worldofpeace changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | NixOS 20.09 Nightingale ✨ https://discourse.nixos.org/t/nixos-20-09-release/9668 | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | https://r13y.com | 20.09 RMs: worldofpeace, jonringer | https://logs.nix.samueldr.com/nixos-dev
hexa- has quit [Quit: WeeChat 2.9]
hexa- has joined #nixos-dev
abathur has quit [Quit: abathur]
cole-h has quit [Ping timeout: 240 seconds]
abathur has joined #nixos-dev
rajivr has joined #nixos-dev
krkini has quit [Remote host closed the connection]
kini has joined #nixos-dev
justan0theruser has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos-dev
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
AlwaysLivid has quit [Ping timeout: 240 seconds]
<siraben> Another newcomer ran into a variation of https://github.com/NixOS/nixpkgs/issues/118481
<{^_^}> #118481 (by fishy, 1 week ago, open): `nix-env --upgrade` upgrades 'nix-2.3.10' to 'nix-2.3.10-x86_64-unknown-linux-musl'
orivej has joined #nixos-dev
cjb has quit [Ping timeout: 268 seconds]
cole-h has joined #nixos-dev
orivej has quit [Ping timeout: 240 seconds]
lopsided98 has quit [Ping timeout: 258 seconds]
lopsided98 has joined #nixos-dev
srk has joined #nixos-dev
jonringer has quit [Ping timeout: 258 seconds]
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
cole-h has quit [Ping timeout: 252 seconds]
__monty__ has joined #nixos-dev
<domenkozar[m]> siraben: yikes
<domenkozar[m]> we really need to address this (at the very least by communicating)
<domenkozar[m]> this is #1 reason people go away from Nix
<domenkozar[m]> (not this bug in particular, but such things where someone follows documentation and it breaks their whole setup)
<domenkozar[m]> I need to run now, but I'll be back in an hour or so
<ekleog> I may be an extremist, but IMO nix-env should be removed altogether, or put behind a scary flag
<domenkozar[m]> yeah I do too, but that's requires a lot of planning and work
<domenkozar[m]> we need a solution for this now
<domenkozar[m]> could we rename nix-2.3.10-x86_64-unknown-linux-musl?
<ekleog> true
<sterni> domenkozar[m]: actually I'm a bit confused why this is picked up at all
<sterni> domenkozar[m]: pkgsStatic doesn't have recurseForDerivations
<sterni> huh graham isn't around
<sterni> maybe nix-env has started to treat recurseIntoAttrs like recurseForDerivations?
<symphorien[m]> there is a nixStatic attribute iirc
<symphorien[m]> > nixStatic
<{^_^}> "<derivation /nix/store/85y7aa01xgqgaa16drn0zasb89chyald-nix-2.3.10-x86_64-unknown-linux-musl.drv>"
<sterni> ohh
<sterni> :|
<symphorien[m]> maybe it's enough to just remove it
<sterni> I would prefer not to
<niksnut> or give it a different priority
<niksnut> or fix the name, it should be 'nix-static-2.3.10' probably
<niksnut> features need to be part of the name, not the version
<symphorien[m]> priority doesn't influence nix-env -u, does it ?
