<Profpatsch>
Ericson2314: what is this drv language you speak of?
<Ericson2314>
Profpatsch: drv files, at first glance. but there's a lot of richness and subtlety going on here, I've been learning recently
<Profpatsch>
Ericson2314: so the semantics of the drv data structure?
<Ericson2314>
yeah
<Profpatsch>
It looked more-or-less straightforward to me, it maps onto processes and the environment
<Ericson2314>
Well the exact way we hash things---maybe that's more beaurocratic than deep
<Ericson2314>
and then content addressed derivations + ret-cont
<Profpatsch>
and you got stuff like passing as file and structuredattrs
<Profpatsch>
Ericson2314: That’s not current semantics though
<Ericson2314>
yeah, the current stuff I've been learning is the hashing subtleties
<Profpatsch>
But I agree that it’s a complication that should be well-specified
<Ericson2314>
exactly how hashDerivationModulo works, remembering old drvs in a funny way
<Profpatsch>
Sounds like a good place for a post-factual (formal?) specification
<Ericson2314>
I am thinking we may use CA derivations as an oppertuniy to clean up some old mistakes, but also we can clever lawyer our way around old mistakes even before that to get remote builders that don't need to trust the client (!)
<Ericson2314>
(for the first part, use special case CA drvs to clean up things unrelated to the content addressing of the outputs)
<Profpatsch>
I don’t think anybody has even started to get a good grasp on the trust model.
<Profpatsch>
And we should fully separate the trust model from the CA improvements, they are completely orthogonal.
<Ericson2314>
Profpatsch: So to be clear I am just talking about the trustworthiness of not-content-addressed builds, and not about secrets in the store or something
<Profpatsch>
To be specific: We can have full content addressed builds without adding a trust model.
<Ericson2314>
absolutely
<Profpatsch>
And we *should* keep those two discussions separate, CA is already way too big as is
<Ericson2314>
Well, I've made a bunch of little independent PRs so far
<Profpatsch>
Even reviewing the RFC by regnat is like a full day of work.
<Ericson2314>
I think the lawyering around the old system stuff untrusted remotes is a great entry point
<Profpatsch>
And then you have to keep up with all updates.
<Ericson2314>
as it doesn't change the data model at all
<Ericson2314>
Yeah I think both RFCs bighting off more at once than we need to
<{^_^}>
nix#3446 (by Ericson2314, 4 days ago, open): Big Derivation refactor
<Ericson2314>
(I need to get it building, yes, but I have some other smaller PRs which clean up more basic tech debt it would be good to land first to make that less complex)
<Ericson2314>
the "big derivation refactor" should give us the vocabulary we need to discuss the changes----hashDerivationModulo was truely a lot of things at once
globin has joined #nixos-dev
justanotheruser has quit [Ping timeout: 272 seconds]
<colemickens>
new Azure bring up instructions have been added to nixpkgs
<colemickens>
We don't really have up-to-date Azure official blogs
<colemickens>
blobs*
<colemickens>
Should we remove this section entirely, or mention the unofficial support and direct them to nixpkgs/nixos/maintainers/scripts/azure-new ?
<garbas>
colemickens: that would be great. could you get your suggestion in a PR?
<colemickens>
I'm preparing it :) Just not sure which option to go with.
<garbas>
colemickens: or we try to upload latest image to azure
<colemickens>
I wrote a nice script that makes it easy for anyone to, but I don't think we can provide the level of support necessary to try to act like we have an official Azure image.
<colemickens>
I think a small mention that unofficial support is documented at <blah blah link> will be fine.
<colemickens>
I'll send a pr shortly
<garbas>
and make it part of the release process, i think worldofpeace was also asking me about this
<garbas>
i hope next week i'll find a time to actually restructure the download page.
<garbas>
colemickens: what would the level of support needed to provide an official azure image?
<colemickens>
imo, we need to cleanup the work I just did so that images have good names that match their nixos labels. we'd need to clean out the existing azure-images.nix (and the other one?) that both point to bad/old images.
