<elvishjerricco>
aleph-: I don't think I've even heard of the elvish shell :P
<aleph->
Heh wrong again
<cole-h>
I sure do love getting unsolicited emails because I contributed to a project on GitHub
<aaronjanse>
cole-h Rust project?
<cole-h>
Yep
<aaronjanse>
All the Redox people got an email
<cole-h>
I got it because of Alacritty.
<aaronjanse>
Ah
<cole-h>
I informed them they are in direct violation of GitHub's Acceptable Use Policies
<aaronjanse>
Thank you
<cole-h>
and am going to report them to GitHub :)
<aleph->
Heh what's this?
maljub01 has joined #nixos-chat
<aaronjanse>
Someone contacted all developers on popular Rust projects for some survey
<bbigras>
those emails are frequents. "surveys" and shit like that
<abathur>
at least it wasn't a pitch to write a book about git+github?
kini has joined #nixos-chat
endformationage has quit [Quit: WeeChat 2.9]
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
waleee-cl has joined #nixos-chat
cole-h has quit [Ping timeout: 260 seconds]
FRidh has joined #nixos-chat
<eyJhb>
aleph-: nftables are atomic, it is "easy" to configure, you can merge multiple rules into one, have one file that dictates all the rules (ie. don't have a .sh script which executes a ton of iptables stuff), etc.
<Ox4A6F>
samueldr: He is informed and sorry.
rajivr has joined #nixos-chat
<eyJhb>
supersandro2000: Might not be so, but did you have to patch systemd to get it to work? :p
<eyJhb>
Also, just plugging this here. NixOS, because `systemd.package = pkgs.systemd.overrideAttrs(old: { patches = old.patches ++ [ ./systemd-test.patch ]; });` <3 <3 so easy to throw in a patch
<patagonicus>
eyJhb: don't know specifically for that one, but for some options you can have multiple Foo=bar lines, so I'd try that.
__monty__ has joined #nixos-chat
Synthetica has joined #nixos-chat
<eyJhb>
patagonicus: Might have a look at that later :)
<eyJhb>
Wondering, how many of your mobile providers support IPv6? Ie. accessing ipv6.google.com
<hexa->
my dudes, this smells very on topic
<Synthetica>
eyJhb: mine doesn't; Tele2 NL
<sphalerite>
Anyone have any fairly-simple-to-do-at-home suggestions for working out why a laptop might be dead? (dead = nothing happens when power button is pressed, regardless of battery presence/mains power connection
<sphalerite>
Anyone have any fairly-simple-to-do-at-home suggestions for working out why a laptop might be dead? (dead = nothing happens when power button is pressed, regardless of battery presence/mains power connection)
<sphalerite>
I've tried reseating the RAM and using only one of the two modules at once, but I would have expected beeps at least if the RAM were dead
<ldlework>
nix should get a bounty program
endformationage has joined #nixos-chat
<leonardp>
sphalerite: is there an LED anywhere that lights up when the charger is plugged in?
<sphalerite>
yeah but I forgot whether it lights up and don't have the charger here >.<
<sphalerite>
what I did get at one point though I'm not sure if it was just an illusion is the power button lighting up very very briefly when I pressed it
<leonardp>
do you have a spare battery you can try? if not, i recommend getting that charger -.-
<LinuxHackerman>
yeah, that's the plan for Monday ^^
<leonardp>
if there is anything lighting up it might be that it's only the inverter of the display
<LinuxHackerman>
no, the display remains off
<LinuxHackerman>
though I'm leaning towards just salvaging what nice removable components there are and giving up on the rest
<LinuxHackerman>
shame, because it's only 4 years old and fairly decently specced (and has a removable battery <3 )
<leonardp>
only dark or off? i had a laptop with a dead inverter once.. if you look very closely you might be able to make out something on the display
<leonardp>
Linux Hackerman: damn that sucks :/
<LinuxHackerman>
pretty sure it's off, since the power LED doesn't light up as it normally would either
<leonardp>
too bad
<LinuxHackerman>
it has a JTAG header position on the board and I'm kind of curious to see if anything interesting's available on there, but it's unpopulated and I have far too many side projects already x)
tomberek has quit [Quit: Ping timeout (120 seconds)]
tomberek has joined #nixos-chat
<ar>
sphalerite: i wouldn't be surprised if EC got stuck somehow, and there's probably a separate small flat (coin cell) battery
<ar>
sphalerite: what laptop is that anyway?
cole-h has quit [Ping timeout: 240 seconds]
<sphalerite>
ar: yeah tried disconnecting that battery and waiting for a while, no luck. It's a thinkpad t470p
<ar>
sphalerite: you might also have a reset pin hole on the bottom
rajivr has quit [Quit: Connection closed for inactivity]
Baughn has quit [Read error: Connection reset by peer]
Baughn has joined #nixos-chat
<sphalerite>
ar: nope, but I did find a button on the board which I tried pressing :D
ravndal has quit [Quit: WeeChat 3.1]
ravndal has joined #nixos-chat
<supersandro2000>
eyJhb: no? Why should I? It just works ™️
<Raito_Bezarius>
how hard would be to have a mechanism to push notifications / emails / whatever when a PR appears in a channel / a critical security update pops?
