* samueldr
opened the fonts menu in inkscape, having forgotten to clean the fonts addition^W folder for yet another time
<samueldr>
guess I didn't need to use that computer for a couple of minutes anyway
<gchristensen>
ugh yeah that menu ...
hexa- has quit [Quit: WeeChat 2.9]
hexa- has joined #nixos-chat
<DigitalKiwi>
seeing as i just got back from one system lock up (wasn't even doing anything this time. think it was chromium. not sure why the oom killers take so long/don't work :()
<DigitalKiwi>
seeing as... what's inkscape menu?
<hexa->
DigitalKiwi: earlyoom :)
<DigitalKiwi>
yeah that one
<hexa->
huh?
<hexa->
the actual "oom killer" is a kernel task
<DigitalKiwi>
it's enabled, doesn't do anything
<hexa->
odd
<DigitalKiwi>
doesn't do anything and/or errors when it does try to
<pie_>
DigitalKiwi: im starting to think we might need to enable a better oomkiller by default in nixos...
<hexa->
systemd-oomd will be there with 248
<pie_>
i never looked into it but i didnt even know theres a built in oom killer
<DigitalKiwi>
systemd processes are the ones that coredump half the time i think :(
<pie_>
and why doesnt it just work
<hexa->
DigitalKiwi: check coredumpctl
<pie_>
why does there need to be a systemd for it
<hexa->
oom killer is supposed to work better with cgroupsv2
<DigitalKiwi>
Sun 2021-04-25 05:28:10 UTC 209207 1000 100 11 present /nix/store/qbmgrdfd3mwryk2hgjiq5hrg0g424pzl-chromium-unwrapped-89.0.4389.114/libexec/ch>
<hexa->
this is a good read on the whole memory/swap/cgroups concept
<DigitalKiwi>
i've tried having swap not havign swap having zram not having zram :(
<samueldr>
DigitalKiwi: the fonts dropdown
<samueldr>
it renders all fonts
<samueldr>
I have too many fonts
<DigitalKiwi>
are you tom 7
<samueldr>
I wish I were
<DigitalKiwi>
would you rather be tom 7 or be friends with tom 7
<samueldr>
neither, I'd prefer be given the chance to be friend, rather then being forced to be
<samueldr>
I don't know, maybe he's a jerk! (unlikely I guess)
<DigitalKiwi>
or option c be parasocial besties
<DigitalKiwi>
(i've only seen one of his videos so lol but) if i were tom 7 i don't think i would enjoy tom 7's video as much as i do by not being tom 7 so not being him > being him
h0m1 has quit [Ping timeout: 260 seconds]
<samueldr>
watched a couple, not sure if I watched them all, at least I watched all their "talks as a video form", and making the videos seem fun!
<DigitalKiwi>
http://tom7.org/lowercase/ it's this one i think in the first few minutes he says he has too many fonts....all of them
<pie_>
did someone finally make a generic STUN/whatever its called tool
<pie_>
is the Global Socket Relay Network "the internet"?
<pie_>
looks cool tho
<pie_>
"Abandon the thought of IP Addresses and Port Numbers. Instead start thinking that two programs should be able to communicate with each other as long as they know the same secret (rather than each other's IP Address and Port Number)." very good.
<aaronjanse>
That very slightly reminds me of cjdns, where your ipv6 address is your public key
<aaronjanse>
Or a fingerprint of your public key
<aaronjanse>
Ofc very different
<pie_>
people and their post-exploitation toolkits :(
<pie_>
* :)
<hodapp>
I always mix up cjdns and cdnjs in my head
<colemickens>
seems like tailscale could copy that feature very easily and to great effect
<colemickens>
not sure if they want to though, there's not really a convert/funnel that way
abathur has quit [Quit: abathur]
<edrex>
tell me why tailscale and not just wireguard?
<edrex>
it seems like folks really like tailscale's product. i know some good people are working there.
<edrex>
they provide some relays for when you're all symmetric NAT i imagine, and they have a nice UX on mobile/desktop?