<colemickens>
there's more restructing of the azure module that would make sense, I don't really want to support an image with the Azure agent, but people might also expect it.
<colemickens>
then I guess a CI job that just runs the two scripts and confirms it can 'ssh' into the VM and do a nix sanity check.
<colemickens>
the default Azure module doesn't enable password-less sudo, but changing that feels like a breaking change. (my default basic image does enable it, however, since it's a new addition).
<colemickens>
sorry, thoughts are really jumbled. maybe I'll open a meta issue on nixpkgs too and we can discuss. I know there are a few other Azure users.
<garbas>
that would actually be the best way forward. having a support for multiple clouds is really going into direction of being more beginner friendly
<garbas>
eg. click here and play with nixos
<garbas>
like we already have for ami
<colemickens>
I have another version, maybe still in git somewhere, that also includes scripts for promoting the uploaded image to "Shared Image Gallery" image, and then that can be made semi-public. Or can be replicated-from.
<colemickens>
But that was a whole other set of minefields, so that got put on the back burner.
<colemickens>
Otherwise, if one wants to make it *really* easy, I think there is a certain amount of hoop-jumping that one has to go through to be as easy as booting an Ubuntu/RedHat/Win VM.
<colemickens>
Anyway, as of today it's as easy as "nix-shell, az login, ./upload-image.sh, ./boot-image.sh" so I'm happy about that.
<colemickens>
Hm, you're involved in the marketing efforts right? You probably have more perspective than I do about the perceived value of an "official image" versus just making it easy for anyone to build their own.
<garbas>
i think for start this is fine. but we need to polish this
<colemickens>
Buildng / uploading can be tough. CPU, kvm, bandwidth.
<colemickens>
Alright, I'll spend more time on this, I'm convinced and jonringer got my first bit merged quickly, so I'm hopeful.
<garbas>
that would be really great. having a really polished support for ami, gcp and az. that is 90% of the marker
<gchristensen>
+1
* garbas
just threw a random high number but should instead you the work "majority" :P
<colemickens>
How is testing/building/owning of AMIs funded?
<colemickens>
Does Amazon give a free account for FOSS? Does someone with authority/notoriety send an email?
<colemickens>
I uh, might be able to ask, and have some credits I can use for development, but ideally this would be a hands-off NixOS org subscription we publish under. (Just thinking of things that might take a while that we should start thinking about as well.)
<gchristensen>
we have an AWS account with a credit card attached
<colemickens>
I think the biggest issue is - the Azure agent is siginificantly out of date. It's also an abomination, and I'm allowed to say that.
<colemickens>
I have gained some python skills, so I'll try to upgrade it again, but I was pretty confused last time I tried to upgrade it.
<drakonis>
this whole discussion was fairly good.
lovesegfault has joined #nixos-dev
teto has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-dev
<emily>
Ericson2314: out of curiosity, what do you find broken about cue? I like a lot of things about it and there are things I could think to point to as warts but I'd be interested in hearing others' criticisms
<Ericson2314>
emily: basically they know their logic programming but not type theory or category theory. Combining "a is a b" and "a <= b" into the same partial order is sketchy and weird to me. Separate mutually recursive lattices seems better
<Ericson2314>
it's like a category of lattices is a 2 cat, and they've smashed it back down into a 1 cat in a weird way
justanotheruser has joined #nixos-dev
<simpson>
Huh, interesting observation.
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
<emily>
Ericson2314: right, the conflation with the more type-y stuff and arbitrary constraints seemed potentially unfortunate (for one the subset of the constraints you can check statically is pretty ill-defined)
<Ericson2314>
indeed!