<symphorien[m]> only the implicit buildEnv behind nix-env
<niksnut> hm yeah
<sterni> #119310
<{^_^}> https://github.com/NixOS/nixpkgs/pull/119310 (by sternenseemann, 9 seconds ago, open): nixStatic: set derivation name to "nix-static-${version}"
<niksnut> we should probably get rid of all 'x86_64-unknown-linux-musl' suffixes since they don't really serve a purpose and break the version field
<niksnut> I can understand including that info for a cross-build, but this is a native package
<niksnut> being compiled with musl is an implementation detail
devhell has joined #nixos-dev
<sterni> niksnut: yeah for pkgsStatic it's leaking the implementation detail that its using the cross infrastructure pretty much
<qyliss> agree
<qyliss> a -static suffix would be nice though to differentiate it from the normal one
<sterni> why do we set crossSystem anyways for pkgsStatic? it isn't set for pkgsMusl
<sterni> the question is does stuff break if we just remove it
<sterni> qyliss: it should be easy to add something to mkDerivation which appends -static to the name if enableStatic
Kritnich has joined #nixos-dev
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
orivej has joined #nixos-dev
<sterni> #119317 fingers crossed this doesn't cause any freak rebuilds
<{^_^}> https://github.com/NixOS/nixpkgs/pull/119317 (by sternenseemann, 34 seconds ago, open): stdenv.mkDerivation: add -static to name if building statically
AlwaysLivid has joined #nixos-dev
AlwaysLivid has quit [Remote host closed the connection]
AlwaysLivid has joined #nixos-dev
AlwaysLivid has quit [Ping timeout: 240 seconds]
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
orivej has quit [Ping timeout: 265 seconds]
<hexa-> > Could not access KVM kernel module: Permission denied
<{^_^}> error: syntax error, unexpected ':', expecting ')', at (string):494:35
<hexa-> check if the user running the test has access to /dev/kvm
<hexa-> needs to be in the kvm group
orivej has joined #nixos-dev
pmy has quit [Quit: WeeChat 3.1]
justanotheruser has quit [Quit: WeeChat 2.9]
<domenkozar[m]> ScottHDev: answered on discourse :)
AlwaysLivid has joined #nixos-dev
rj has joined #nixos-dev
jonringer has joined #nixos-dev
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
AkechiShiro has quit [Quit: WeeChat 2.9]
<gchristensen> I'm a little surprised there hasn't been more feedback w.r.t. the style change on hydra.nixos.org
<cransom> oh. huh. i'm not there often. it's weirdly less dense now and bigger fonts.
<gchristensen> maybe shouldn't tempt fate :)
rj has quit [Ping timeout: 240 seconds]
<qyliss> gchristensen: I don't think the width of this table should be constrained: https://hydra.nixos.org/project/nixpkgs
<qyliss> the dates are forced onto two lines even on my big monitor so I don't get to see many rows
<qyliss> this is awesome work though. thank you so much for everything you're doing with Hydra
<qyliss> (it might have always been like this but the big font makes it more noticeable)
<gchristensen> this was from twhitehead, in a PR from september of 2019 :') https://github.com/NixOS/hydra/pull/677
<{^_^}> hydra#677 (by twhitehead, 1 year ago, merged): Javascript libraries update
<gchristensen> that wrapping is a bit weird, maybe if those were made in to relative dates it'd be more usable and use less space? I think there are good reasons to constrain widths but I'm no expert, just parroting what I think I've heard
MichaelRaskin has quit [Ping timeout: 240 seconds]
<qyliss> there are good reasons to constrain widths of prose, yeah
<qyliss> I don't think it applies to multi-column tables
rj has joined #nixos-dev
<qyliss> Relative dates make it hard to see when things actually happened, and they can be misleading with rounding
<gchristensen> yeah
<qyliss> maybe I can put my money where my mouth is and make a PR
<qyliss> everything else I wanted to do today is blocked on rebuilding stdenv anyway
<gchristensen> (♥‿♥✿)
<qyliss> I do want to improve my Perl...
<gchristensen> I've found it fairly reasonable once I figured out how to write tests
<gchristensen> I can't write valid perl without starting with tests
<qyliss> I like what I've seen of Perl but I'm not confident in it
<gchristensen> start with a test :)
<qyliss> tbh I think we could probably just unconstrain every page on Hydra
<gchristensen> I'll have to defer to samueldr on this
<qyliss> it's all tables and preformatted text that would just scroll off to the side if it was width-constrained
tv has quit [Ping timeout: 260 seconds]
rj has quit [Remote host closed the connection]
rj has joined #nixos-dev
<cransom> +1 on that. my web browser is on a portrait rotated monitor. anything left in the margins for me is wasted space, especially since it's already narrower than typical usage.