<edrex>
i mean for hacker-in-garage. for orgs it makes sense to pay someone for support
<colemickens>
if you set it up you'll understand
<colemickens>
it takes less than 5 minutes to on board literally every device in my life
<colemickens>
nat punching works, always, no thoughts
<colemickens>
I never open ports to my dev VM in the cloud anymore, tailscale's got my back
<colemickens>
I will switch to an open source alternative (headscale, innernet, something) but it's about a billion times nicer than managing that stuff by hand (which I did for a year or two)
<colemickens>
(in this context, tailscale already has the "Global Socket Relay Network" in their backbone/NAT-punching offering, folks in this space might already have it installed, they have all of the pieces there, etc, but not sure ephemeral connections fit with their product, might be confusing/muddling security concerns etc)
<colemickens>
`innernet` is neat, it's a rust-y alternative (that supports kernel WG unlike tailscale [yes, I notice])
<DigitalKiwi>
wonder how many less hard drives we would need if everything was in the same cloud
<samueldr>
I guess you mean "with dedup, and no compression"?
<colemickens>
and its api server hides behind wireguard, so the only port open in the entire network is the "api servers" wireguard port. so if you trust WG security, you can feel somewhat okay about its apiserver security. they onboard clients with a one-time-use WG key and then rotate
<DigitalKiwi>
dedup yeah
<colemickens>
encrypt yo data
<colemickens>
what dedup
<samueldr>
DigitalKiwi: I assume a lot of "linux isos" space would be reclaimed
<DigitalKiwi>
just imagine if instead of every company storying multiple copies of everything they know about you they all just had one!
<samueldr>
interesting thought idea indeed
<samueldr>
uh, thought experiment
addcninblue has quit [Ping timeout: 265 seconds]
<samueldr>
I was distracted, trying to track down a mostly black micro sd card on my black desk
<samueldr>
among black cables and black accessories
<DigitalKiwi>
they could even then sell you* an access key and you* could use the data in an actually useful way (*)anyone who wants it
<DigitalKiwi>
turn off the lights get someone to take a picture and you got yourself a jigsaw puzzle scene
<samueldr>
I found it in the end
<DigitalKiwi>
the white album, polar bear in blizzard, black cat at night, samuel in the home lab
<colemickens>
I'm now in search of nixos users who may've done things like poke extra entries into grub that chainload ISOs from the nix store? someone must have done this before.
<pie_>
colemickens: "doesnt sound that hard"? x)
<DigitalKiwi>
colemickens: i might be wrong but i think zfs has dedup even encrypted
<colemickens>
pie_: my thinking too, we'll see if I'm happy in a half hour or sad in three
<DigitalKiwi>
both
<pie_>
colemickens: i recommend building vms to test this kind of stuff
<colemickens>
DigitalKiwi: I hope I'm not just ruining a joke, but the hope is that customer's are encrypting their data. (presumably with keys they aren't sharing)
<DigitalKiwi>
in a half hour you will have found someone and in 3 hours realize it doen't work anymore
<colemickens>
hence there would be no de-dupe
<samueldr>
I just want to say that gnome-disks is an amazing tool all around
<samueldr>
not the most in-depth, but just enough and a bit more to be really useful
<samueldr>
mounting disks, but also "attaching" (mounting) disk images
<pie_>
now combine it with gparted
<samueldr>
pie_: gnome-disks does partitions too!
<samueldr>
but not in-depth
<pie_>
good to know, ive always been annoyed by having to do it by hand <samueldr> mounting disks, but also "attaching" (mounting) disk images
<pie_>
we need more tui tools
<samueldr>
more interactively driven interfaces, yes
<samueldr>
it's so much more convenient than doing it by hand, for mounting disk images
<samueldr>
or even mounting partitions on your usb or sd card!
<DigitalKiwi>
colemickens: think like ipfs/nix paths w/e that are content addressable and how many millions of copies everyone has of everything
<colemickens>
sure, I like CASy things :)
<colemickens>
but again, short of it being public, it's not really shareable?