<Ericson2314>
lack of static checks is good canary in the coal mine
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
drakonis has quit [Quit: WeeChat 2.7.1]
bhipple has quit [Remote host closed the connection]
lovesegfault has quit [Ping timeout: 252 seconds]
FRidh has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
FRidh has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos-dev
MichaelRaskin has quit [Ping timeout: 264 seconds]
orivej has quit [Ping timeout: 264 seconds]
__monty__ has joined #nixos-dev
<qyliss>
How do I fix a hackage package that is incorrectly marked as broken? The file it's defined in looks generated, so wondering if there's something special to do.
<Taneb>
qyliss: edit configuration-hackage2nix.yaml and then wait until Friday evening
<Taneb>
(when peti does the updating)
_e has joined #nixos-dev
globin has quit [Quit: o/]
globin has joined #nixos-dev
<srk>
qyliss: are you sure it isn't broken (with ghc883)?
globin_ has joined #nixos-dev
globin has quit [Ping timeout: 252 seconds]
phreedom has quit [Remote host closed the connection]
<domenkozar[m]>
pulseaudio always brings a bit of windows feel to linux
<pie_[bnc]>
gotta love accumulating errors
<pie_[bnc]>
having kids is rebooting
<flokli>
gchristensen: shouldn't these errors go to stderr, and the json to stdout?
<gchristensen>
flokli: nix-env's output goes to stdout
<domenkozar[m]>
huh did X stop respecting .xsession?
<gchristensen>
the json dump should go to a named file or a third pipe or something
globin has joined #nixos-dev
nix-build has joined #nixos-dev
dottedmag_ has joined #nixos-dev
dottedmag_ has joined #nixos-dev
dottedmag_ has quit [Changing host]
teozkr has quit [Ping timeout: 240 seconds]
{^_^} has quit [Read error: Connection reset by peer]
dottedmag has quit [Quit: QUIT]
bridge[evilred] has quit [Ping timeout: 240 seconds]
abbradar[m] has quit [Ping timeout: 240 seconds]
timokau[m] has quit [Ping timeout: 240 seconds]
yorick has quit [Ping timeout: 240 seconds]
Mic92 has quit [Ping timeout: 240 seconds]
thefloweringash has quit [Ping timeout: 240 seconds]
cocreature has quit [Ping timeout: 240 seconds]
janneke has quit [Ping timeout: 240 seconds]
talyz has quit [Ping timeout: 240 seconds]
aranea has quit [Ping timeout: 240 seconds]
janneke has joined #nixos-dev
dottedmag_ is now known as dottedmag
teozkr_ has joined #nixos-dev
Mic92 has joined #nixos-dev
janneke has joined #nixos-dev
janneke has quit [Changing host]
tokudan_ has joined #nixos-dev
cocreature has joined #nixos-dev
bridge[evilred] has joined #nixos-dev
talyz has joined #nixos-dev
yorick has joined #nixos-dev
aranea has joined #nixos-dev
tokudan has quit [Ping timeout: 240 seconds]
abbradar[m] has joined #nixos-dev
globin_ has joined #nixos-dev
timokau[m] has joined #nixos-dev
thefloweringash has joined #nixos-dev
globin has quit [Ping timeout: 272 seconds]
globin_ is now known as globin
globin has quit [Changing host]
globin has joined #nixos-dev
<domenkozar[m]>
is there a sane alternative to dunst (desktop notifications)? I'm mainly looking for something that supports input from user (i.e. a password/pin)
<domenkozar[m]>
ah, it gets inited on second call?
<mkaito>
stupid question, but what's this hash format and how do I get something I can paste into `sha256` from it? sha256-YWbg055ntjOP4S3JKg+8+zGiosPYDlieNuw6vSFLer4=
zimbatm[m] has left #nixos-dev ["Kicked by @appservice-irc:matrix.org : Idle for 30+ days"]
<adisbladis>
mkaito: Looks like base64
<adisbladis>
Which is supported by nix
<mkaito>
interesting
<adisbladis>
We have support for "sha256:_base64hash_", I'm not sure if "sha256-_base64hash_" works
<adisbladis>
Try just passing `hash = "sha256-YWbg055ntjOP4S3JKg+8+zGiosPYDlieNuw6vSFLer4=";`
<mkaito>
that seems to work, actually
<mkaito>
thanks
<adisbladis>
np :)
gleber_ has joined #nixos-dev
julm has quit [Quit: leaving]
julm has joined #nixos-dev
<niksnut>
domenkozar[m]: right, DownloadItem() can be used multiple times
greizgh has quit [Quit: No Ping reply in 180 seconds.]
