos-chat
<gchristensen>
andi-: if I may suggest, swapping the } and { eys
<__monty__>
gchristensen: Is this a ploy to sew the seeds of mass confusion among nix contributors?
<ar>
/39/46
<gchristensen>
no
<gchristensen>
on his keyboard, pressing the top left button inputs a ], and the top right bottun inputs a [
<__monty__>
I can see how that can be confusing.
<ar>
what keyboard is that?
<__monty__>
Another kinesis advantage probably?
<gchristensen>
yeah
<etu>
Oh, it's donate to nixos day
<etu>
:)
<__monty__>
Where do I send my kidney cooler?
<adisbladis>
__monty__: I know a guy
psyanticy has joined #nixos-chat
<eyJhb>
The swedish mafia.
<eyJhb>
It is always nice getting a contract, where you explicit are hired to work on your opensource project and have agreed that they will not own the code at all, etc. and then there is a "we own everything you work on in work hours" thing in the contract
<eyJhb>
God damn it
<__monty__>
Doesn't actually matter. They agreed to contribute it under a foss license. Not like they can take that back.
<adisbladis>
eyJhb: Wouldn't one clause invalidate the other?
<adisbladis>
Also it doesn't really matter who owns the copyright if the code is in the open anyway?
<eyJhb>
I only have the spoken word on that... So that wouldn't make much sense to enforce
<__monty__>
adisbladis: Not necessarily. The former might be talking about outside of working hours. Which isn't valid in an employment contract afaik? Might depend on labor laws.
<eyJhb>
Not really, as 1 years of work would fall to them within a day and I could get fired right after
<__monty__>
What do you mean?
<__monty__>
Any code you've pushed to a foss repo is covered by its license.
<ar>
speaking of fun clauses in contracts, doesn't google have a "we own everything you do while employed, incl. in your free time"?
<eyJhb>
__monty__: regarding who holds the copyright to the software. I wouldn't want to have spent so much time and energy on it, and then just "give" it to them where I could potentially get fired
<eyJhb>
I actually think this clause would try to take-over such a thing __monty__
<eyJhb>
Sounds about right ar. Seems the same for Disney :p
<__monty__>
eyJhb: What do you mean? They literally can't. There's no such thing as "unpublishing" contributions.
<__monty__>
And copyright claims wouldn't hold up.
<eyJhb>
__monty__: Google Translate `This applies whether or not these can be registered under the laws of patents, designs, copyrights, designs, trademarks or not.`
<eyJhb>
But yeah, it doesn't specify existing things
<__monty__>
That's super far-reaching and I can't believe it'd hold up if actually contested in a european court.
<eyJhb>
Yup, it really is. That is why I do not like that being in my contract...
<eyJhb>
Also they have some forced labor which I cannot deny in that contract under any circumstance
<eyJhb>
Which is also weird :p
<__monty__>
Which might work out to your advantage. Because I *think* if you can show that clause is not valid under labor laws, all other clauses are disregarded too, making the contract null and void and return *all* of the copyright to what you produced to you.
<__monty__>
I'd hire a lawyer if I were you, seriously.
<eyJhb>
The joys of uni! Lets find a professor which can guide me :D
<__monty__>
Just go to a big law firm to have the contract checked and explained to you.
<__monty__>
Uni profs might not be quite honest about uni contracts.
<eyJhb>
It isn't a uni contract, it is a external private company
<FaeFly>
I understand things like that can often be negotiated away (and yes, likely not enforceable in EU, but not a lawyer etc...)
<eyJhb>
Also, as long as I don't talk to HR at uni, professors are really honest at not liking HR :p
<eyJhb>
Yeah, I hope so at least. I have written to them and asked about it.. Actually one of the first contracts I have read, even though I have like signed 10+ now I think
<andi->
gchristensen: go ahead, I have backups :P
ddima has quit [Quit: Lost terminal]
drakonis1 has quit [Ping timeout: 245 seconds]
drakonis has quit [Ping timeout: 276 seconds]
drakonis has joined #nixos-chat
drakonis1 has joined #nixos-chat
endformationage has joined #nixos-chat
<kraem>
ar: do you mind sharing what options you set for iwd to function properly? are you using it's builtin dhcp functionality?
<ar>
kraem: well, userControlled.enable = true; is irrelevant
<kraem>
ah does it need the systemd.services.iwd part for it to have access to openresolv?
<etu>
Hmm, smartcards/gpg seems to be broken on latest unstable :/
<etu>
Can't decrypt or sign anything
<etu>
but gpg --card-status shows me the smartcard
<adisbladis>
etu: Oh!
<adisbladis>
I know this one!
<adisbladis>
I hacked it locally when I didn't have an internet connection and forgot about it
<adisbladis>
Something changed with how pinentry works, I'm still not sure about the exact root cause
<adisbladis>
etu: My "fix" was to hack home-manager http://ix.io/20vo
<adisbladis>
Not pretty at all :/
<etu>
:(
<kraem>
ar ^ :)
<adisbladis>
I have no idea why gpg-agent suddenly started looking up pinentry relative to itself
<ar>
kraem: yes, without "path = [ pkgs.openresolv ];" it won't find the resolvconf binary, and without serviceConfig = { ReadWritePaths … it won't be able to: acquire a lock at /run/resolvconf/lock; resolvconf won't be able to update resolv.conf
<pie_>
<y> x: the other way round. debian/ubuntu/other distros with properly implemented CM feel like a poor substitute of nixos
<pie_>
<x> maybe it's called mirageos
<pie_>
<x> y: i want the operating system that nixos feels like the poor substitute for, where everything is just declarative to begin with
<pie_>
<x> y: yes, I know. I'm saying that even NixOS is itself a hack, like you said of CM in general
<pie_>
<x> because we don't have operating systems that just expose declarative objcap configuration and orchestration APIs in the first place
<waleee-cl>
x do realize that there's a real thing called mirageos?
<gchristensen>
objcap I agree with
<gchristensen>
but I disagree that nixos is a hack, obviously :P
<ar>
pie_: i'm here and i'm ok with being quoted here
<ar>
pie_: the other person is here too
<MichaelRaskin>
gchristensen: NixOS module system does give an impression of «now let's spend a ton of effort to emulate distributed shared-state imperative computing in a purely functional language»
<gchristensen>
sure
<savanni>
Hey, y'all. in case you are interested, I created a repository for my system-level NixOS configuration. Probably interesting to folks running Ryzen thinkpads.