<Raito_Bezarius>
as far as I understand, most of the work has been done in qyliss's PR tracker and it can be extended right?
<hexa->
probabl<
<hexa->
most everything security is tagged as such
<hexa->
but tags are only on pull requests, not on commits
<Raito_Bezarius>
is that an issue that tags are only on PRs and not on commits?
<Raito_Bezarius>
I would guess the nice primitive would be to say something like once PR X lands in channel Y, send push notification to Zs
<Raito_Bezarius>
and then to do it for all PR X such that they're tagged security
<hexa->
sure
<hexa->
if PR lands in nixos-20.09 (as opposed to release-20.09)
<Raito_Bezarius>
what's the exact difference between those?
<Raito_Bezarius>
is it something like nixos-20.09 is the backport channel and release-20.09 is the frozen channel?
<hexa->
nixos-20.09 is whats in the channel
<hexa->
it is after release-20.09 went through hydra
<Raito_Bezarius>
alright
<hexa->
committers never push to nixos-20.09
<Raito_Bezarius>
is it ofborg or hydra's responsibility to do it?
<hexa->
hydra
<hexa->
ofborg is the ci runner integrated into githubs pr system
<Raito_Bezarius>
alright
nf has joined #nixos-chat
NinjaTrappeur has quit [Quit: WeeChat 3.1]
NinjaTrappeur has joined #nixos-chat
jess has quit []
<bbigras>
the Starsector game had a big update yesterday. It's not free but it works with steam-run.
ece has quit [Read error: Connection reset by peer]
ece has joined #nixos-chat
spacekookie has quit [Quit: **aggressive swooshing**]
spacekookie has joined #nixos-chat
<aleph->
eyJhb: Interesting. Wasn't there someone who wrote a nftables module?
spacekookie has quit [Quit: **aggressive swooshing**]
spacekookie has joined #nixos-chat
<eyJhb>
aleph-: I think so! But it tried to reimplement the quirks of the current module...
<eyJhb>
As a drop-in replacement
<aleph->
Blegh
<samueldr>
Ox4A6F: no need to be sorry, I was only saying :)
<eyJhb>
I have used networkd + nftables very little, but I enjoy it very much so far
<eyJhb>
Same goes for IPv6. Trying to push my phone company to support it, when I am on mobile data... Annoying that they don't
<bbigras>
"Trying to push my phone company to support it" good luck with that. No sure the "comonnnn" argument would work.
<supersandro2000>
WTF
<bbigras>
?
<supersandro2000>
Germany is place 3 in IPv6 adoption
<supersandro2000>
that can't be the result of cable companies or the state
<supersandro2000>
probably some big corp pushed for it
<bbigras>
"i find your lack of faith disturbing"
<supersandro2000>
but at least when it is about bandwidth Germany is way behind
<supersandro2000>
and the biggest provider created his own Internet Exchange almost no one peers with so going over Level 3 in the US is sometimes faster for connections to other parts of Europe
<supersandro2000>
bbigras: mine? Faith that the gov does internet right? 😂
<bbigras>
haha. yeah. I was joking. the quote just sounded good for your lack of faith in the gov
<supersandro2000>
lets be realistic. There is room for improvement
<leons>
supersandro2000: what are some of your reasons to conspire for "big corp" pushing towards IPv6?
<bbigras>
I'm sure. every gov are so big they must be huge clusterufcks
<supersandro2000>
leons: Germany is way behind in Internet speed and mobile coverage and it is over priced
<supersandro2000>
it more plausible that IPv6 is an accident than intentional
<samueldr>
it's likely that mobile internet providers (cellphone internet) use IPv6 with a NATed IPv4 or something like that, if it's like here
<samueldr>
that alone could be inflating the numbers
<samueldr>
(I might be using the wrong terminology when I say NATing IPv4)
<samueldr>
I hate networking, and networking hates me just as much
Baughn has quit [Ping timeout: 260 seconds]
<LinuxHackerman>
Yeah my home internet is native IPv6 with IPv4 only tunnelled (I think it's called DS-Lite, not to be confused with the nintendo handheld)
<supersandro2000>
I think mine, too
<supersandro2000>
maybe because you can just do IPv6 software side and do not require new cables, hardware etc
<LinuxHackerman>
no, the other way round. IPv4 is tunnelled in IPv6.