nbp_ has quit [Ping timeout: 268 seconds]
Shados_ has joined #nixos-dev
Shados has quit [Ping timeout: 260 seconds]
avn has quit [Ping timeout: 260 seconds]
multun has quit [Ping timeout: 260 seconds]
avn has joined #nixos-dev
nbp has joined #nixos-dev
greizgh has joined #nixos-dev
edef has quit [Ping timeout: 252 seconds]
matthewbauer has quit [Ping timeout: 240 seconds]
Irenes[m] has quit [Ping timeout: 240 seconds]
worldofpeace has quit [Ping timeout: 240 seconds]
layus[m] has quit [Ping timeout: 240 seconds]
colemickens has quit [Ping timeout: 240 seconds]
<globin>
zimbatm, manveru ^
<teto>
what does the delta represnet in this lib.debug.traceSeq "CSApprox" = <δ:vimplugin-csapprox-2013-07-27> ?
Guest32855 has joined #nixos-dev
matthewbauer has joined #nixos-dev
Irenes[m] has joined #nixos-dev
colemickens has joined #nixos-dev
worldofpeace has joined #nixos-dev
Mbarak has quit [Remote host closed the connection]
layus[m] has joined #nixos-dev
MichaelRaskin has joined #nixos-dev
Shados has joined #nixos-dev
Shados_ has quit [Ping timeout: 246 seconds]
greizgh has quit [Ping timeout: 240 seconds]
greizgh has joined #nixos-dev
cole-h has joined #nixos-dev
Mbarak has joined #nixos-dev
Mbarak has quit [Remote host closed the connection]
edef has joined #nixos-dev
<julm>
teto: the derivation type I guess, looking at lib/generators.nix
<teto>
julm: indeed right ty
greizgh has quit [*.net *.split]
Shados has quit [*.net *.split]
pkolloch[m] has quit [*.net *.split]
domenkozar[m] has quit [*.net *.split]
sorear has quit [*.net *.split]
masaeedu[m] has quit [*.net *.split]
tokudan[m] has quit [*.net *.split]
sphalerite has quit [*.net *.split]
Shados has joined #nixos-dev
greizgh has joined #nixos-dev
sphalerite has joined #nixos-dev
masaeedu[m] has joined #nixos-dev
tokudan[m] has joined #nixos-dev
sorear has joined #nixos-dev
domenkozar[m] has joined #nixos-dev
pkolloch[m] has joined #nixos-dev
Guest32855 has quit [Quit: WeeChat 2.7.1]
justanotheruser has joined #nixos-dev
Ox4A6F has quit [Ping timeout: 240 seconds]
Dandellion has quit [Ping timeout: 240 seconds]
pkolloch[m] has quit [Ping timeout: 240 seconds]
domenkozar[m] has quit [Ping timeout: 240 seconds]
masaeedu[m] has quit [Ping timeout: 240 seconds]
tokudan[m] has quit [Ping timeout: 240 seconds]
Irenes[m] has quit [Ping timeout: 252 seconds]
roberth has quit [Ping timeout: 246 seconds]
rnhmjoj has quit [Ping timeout: 246 seconds]
aanderse has quit [Ping timeout: 246 seconds]
jonge[m] has quit [Ping timeout: 246 seconds]
jtojnar has quit [Ping timeout: 246 seconds]
bachp has quit [Ping timeout: 246 seconds]
emily has quit [Ping timeout: 246 seconds]
colemickens has quit [Ping timeout: 240 seconds]
matthewbauer has quit [Ping timeout: 246 seconds]
abbradar[m] has quit [Ping timeout: 246 seconds]
mkg20001 has quit [Ping timeout: 246 seconds]
worldofpeace has quit [Ping timeout: 252 seconds]
thefloweringash has quit [Ping timeout: 256 seconds]
ma27[m] has quit [Ping timeout: 256 seconds]
Nyanloutre[m] has quit [Ping timeout: 256 seconds]
bennofs[m] has quit [Ping timeout: 256 seconds]
layus[m] has quit [Ping timeout: 260 seconds]