<DigitalKiwi>
but even like the encrypted user data like if i have the same file on dropbox and google cloud and nextcloud and really it's all just a download from wherever pdfs are acquired lol
<DigitalKiwi>
veered off there; but i was getting at is i could use the same encryption key multiple places possibly
<colemickens>
I guess I'm not sure what we're getting at, it's sort of all just moving cheese a bit
tomberek has quit [Quit: Connection closed]
<DigitalKiwi>
one giant wheel of cheese and everyone can use it
<DigitalKiwi>
keyword can; as they say in the stonks options "the right but not the obligation"
<DigitalKiwi>
one example; think of how many dozens of copies of an image posted to social media get stored by how many hundreds of computers on all of the sites that it gets reposted to and then cached as it spreads out on all of the cdns and isps
<samueldr>
oh, don't worry
<DigitalKiwi>
..and then resized and screen shotted and no longer the same file ;_;
<samueldr>
they don't store a copy
<samueldr>
exactly
<samueldr>
it'll also get recompressed if the image is used outright
<pie_>
can two nix daemons run concurrently safely?
<DigitalKiwi>
it seems you are uploading an already efficient and high quality png; would you like this shit jpeg instead? of course you do
<pie_>
sqlite dbs are generally locked for concurrency afaik?
<DigitalKiwi>
i suspect it could dead lock
<samueldr>
pie_: wrong channel to ask :) I suggest #nixos-dev
<samueldr>
unless you were shitposting ;)
<pie_>
am i ever not? :(
<samueldr>
hard to tell!
<pie_>
this was the shitpost <pie_> am i ever not? :(
<samueldr>
hard to tell!
<DigitalKiwi>
i'm not certain it's happened to be before but i've definitely had to kill nix-daemons after giving up waiting
<pie_>
xD
<samueldr>
;)
<pie_>
im still vaguely pondering if i can use the unstable daemon without upgrading my machine...
<DigitalKiwi>
i had about a week on irc after switching to flakes of being disappointed with all of the things i can't do with nix profile that i could with nix-env and not being able to use nix-env and a few other annoyances i wrote bash aliases for lol
<edrex>
colemickens: i'll have to give it a spin, thanks
<colemickens>
edrex: oh I forgot, sharing feature is killer
<cole-h>
aaronjanse: Thanks, seems obvious in retrospect :P
<colemickens>
edrex: I can send you a link, you click, all of the sudden you have access to one of my peers
<aaronjanse>
No problem cole-h
<edrex>
oh i do like sharing things, and it is a pita
<colemickens>
I can revoke whenever, etc.
<edrex>
how would you go about just sharing one service on a peer?
<colemickens>
I think it might support that? They have some feature where it can know about individual services and then they do have some sort of acl/firewalling. I don't use any of that though (specifically because I do not want to stay on tailscale long-term)
<colemickens>
in fact, I'm going to play around with an 'innernet' module right now
<edrex>
oh i see, they have an ACLs thing
* edrex
looks up innernet
<edrex>
i like things that are half-baked and difficult
luxemboye has quit [Remote host closed the connection]
luxemboye has joined #nixos-chat
<pie_>
oof
<abathur>
anyone aware of a tool that can do (or term that describes?) something like "take a list of directories as input and output a directory that is roughly their union, concatenating all identically-named files into a single output file of the same name"?
<Ke>
concatenating the files is something I have never seen
<Ke>
rsync can add files by default
<Ke>
but it overwrites, not concatenates
<abathur>
yeah
<abathur>
I assume it's a niche need
<abathur>
the best metaphor I have atm is akin to a collation setting on a printer/copier, but there are other unrelated uses of collation so I suspect that won't be fruitful
<Ke>
I guess maybe rsync with some no overwrite and remove source might work for the start
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-chat
midchildan_ has quit [Read error: Connection reset by peer]
steveeJ has quit [Read error: Connection reset by peer]
raboof has quit [Read error: Connection reset by peer]
davidtwco has quit [Read error: Connection reset by peer]
davidtwco has joined #nixos-chat
midchildan_ has joined #nixos-chat
raboof has joined #nixos-chat
steveeJ has joined #nixos-chat
<thibm>
"Clang 7 seems pretty old"
* thibm
is digging into old nixpkgs checkouts to find old clang versions (3.4, 3.1) to understand why my code does not compile in this proprietary tool that uses such old compilers…
<thibm>
BTW, it would be awesome to add git tags in a local nixpkgs git repo where hydra tried to build packages. I'm using a 2 year old checkout, I don't care about the precise commit as long as I can get the most binary substitutes for the packages I want
<thibm>
Is there a way to achieve that? (Through git tags or any other method…)
<eyJhb>
Someone here told me to never use w3schools for reference, not sure who.. But I guess having this on a webpage hurts then `<link rel="stylesheet" href="https://www.w3schools.com/w3css/4/w3.css">`
<__monty__>
Maybe that's available from w3c themselves?