<gchristensen> would love for an expert at this get involved
tv has joined #nixos-dev
tv has quit [Client Quit]
tv has joined #nixos-dev
orivej has quit [Ping timeout: 268 seconds]
MichaelRaskin has joined #nixos-dev
<MichaelRaskin> Maybe people who care about font choice have force-overrides in the browser already
<qyliss> The development version of Bootstrap has a bunch of different container sizes, so we could just change to a bigger one
<qyliss> But the version we're using doesn't :(
<qyliss> oh actually we'd just need to go up one minor version
AkechiShiro has joined #nixos-dev
justanotheruser has joined #nixos-dev
pmy has joined #nixos-dev
orivej has joined #nixos-dev
AlwaysLivid has quit [Ping timeout: 240 seconds]
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
mkaito has joined #nixos-dev
mkaito has joined #nixos-dev
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
cole-h has joined #nixos-dev
qyliss has quit [Quit: bye]
qyliss has joined #nixos-dev
<samueldr> gchristensen: it's probably good to switch to container-fluid
<samueldr> because yes, this is an *app* more than *contentful* web
<samueldr> and if there is a place where constraining width helps, we can, at this location, add what is needed
pbogdan has quit [Quit: ZNC 1.8.2 - https://znc.in]
mkaito has quit [Quit: WeeChat 3.1]
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
justan0theruser has joined #nixos-dev
justanotheruser has quit [Ping timeout: 250 seconds]
ScottHDev has quit [Ping timeout: 265 seconds]
rajivr has quit [Quit: Connection closed for inactivity]
<gchristensen> domenkozar[m]: what is the domain for your precommit thing?
lukegb has quit [Quit: ~~lukegb out~~]
kini has quit [Remote host closed the connection]
lukegb has joined #nixos-dev
kini has joined #nixos-dev
<domenkozar[m]> https://pre-commit.com/ ?
<gchristensen> ah yeah thanks
plumm has quit [Quit: Textual IRC Client: www.textualapp.com]
orivej has quit [Ping timeout: 240 seconds]
plumm has joined #nixos-dev
sid_cypher has joined #nixos-dev
srk has quit [Ping timeout: 260 seconds]
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
* samueldr grumbles
<samueldr> the "unstable-YYYY-MM-DD" thing is getting to me a bit
<samueldr> because I have a package that will never have a release
<samueldr> it's not really a "package"
<samueldr> it's a collection of firmware files
<supersandro2000> just don't make it a package
<samueldr> it adds no value to append unstable to the package name or prepend unstable to the version
<samueldr> supersandro2000: what should it be then?
<supersandro2000> if it is a derivation for in some directory it does not need a version
<supersandro2000> which is not exposed in top-level
<samueldr> then the end-user can't override it
<supersandro2000> should they?
<samueldr> maybe
<supersandro2000> then it needs a version
<samueldr> I don't know what the user is doing
<supersandro2000> there is an unstable updater
<supersandro2000> FYI
<samueldr> an unstable updater?
<samueldr> I don't see how it matters with what I was saying
<samueldr> I know the current thing is to add unstable to `pname`
<samueldr> and I'll do it because "it's how we do it"
<samueldr> but it adds no value in a package which is a collection of firmware files that will never see a released version
<supersandro2000> the current thing is to add unstable to version and to pname if there are two packages in nixpkgs
<supersandro2000> but it needs a version :P
<samueldr> sure, and the date is fine there
<samueldr> but I don't see any value in adding unstable
<samueldr> and it's clearly written
<samueldr> >> If a package is not a release but a commit from a repository, then the version part of the name must be the date of that (fetched) commit. The date must be in "YYYY-MM-DD" format. Also append "unstable" to the name - e.g., "pkgname-unstable-2014-09-23".