timokau[m] has quit [Ping timeout: 260 seconds]
codyopel has quit [Ping timeout: 260 seconds]
dtz has quit [Ping timeout: 260 seconds]
arcnmx has quit [Ping timeout: 260 seconds]
vaibhavsagar has quit [Ping timeout: 260 seconds]
alienpirate5 has quit [Ping timeout: 260 seconds]
rycee has quit [Ping timeout: 260 seconds]
Ericson2314 has quit [Ping timeout: 260 seconds]
multun has joined #nixos-dev
drakonis has joined #nixos-dev
puck has quit [Read error: Connection reset by peer]
FRidh2 has joined #nixos-dev
FRidh has quit [Ping timeout: 246 seconds]
puck has joined #nixos-dev
<pie_[bnc]>
does nix have any sort of FHS specification?
<pie_[bnc]>
i just discovered XDG_DATA_DIRS has /run/current-system/sw/share for example
<pie_[bnc]>
nothing wrong here, but for example do we have a document that also describes the layout of /run/current-system?
sphalerite has quit [Ping timeout: 240 seconds]
NinjaTrappeur has quit [Ping timeout: 272 seconds]
sphalerite has joined #nixos-dev
sphalerite is now known as Guest62967
rycee has joined #nixos-dev
<pie_[bnc]>
If not I'll open an issue about creating such.
NinjaTrappeur has joined #nixos-dev
<pie_[bnc]>
I'm currently trying to learn how X11 config works, so if anyone has wording suggestions; https://bpaste.net/UFWQ
<teto>
pie_[bnc]: I dunno if the document already exists but that would be useful. grepping nixpkgs for environment.pathsToLink would give some hints already
<pie_[bnc]>
ok well ill just open the issue i guess
<pie_[bnc]>
MichaelRaskin: do you mind if i start pinging you in x11 stuff i run into?
<pie_[bnc]>
*on x11
<MichaelRaskin>
Hmm, not sure I actually use this part of the module
<MichaelRaskin>
Yep, I do not use this part and never cared about it. But maybe one of the sections I added to config are equivalent to that
<pie_[bnc]>
i dont even know what it does, im just going through it and noticed this as i passed by
<pie_[bnc]>
hence my question of youre ok with pings about random x11 stuff
<pie_[bnc]>
so anyway, if anyone else knows... :P
<MichaelRaskin>
In general it adds some config as if added to xorg.conf
<MichaelRaskin>
As usual, conf.d is for easier management of .conf snippets without file-level conflicts
<MichaelRaskin>
As it is about evdev, it probably tells X.org to enable some of the evdev devices using evdev driver
<MichaelRaskin>
Details might vary
ixxie has joined #nixos-dev
<samueldr>
Imagine I have a file with `{ a = throw "hi"; }`, that fails nix-instantiate witht the error clearly written, shouldn't hydra-eval-jobs also print that thrown error?
<pie_[bnc]>
modules with big let expressions should move that stuff into a local library that can be imported elsewhere
lovesegfault has joined #nixos-dev
<pie_[bnc]>
if people really want to make core changes to nix, maybe we should do something about overridability and let expressions...which is still mostly solvable by changes in engineering methods...
<pie_[bnc]>
unless im missing something blatant, most functional languages dont have a concept of overrides
<pie_[bnc]>
but we do. so why do people insist on large opaque let expressions.
<pie_[bnc]>
am I crazy?