<eyJhb>
No clue, I think they guy who is currently using this, just viewed the source of a playground, and then copied the link from there
tomberek has quit [Quit: Connection closed]
<matthewcroughan>
At work, I asked why we can't build the older version of our distribution.
<matthewcroughan>
Something about a gitRepo manifest change.
<matthewcroughan>
"Use the latest" they said.
<matthewcroughan>
"Don't go back" they said.
<matthewcroughan>
"You should never use old software" they said.
<Guest80259>
Those who cannot reproduce their history are doomed to repeat it. :>
<__monty__>
TIL bird collisions *do* happen.
<MichaelRaskin>
You mean bird-bird collisions?
<__monty__>
Yeah.
<eyJhb>
Video evidence __monty__ ?
<__monty__>
Surprisingly birds falling from the sky happens to fast for me to go inside and get an ipad to record it on or something.
<eyJhb>
What about a mobile?
<eyJhb>
:D
<eyJhb>
I don't require the videos to be recorded on a iPad
<__monty__>
Don't think what my T303 records can respectfully be called video.
<__monty__>
And I don't carry it around anyways.
<__monty__>
Ipad is not my best option for video. Just my only one.
<eyJhb>
Sony Ericsson T303?
<__monty__>
Yep.
<__monty__>
Been rocking that bad boy for over a decade.
<eyJhb>
Did you go back, or did you just never upgrade?
<eyJhb>
Three years more, and my OPO will be a decade as well :p
<__monty__>
That's still my first phone. Hasn't quite stopped functioning so I have no reason to get a new piece of e-waste.
<eyJhb>
I've had a couple, but most used. Had a w660i. Loved that phone.
<infinisil>
A device that could record the last ~10 seconds on a button press of your vision + sound would be revolutionary
<infinisil>
No longer would people have to hope that they are recording just as something amazing happens
<infinisil>
We'd get a ton more footage of cool things
<__monty__>
Ubiquitous recording has major privacy implications though.
<__monty__>
I mean, it already does. This'd make it even worse.
<samueldr>
we'd get the footage of the end of something maybe cool happening
<eyJhb>
I think samueldr is on point here, that's what is important!
<eyJhb>
Also, offline?
<eyJhb>
And not contacting remote servers? :D
<eyJhb>
Take Google Glasses, cut out everything Google + Antennas, put a USB-C connector, and then make a rolling recorder
<eyJhb>
infinisil: I want it.
<infinisil>
samueldr: I was thinking realistically it would be hard to have such a rolling recorder for longer events
<infinisil>
But yeah, ideally have it be as long as possible
<infinisil>
eyJhb: I had in mind like some futuristic device that could hook into our brains directly :P
<eyJhb>
Ehmm.. Have you seen the short film about this concept??
<samueldr>
dashcams can go for multiple minutes on a rolling recording, or even record on the media "forever" in rolling mode
<eyJhb>
Somewhat this concept.
<eyJhb>
Seeing if I can find it, but this network SUCKS... Or rather, the WiFi does.
<eyJhb>
Can't find it :/
<__monty__>
I wonder what TARGET_OS_WIN32 is for in Apple's Libsystem.
<samueldr>
itunes?
<__monty__>
Things like iTunes on Windows? But then why no WIN64?
<samueldr>
probably because win32 is synonym with "not 3.11"
<__monty__>
TARGET_CPU_* includes MIPS, Sparc and Dec Alpha. Don't think they ever produced software for such architectures, right? Wonder if that shows what they've experimented with 🤔
<philipp[m]>
"All 32 users of our windows stuff"?