<LinuxHackerman>
Oh yeah IPv6 compatible cable is important hahahahaha
<Raito_Bezarius>
in France, Bouygues implements IPv6 by default on mobile network
<Raito_Bezarius>
with NAT64/DNS64 also
<Raito_Bezarius>
no native IPv4
<Raito_Bezarius>
that's pretty cool
<Raito_Bezarius>
I don't need to run WireGuard on LTE :>
<supersandro2000>
LinuxHackerman: but is it IPv4 compatible, too?
<LinuxHackerman>
All my IPv4 is wrapped in IPv6.
<Raito_Bezarius>
LinuxHackerman: so all DNS resolutions are some special IPv6?
<Raito_Bezarius>
or do you receive an IPv4 on your end devices?
<LinuxHackerman>
Raito_Bezarius: no, on the end-user device side it all looks the same as usual
<Raito_Bezarius>
alright
<LinuxHackerman>
but my router wraps the IPv4 packets in IPv6 before sending them into the modem
<LinuxHackerman>
(and the IPv6 in PPP, but that's another story)
<LinuxHackerman>
yep exactly, DS-Lite as I mentioned above :)
<LinuxHackerman>
supersandro2000: cables don't care what you send over them :p
<LinuxHackerman>
at least on that level
<leons>
Linux Hackerman: yeah, unfortunately standards often do. for example DOCSIS is not a raw layer-2 Ethernet transport and as such, IP protocol matters
<leons>
supersandro2000: "it more plausible that IPv6 is an accident than intentional" what do you mean by that?
<supersandro2000>
LinuxHackerman: hoax! It is very important that my cable is IPv6, IPv4 and Cisco certified!
<LinuxHackerman>
Leon: sure, but that's still not on the level of the cable. That's just a protocol going across lots of layers, but in the end it's still a copper wire and _that_ part really doesn't care. It's just what you put at the ends of the cable that matters.
<leons>
supersandro2000: Did routers decide to just add 96 bits to IP address and change the EtherType as a surprise?
<supersandro2000>
leons: I just randomly guess that it happened by accident that IPv6 is so good here because internet normally sucks and Germany is a bit behind
<leons>
Linux Hackerman: sure, but when you're on that level you can encode literally anything on Glass/Copper/EM spectrum :)
<Raito_Bezarius>
supersandro2000: that might also be the work of a regulation body :p
<Raito_Bezarius>
in France, ARCEP pushes IPv6 for all operators through regulation
<supersandro2000>
for example companies still put copper in the ground for vectoring
<supersandro2000>
instead of fiber
<LinuxHackerman>
Leon: sure. Key point being you don't have to tear the walls open to install IPv6-capable cable, because _that_ would be a real nightmare.
<supersandro2000>
Raito_Bezarius: recent regulations are more on how we limit and surveillance the internet but might be
<leons>
supersandro2000: nothing like this happens by accident. We had a monopoly internet provider here in Germany for a long time, especially when large prefixes were delegated. With the uprising of cable internet through Unitymedia (now Vodafone monopoly), those couldn't get sufficient v4s for every customer
<Raito_Bezarius>
supersandro2000: in Germany?
<leons>
So it's really just a question of demand
<LinuxHackerman>
supersandro2000: oh yeah I absolutely hate them for that. I even have fibre into the basement here, but the last 10m of my internet connection are good old phone wire with DSL on it.
<Raito_Bezarius>
nevertheless, there is a EU IPv6 taskforce, so it's not all *that bad*
<Raito_Bezarius>
does not work with Google Translate :(
<supersandro2000>
Seehofer, a great bavarian politician, wants to make ID verification mandatory for mail and messenger provider
Baughn has joined #nixos-chat
<Raito_Bezarius>
Ha :/
<Raito_Bezarius>
Well, we get the same kind of stupidity here also
<Raito_Bezarius>
Also for porn sites :>
<supersandro2000>
something like: you want to chat on IRC? First send us your address, a picture of your ID card and then we get back to you in a few weeks
<leons>
But the terrorists!!!!!
<Raito_Bezarius>
help the police to be even more lazy
<Raito_Bezarius>
</sarcasm>
<supersandro2000>
just read the CCC statements about it
<supersandro2000>
I don't know if they have them in English but they tear the government regularly apart when they want to do new internet laws
<supersandro2000>
well there is an English page but most posts are not translated...