<gchristensen>
(nor does Nix-the-language)
<pie_[bnc]>
yeah
<MichaelRaskin>
let expressions are evaluation-performance-cheap
<pie_[bnc]>
*mild exasperation*
<pie_[bnc]>
^ not @ MR
<pie_[bnc]>
MichaelRaskin: I don't think that's why people use them.
<pie_[bnc]>
please someone tell me I'm crazy yes/no
<pie_[bnc]>
my statement: we should not be using opaque let expressions because they make overriding things hard
<gchristensen>
generally agreed
<gchristensen>
but making it available for override makes it much more expensive, too
<infinisil>
pie_[bnc]: Is `some.option = [ foo bar baz ]` fine? If so, is `let values = [ foo bar baz ]; in { some.option = values; }` fine?
<infinisil>
One big reason people use let expressions is for readability
<pie_[bnc]>
infinisil: im not going to draw a line here and now about the size of things its not ok to copy and paste
<infinisil>
Just an example
<pie_[bnc]>
infinisil: ive run into cases where I wanted to filter out a specific value from a list, and indeed i can access it with mkIf config.whatever; so one case is doing something with computed values and then regenerating the expression
<pie_[bnc]>
well
<pie_[bnc]>
i dont really see a good way to express this. its an override. you want to do what you do with overrides.
<pie_[bnc]>
except without copy pasting a bunch of code
<pie_[bnc]>
its better ergonomically and you dont have to keep a kludge updated
<pie_[bnc]>
this is why i feel like i must be crazy its super obvious to me?
<infinisil>
I mean yeah sometimes thing should probably be overridable. But in general one can't say that all things defined in a `let` should be overridable
<gchristensen>
I think it is better to keep Nix simpler and understandable, and copy-paste some code when you need to do something which isn't easy to override
<pie_[bnc]>
you could put large let expressions in another file and import them.
<pie_[bnc]>
doesnt require changes to nix. adds more files.
<gchristensen>
seems decent
<pie_[bnc]>
now go do it to the entire module system :p
<pie_[bnc]>
s/system/set/
<pie_[bnc]>
i bet noone would merge my PRs
<infinisil>
pie_[bnc]: Got a good example for this?
<pie_[bnc]>
infinisil: well, I want to say: all of them, but I don't think you would like that
<pie_[bnc]>
and yes having it in a separate file is annoying
<drakonis>
i smell lisp magic again
<infinisil>
pie_[bnc]: *All* NixOS modules in nixpkgs?
<pie_[bnc]>
not sure if thats warranted, but implicit scope things yeah..
<pie_[bnc]>
s/but/with/
<pie_[bnc]>
s/yeah/yeah I don't know../
<infinisil>
Yeah that sounds reasonable there, if somebody needs it
<pie_[bnc]>
thats a lot of arguments to pass *shrug8
<pie_[bnc]>
or you call another full module system evaluation with your changes?
<pie_[bnc]>
:P
<pie_[bnc]>
infinisil: dont need it till you need it
<pie_[bnc]>
though thats a bit slippery slopey into overengineering maybe
<infinisil>
I think we should focus our efforts on things that are actually needed :)
<infinisil>
And there's a lot
<infinisil>
Actually needed *now*
<pie_[bnc]>
:PPP
<pie_[bnc]>
and i think people need to start improving their engineering practices *now* :p
<pie_[bnc]>
(and always for that matter, but sometimes you have to get some actual features done)
<infinisil>
pie_[bnc]: Got any concrete things people should improve on?
<gchristensen>
what was that FOSS doodle.com thing?
<drakonis>
framadate
<pie_[bnc]>
infinisil: expose the contents of let expressions, write more functions
<pie_[bnc]>
*write more explicitly scoped functions
<infinisil>
What are explicitly scoped functions?