<samueldr>
maybe as NeXT?
<philipp[m]>
How is it nowadays with windows? Are most applications finally 64bit?
<abathur>
repeat from last night: anyone aware of a single tool that does something like "given a list of directories as input, output a directory that is roughly their union--with all identically-named files concatenated into a single output file of the same name"? (or know of any term for this operation?)
<philipp[m]>
Sounds like something I'd try to do in a few lines of bash.
<abathur>
(not particularly worried about finding a way to script it)
<__monty__>
samueldr: Hmm, Next might explain Sparc, yeah. Can't find references to NextStep on Dec Alpha or MIPS though.
<philipp[m]>
I mean, of course it didn't help with a setup that contains X.
vulpine is now known as Fuechschen
<gchristensen>
is it janky as heck to host a public webservice for people to run integration tests against, instead of figuring out how to make CI start up the service?
<samueldr>
yes?
<gchristensen>
mehh
<gchristensen>
any good Go devs 'round? I have some questions about good practice
<lukegb>
define "good" :P
<abathur>
Dear Santa,
<hodapp>
I write Go but whether it's 'good practice' is up for debate
<lukegb>
in the Santa context I'd probably put myself in the "nice" bucket rather than the "naughty" bucket
<gchristensen>
familiar with Go enough to be aware of what normal looks like :)
<hodapp>
go ahead and ask
<gchristensen>
I want to call an API which publishes an OpenAPI 3.0 spec from go, I found https://github.com/deepmap/oapi-codegen#overview and it looks pretty good, but it feels weird to need to `go get` this thing every time I want to regenerate the spec, is that normal? is there a more normal way to integrate this?
<hodapp>
if the 'go get' is just to get the same oapi-codegen binary every time, why would you need to re-run it every time?
<gchristensen>
well maybe I'm overthinking it, but I was imagining doing this in CI for example
<lukegb>
Presumably you could grab it from nixpkgs instead
<gchristensen>
but this is totally normal, then?
<gchristensen>
not some sort of, like, "dev dependencies" thing that might track this
<gchristensen>
I'm not stressed about it, just curious what normal is
<lukegb>
I don't think Go usually tracks it; it has the //go:generate stuff but it's kinda... poor, and requires that you have the tools on your PATH anyway
<hodapp>
'go get' was the standard way to do all Go dependencies prior to Go modules in Go 1.11 and onward
<lukegb>
I suspect usually people write Makefiles to generate these things...
<lukegb>
I've seen a few projects with Makefiles to go get all the dev-only dependencies for doing things like codegen
<gchristensen>
that seems right
<gchristensen>
cool, thanks :)
jdnixx-M has quit [Quit: Bridge terminating on SIGTERM]
ky0ko1 has quit [Quit: Bridge terminating on SIGTERM]
ky0ko has joined #nixos-chat
rj has joined #nixos-chat
jdnixx-M has joined #nixos-chat
endformationage has joined #nixos-chat
rj has quit [Ping timeout: 240 seconds]
waleee-cl has joined #nixos-chat
rj has joined #nixos-chat
tomberek has joined #nixos-chat
abathur has quit [Quit: abathur]
abstrn has quit [Ping timeout: 240 seconds]
cole-h has joined #nixos-chat
abstrn has joined #nixos-chat
<__monty__>
Is there a way to figure out which header file is getting included, i.e., the absolute path?
rj has quit [Ping timeout: 240 seconds]
rj has joined #nixos-chat
<pie_>
__monty__: for what
<pie_>
c/++?
* pie_
tries to remember
<pie_>
__monty__: i think the usual debugging flags are something like -v -M and -Mb or something like that
<pie_>
i think those might have been for defines though
<pie_>
actually hold on its coming back to me, yea i think
<Synthetica>
Really useful for when you just quickly want to hack on something
<aaronjanse>
Yeah
addcninblue has joined #nixos-chat
<mjlbach>
At that point I'd rather be in a shell for incremental compilation
<Synthetica>
If you want to build something big or more than a few times, yeah, but this way you don't even have to know what build system they're using
Jackneill has quit [Read error: Connection reset by peer]
<samueldr>
>> (Vote : Pour 120, Contre 0, Abstention 0)
<samueldr>
a "right to repair" bill in Québec has been voted to be adopted
<Guest80259>
Nice!