<supersandro2000>
but we had gems like 136 000 corona test results publicly available with personal data of the testers
<supersandro2000>
in government paid software which does not meet basic security requirements
<supersandro2000>
the government might do great things but especially lately it is over shadowed by such accidents in my mind
FRidh has quit [Quit: Konversation terminated!]
luxemboye has quit [Remote host closed the connection]
<samueldr>
sphalerite: in my experience using the he.net IPv6 tunnels made youtube basically unusable
<samueldr>
I wonder where the ~$150/mo for AWS hosting number comes from
<ekleog>
like at $nonprofit we have a ML server that already passed 1834 emails today that resulted in 6222 deliveries, that is a VM extracted from the hard drive of a dead old server, that runs on top of a several-year-old machine
<samueldr>
(I know, from zimba, but where does it come from there?)
<joepie91>
hosting on AWS seems like a bad idea
<ekleog>
I mean if that VM is worth more than $15/mo I don't know where the money is going
<joepie91>
ekleog: bandwidth, probably
<samueldr>
nickel and diming, the AWS way
<joepie91>
bandwidth is traditionally the way cloudyclouds screw you
<joepie91>
$50-$100/TB is typical, vs. $1-$5/TB in the rest of the hosting industry
<ekleog>
(oh and this VM's CPUs are like <5% used, and it consumes like 3G of its 6G of RAm, so there's quite a bit of headspace to host a webui there)
<ekleog>
it's pretty impressive how these numbers went up
<joepie91>
err, oops
<joepie91>
I meant to highlight samueldr :P
<ekleog>
i mean you did answer my question :p
<samueldr>
I didn't need an answer
<samueldr>
so you accidentally did that right
<ekleog>
\o/
<joepie91>
lol
<samueldr>
I really meant, how was the number computed, I kind of see how it can be $150/mo, but it'd have been nice to see the breakdown
<ekleog>
i wouldn't have expected a web ui + ML to cost as much as $150 in AWS TBH, even counting absurd egress prices
<samueldr>
"a web ui" is downplaying what this is
<joepie91>
right
<samueldr>
discourse is "architected" in a way where it's basically a big mess hard to package IIRC
<ekleog>
oh maybe if you actually run it with one AWS VM per discourse microservice it'd cost that much just because the smallest VMs would still be too small?
<samueldr>
so it wouldn't surprise me that, while it helps scaling at some levels, it's also "too big" for small scale setups
<ekleog>
(at $nonprofit we also had a self-hosted discourse running on a similarly old machine, though no one used it so we let it die and it never saw much traffic)
<samueldr>
I also don't know what the actual traffic is for the discourse instance
<samueldr>
instance*
<samueldr>
uh.... I wrote it right the first time...
<ekleog>
but you wrote it right the second time too so it's still right :3
<samueldr>
I felt my fingers fumble on the keys as I hastily pressed enter, so I guess I fumbled it right
<samueldr>
maybe the traffic volume is just non-trivial enough, too
<ekleog>
that's possible, but we're around 500 non-ML-mode notifications a day / 60 new messages a day IIUC zimbatm's message, so… probably not _that_ big?
<samueldr>
I was thinking read traffic
<ekleog>
(now if we count manpower we'd definitely dwarve the resource cost so the whole question is a bit moot, but still it's interesting :))
<ekleog>
hmm but then people would go there manually?
<ekleog>
you mean there are people who go on websites without having received a notification? 😱
<samueldr>
yes
<Raito_Bezarius>
sphalerite: samueldr: I actually took the route, fixing locally Internet
<Raito_Bezarius>
and route the packets using HE or the operator's native IP depending on the target
<Raito_Bezarius>
that works well more or less fine after many hours reading about what the fuck are VRFs and how they work in the linux kernel
<Raito_Bezarius>
but the biggest annoying thing atm is that this operator, who at some time offered the same for IPv4 (and has ceased since the move to IPv6), do not offer any reverse DNS zone delegation, nor any way to configure it :(
<Raito_Bezarius>
on their issue tracker, the issue is open for 8 years :/
<Raito_Bezarius>
since 8 years*
julm has joined #nixos-chat
cole-h has joined #nixos-chat
__monty__ has quit [Quit: leaving]
<veleiro>
Q: what was the first mainstream language that used a package manager?
<veleiro>
it seemed to me to start being a trendy thing with js/ruby/python
<leah2>
ruby gems?
<gchristensen>
CPAN became a thing in 1993
<leah2>
gems is 2003, npm is 2010
<leah2>
cpan had an installer, but not really a manager?