<samueldr> append unstable to the package name
<supersandro2000> but the package has no release numbers and just using the date would imply it has
<supersandro2000> samueldr: that sentence does not solve the question because it was written before pname + version
<sterni> samueldr: just add unstable to version only
__monty__ has quit [Quit: leaving]
<sterni> iirc pname = "something-unstable" just causes a lot of problems
<sterni> (“problems” as in repology doesn't like it)
<supersandro2000> so its ambiguous and people use pname for the package name all over the place which breaks if you change it to *-unstable
<supersandro2000> also we would technically add and remove packages when changing pname and pname should match the attribute name
<samueldr> then fix the documentation
<supersandro2000> but nix is not smart enough to correctly parse unstable-XXXX-XX-XX in some functions
<supersandro2000> samueldr: I could just push it straight to master and hope no one notices it
<samueldr> no
<supersandro2000> otherwise I need to get some more wood and nails
<supersandro2000> and we need to build that bikesheed
<samueldr> but again, I wrote YYYY-MM-DD for the version because we need a version, but it's not "unstable", so I don't think unstable adds any kind of value
<supersandro2000> but YYYY-MM-DD is the version format of some programs who thought semver is "complicated" and we can't create version numbers for programs which do not exist
<samueldr> I don't follow how that's relevant
<samueldr> we're not doing maths with version number across unrelated packages
<supersandro2000> If you set the version number to YYYY-MM-DD it implies that the packages has versioned releases
<samueldr> but it's not even really a package, it's a derivation with "data"
<samueldr> you won't be installing this
<samueldr> it's a transient dependency
<samueldr> if we had a way to not have them listed as packags, but present for callPackage dependency injection, it would be that
<evils> supersandro2000: are you arguing an ISO8601 date on its own should only be used when upstream used that as a release name?
<evils> s/release name/release version string/
<supersandro2000> Can we make packages not installable?
<samueldr> we don't even have a clear definition of "package" really
<supersandro2000> evils: no, that is how it is.
<evils> supersandro2000: there no defined "how it is" there's how mergers accept stuff
<supersandro2000> but do you want to get that booting thingy done or continue talking about version numbers?
<supersandro2000> evils: but isn't that the defacto standard?
<evils> i don't think there is a defacto standard in this case
<evils> i think a date string is a perfectly fine identifier for an arbitrary commit; sometimes i prefer a git short commit hash to identify a commit though
<samueldr> anyway, I wanted to take the pulse of multiple contributors on *requiring* the "unstable" moniker without any judgement even if it brings no value
<samueldr> but I'll continue working on my "boot thingy" instead
<samueldr> as I don't think the discussion will be fruitful currently
<evils> sorry if i disrupted the discussion by jumping in; i've been annoyed by this since it was suggested on a PR of mine 3 months ago and i consider it an impasse; but i haven't put serious thought into it
<supersandro2000> evils: it is not allowed to reference git commits in fetches with short hashes
<evils> supersandro2000: i meant using a git short commit as a version string; usually clearer that it's not an upstream endorsed release
<supersandro2000> because you can brute force bad commit short shas using forks
<supersandro2000> yeah, could be but changing everything is a bit of work I avoid for now
<evils> to me: semver tag implies some stability, a date just says "it's whatever it was on this date" (as determined by the packager or upstream) and a git short commit has almost no meaning (besides implying that that commit isn't tagged) (though i do have a local change that uses a short commit hash of a tagged commit...)
<evils> supersandro2000: ^ and good night :)
rj has quit [Ping timeout: 240 seconds]
<supersandro2000> if pytest uses a format of date versioning it cannot be whatever was on that date
<supersandro2000> otherwise it would be always broken
avocadoom has joined #nixos-dev
kini has quit [Remote host closed the connection]
kini has joined #nixos-dev
supersandro2000 has quit [Killed (kornbluth.freenode.net (Nickname regained by services))]
supersandro2000 has joined #nixos-dev
globin has quit [Ping timeout: 240 seconds]