<samueldr>
hi Guest80259
<samueldr>
probably want to re-assume your normal identity
<Guest80259>
<cynicism informedness="low">because availability of tech stuff wasn't bad enough there already?</cynicism>
<samueldr>
oh, you're a hacker man, that's why
<samueldr>
yeah
Guest80259 is now known as LinuxHackerman
<samueldr>
that'll give the "don't sell in Québec" checkbox that must already be present in national merchants a run for its money
<hexa->
LinuxHackerman: i commented on your routing gist
<samueldr>
it must already be present because it's a normal occurence to see few products unavailable
LinuxHackerman is now known as Guest80427
<samueldr>
heh
<Guest80427>
And thank you for the info, even if it came out as "hi Linux Hackerman" on the other side of the bridge. UX 10/10
<samueldr>
great success!
<Guest80427>
hexa-: thanks!
<samueldr>
Guest80427: you're now Guest_80427 without underscore
<samueldr>
great success was sarcastic
<samueldr>
the proposed amendments are not the best they could be, but at the very least it's a bit more than the bare minimum
Guest80427 has joined #nixos-chat
Guest80427 has quit [Changing host]
<samueldr>
forcing standardized testing of equipment, and notes according to those standardized tests to be advertised with the product
<samueldr>
now I need to lobby for software right to control to be part of those criteriæ!
<Guest80427>
hexa-: uuuuh… why would the firewall, which when unconfigured on nixos allows forwarding everything, affect this? I mean, it seems fairly clear to me that it does, but that confuses me a great deal
<Guest80427>
And especially in such an unreliable way
<samueldr>
Guest80427: your quit/reconnect didn't help with your nickname :)
<__monty__>
I get the same output as for -E and it's just -Wundef-prefix errors.
<__monty__>
And the unused argument error about -dI.
<__monty__>
No, not error, warning.
<pie_>
__monty__: for th eoneliner as well?
<__monty__>
Actually, yes. That's a little unexpected.
<pie_>
ok so presumably something is fucky with your env
<pie_>
try --pure
<pie_>
if it still broken try clearing your NIX_PATH or do whatever to make sure no overlays are applied (I dont know how that works actually and if clearing the NIX_PATH will fix that)
<pie_>
goal is to get to a working state and then figure out whats different on your system
<__monty__>
Same result. But maybe that's because of the sandboxing situation on darwin. Not sure whether that affects --pure?
<pie_>
oh
<pie_>
doh
<pie_>
youre on mac
<pie_>
who knows whats going on there
<pie_>
make sure its running the right clang?
<pie_>
clear PATH maybe?
<pie_>
even then this should apply for all clangs so who knows whats going on
<__monty__>
The clang on my PATH is a different version.
<pie_>
does the other version have it
<pie_>
did you check is its listed in th e--help?
<pie_>
you did check the version earlier and said tha tit has it, did you actuall check it by running the exeutable or by "knowin" what version youre supposed to be using
<pie_>
or do i remember wrong?
<__monty__>
The clang on my PATH does give a lot more output on that oneliner but the warning about -dI being unused is still present.
<__monty__>
Both --helps have it.
<pie_>
hmhmhm
<pie_>
ok surely knowing this is progress somehow
* pie_
ponders
<pie_>
__monty__: ok so what does `which clang` give you, and can you get into a clean shell? does --pure work on mac?
<pie_>
well really, i guess `which clang` wont tell me much. i guess the next thing id do is check if its the clang wrapper script, and check how i enable debugging the wrapper script
<pie_>
IIRC something like NIX_DEBUG=1 or whatever, and then stare at the output and see if its passin the argument wrong or something
<pie_>
that should (?) also give you a commandline that you can call manually
<pie_>
you can then try minimimizing hat
<pie_>
that
<pie_>
reasonable?