<gchristensen>
framadate is so hard
<gchristensen>
I really wish I could have people just fill out the times they are able to be there, and not require me to fill in options ahead of time
<drakonis>
hoo boy framadate looks like nobody has touched it since 2010
<gchristensen>
I am but one person, who am I to create the choices
<drakonis>
it even has some weirdly specific license
<infinisil>
gchristensen: when2meet works like that
<pie_[bnc]>
infinisil: call a function with module system variables instead of using them directly i guess
<gchristensen>
when2meet is perfect! thanks!
<infinisil>
gchristensen: (though I don't think when2meet works on mobile, if that's your target audience)
<worldofpeace>
yeah, framadate is torture to use. when2meet looks pretty good
<samueldr>
I'm still waiting on an actual failure to finish evaluating to know whether it works for real or not, but my simplified test case does mirror the errors on stderr again with https://github.com/NixOS/hydra/pull/729
<pie_[bnc]>
gchristensen: a possible alternative i forgot to mention: expose an attribute where you iherit all the stuff in the let expression. which is also kind of ugly because you have to copy all the names
<tilpner>
But also good, because now you get to choose your public API (and what deserves changelog entries when refactored)
<Ericson2314>
this is about the overriding rfc?
<Profpatsch>
gchristensen: pie_[bnc] I also had the epiphany that often it’s a lot more work to override something / make it overridable than it is to just copy and paste.
<Profpatsch>
Alternatively, factor it out. That’s why nixpkgs is a monorepo. You can just touch other people’s code OoO
<Profpatsch>
(which flakes would make a lot harder, but I digress)
<Ericson2314>
the overriding problem i becaue nixpkgs is trying to be both the library and the example program at once
<Ericson2314>
<Profpatsch "Alternatively, factor it out. Th"> absolutely
<pie_[bnc]>
Profpatsch: and if people are willing to merge mor lib functions :p
<Ericson2314>
yes please, general purposes programming language!
<pie_[bnc]>
Ericson2314: didnt see anything about an overriding rfc
<Profpatsch>
well, factoring out has two steps in general
<pie_[bnc]>
I dont immediately see how this works <Ericson2314> the overriding problem i becaue nixpkgs is trying to be both the library and the example program at once
<Profpatsch>
2) give a function argument to make a template
<pie_[bnc]>
sounds to me like nixpkgs isnt trying hard enogh to be a library
<Ericson2314>
the one time I do use overriding is more "vertical" than "horizontal", e.g. I want to change the call to regular mkDerivation in a python build*Package package
<Ericson2314>
pie_[bnc]: Basically, there should be a way to have a Nixpkgs with 0 fixed points, which today's nixpkgs is made from
<pie_[bnc]>
Ericson2314: oh geez this rfc sounds like something i should really dig into but i dont have the capacity for it...
<Profpatsch>
Also, lol at config languages which don’t have functions but allow for “interpolation”
<Ericson2314>
so you gotta pass the binutils to the packages that need it, yourelf
<Ericson2314>
lol indeed
<pie_[bnc]>
Ericson2314: i think profpatsch really wants that
<gchristensen>
:/ it looks like I have made ofborg even worse in my attempts to make it better. unfortunately Emily and I are getting quite hungry and I'm getting grumpy, so I'm going to go make / eat some dinner, and then fix it.
ekleog has joined #nixos-dev
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos-dev
<cole-h>
The food will get your brain going
<Profpatsch>
or it will make you sleepy
orivej_ has joined #nixos-dev
orivej has quit [Ping timeout: 260 seconds]
<worldofpeace>
Profpatsch: that depends if you are super diverting blood to focus on digestion (lowering body temperature)
* Profpatsch
is super diverting alcohol to his neural net to increase focus while coding
<ma27[m]>
prost!
<Profpatsch>
Weltenburger Kloster Urtyp Hell <- in Germany so finally good beer again
<worldofpeace>
lol, I have no idea what I meant be super diverting. But it's a moment, I'm going to say it all the time now.
<ryantm>
I searched for the term, worldofpeace #soundslegit
<MichaelRaskin>
There is also supper-diverting, I guess?
justanotheruser has quit [Ping timeout: 256 seconds]