<__monty__>
Surely the wrapper wouldn't warn about unused arguments? So I think the argument is getting to clang. Also the one on my PATH isn't wrapped and gives the same warning.
<__monty__>
So there's something about -dI on a mac.
<__monty__>
Or on my mac at least.
<pie_>
__monty__: yes but it may be ending up in the wrong place or idk
<pie_>
im just trying to reduce to a source or osmething
<pie_>
is there an attr for getting an unwrapped clang
<__monty__>
The clang on my PATH isn't wrapped. So the argument isn't in the wrong place, given it works for you.
<pie_>
clang.cc i guess
<pie_>
__monty__: ok sure but where does the one on the path come from
<pie_>
in the secod one i added an a to the heaader name
<pie_>
when i try them both, the first works and the second gives the clang-7: warning: argument unused during compilation: '-dI' [-Wunused-command-line-argument]
<__monty__>
The problem is there's multiple headers with the same name and I need to find out which one is included.
<__monty__>
I don't get an error about the header being missing. I get errors that seem to indicate the correct header isn't being included.
<pie_>
all im saying is it looks like errors before something orother prevent you from using -dI
<pie_>
or at least _that also_ yields that message
<pie_>
__monty__: can you paste your full error
<pie_>
__monty__: another way to do this is get the includes path and order from clang -c and manually search for the file names
<pie_>
look for the stuff like `#include <...> search starts here:` in the clang -v output
<red[evilred]>
not sopacex - but looiks like ULA is good for launch shortly
<red[evilred]>
,spacex
<red[evilred]>
however you trigger the notification - I fogot
<__monty__>
,launch
<{^_^}>
Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson
<pie_>
oh f*** looks like it still chokes too early with -H
<pie_>
well, thats if i tell it to try to load an include that doesnt exist, maybe it works for you
<__monty__>
pie_: I'd tried that flag earlier and it was useless. BUT now it says something about include depth so I think that confirms my suspicion there's a circular include.
<pie_>
can you paste it
<infinisil>
__monty__: pie_: This sounds offtopic for -chat tbh
<__monty__>
We can move. #nixos, pie_?
<pie_>
ok lets go to main infinisil wants to play with the rockets :P
<__monty__>
Fwiw, it started out with a short and, I assumed commonplace, question.
<__monty__>
Unrelated to nix that is.
<pie_>
its never that simple
<samueldr>
yeah, you're not in trouble... this time ;
<samueldr>
;)*
<pie_>
also didnt know its not a nix question xD
kraem[m] has joined #nixos-chat
<infinisil>
I think -chat in general shouldn't be for troubleshooting, imo at least
<pie_>
im a repeat offender
<__monty__>
I wasn't expecting anyone to go as far as helping me troubleshoot this, fwiw.
<gchristensen>
+1 off topic
<infinisil>
/invite pie_ #irc-jail
<pie_>
bonk
<joepie91>
no that's other jail
<pie_>
do not pass go
<pie_>
do not collect 200 $
<joepie91>
yes that one
<infinisil>
Damn, all those "Go"s
<gchristensen>
a little legacy compared to spacex, but pretty thrilling
<infinisil>
I love it
<infinisil>
Shows how many people and teams are working together
<{^_^}>
T-2:00! https://www.youtube.com/watch?v=4C_HRGrtBjg: Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson
<pie_>
who needs pxe when you can boot over dns
<infinisil>
Launch!
<Taneb>
!!!
<aaronjanse>
How can I add myself to get pinged by this bot?
<aaronjanse>
,plz ping me
<gchristensen>
,launch = Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson aaronjanse
<{^_^}>
launch redefined, was defined as Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson
<pie_>
i dont know the appropriate terminology, there was some talk some days/ weeks ago in here about kexec or something involving efi? id like to re/boot into the uefi firmware or something instead of a norma full reboot, is there something for that?
<pie_>
anothe rthing would be that im running a pxe installer to install a second system and bootloader and some partitions,and it would be nice for the moment if i can avoid nuking the existing bootloader, can i boot into the second grub somehow from the running system?