gchristensen changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | 18.03 release managers: fpletz and vcunat
<Sonarpulse> shlevy: can we remove old jank builtins from nix?
<Sonarpulse> there's lots of stuff it would be nicer to....not add to hnix :D
<Sonarpulse> __override for example
<Sonarpulse> I dunno what the back compat story is for nix
<shlevy> Sonarpulse: I'd just make a call in hnix
<shlevy> Decide what you want to support from nixpkgs and call it a day
<Sonarpulse> shlevy: as in removing things from nix is unlikely?
<Sonarpulse> right now the goal is everything
<Sonarpulse> there is more effort recently
<shlevy> Not so much unlikely as unnecessary up front
<shlevy> If hnix works well for people's use cases
<Sonarpulse> makes a lot of sense
<shlevy> Then that's a good argument to remove :)
<Sonarpulse> how about deprecation?
<Sonarpulse> __override is already unused in nixpkgs
<Sonarpulse> fwiw
<shlevy> yeah, I think that makes sense. Some method to show a warning if a builtin is used, then we can add things to that list
<Sonarpulse> cool
kmicklas has joined #nixos-dev
<Sonarpulse> shlevy: <shlevy> yeah, I think that makes sense. Some method to show a warning if a builtin is used, then we can add things to that list
<Sonarpulse> then worry about removing later
<Sonarpulse> also, btw, we are interested in banning hashless builtins.fetch* but I forget how that's done without also banning IFD
<shlevy> pure-eval?
<shlevy> But that also bans local path reference without hashes/builtins.path
<Sonarpulse> ah true
<Sonarpulse> yeah I appreciate the argument that impure path is no worse
<Sonarpulse> hmm
<Sonarpulse> interesting to think about
<Sonarpulse> shlevy: store path is ok, right?
<shlevy> By the way, what's the impetus for the recent interest?
<shlevy> Sonarpulse: Yeah
<Sonarpulse> shlevy: you'd best ask ryan
<Sonarpulse> but one thing is an inotify thing to automatically reeval
<Sonarpulse> kind of a ghcid for nix
<shlevy> Aaah cool
<Sonarpulse> shlevy: so store path without hash is fine, right?
<Sonarpulse> cause it's already hashed?
<shlevy> Yeah, if specified via builtins.storePath
<shlevy> I don't know if it actually is in current Nix
<shlevy> but it should be
<Sonarpulse> err what about ./foo/bar from /nix/store/.../asdf.nix
<Sonarpulse> that should be, but who knows, you are saying?
<shlevy> What I was specifically referring to was (builtins.storePath /nix/store/.../foo/bar)
<Sonarpulse> oh i forget about that one
<shlevy> but I can see an argument for supporting that case too... I *think* there's some logic for it in at least the restricted eval case, not sure about pure eval
<Sonarpulse> ok
<shlevy> Yeah, if you import from derivation all the paths there are put into allowed-paths
<Sonarpulse> yeah i don't even know how those settings are set
<Sonarpulse> I'm very out of the loop on those modes
<shlevy> So it should work with import-from-drv and even with references inside of builtins.storePath
<Sonarpulse> shlevy: what if we reuse those modes for the deprecated builtins?
<shlevy> ?
<kmicklas> shlevy: Sonarpulse tells me there's some kind of "strict" evaluator mode
<Sonarpulse> kmicklas: just thought of that for the deprecated builtins as I said things allowed based on what you said for store paths
<shlevy> You mean disallow deprecated builtins in pure-eval? That seems completely orthogonal
<kmicklas> I'm interested in using that to deprecate __overrides if that's possible
<Sonarpulse> shlevy: make a new thing like pure-eval
<Sonarpulse> orthogonal "anti-feature flag"
<shlevy> I think we should distinguish deprecated from unwise-for-other-reasons
<Sonarpulse> worry about the warning case
<shlevy> Anyway, this is probably better in a concrete RFC
<Sonarpulse> oh definitely
<Sonarpulse> but good to know the code for the PR before the RFC maybe :D
<Sonarpulse> shlevy: https://github.com/NixOS/nixpkgs/pull/38485 btw could probably use a re-review
<Sonarpulse> I realized how I can just instantiate some things in nixos/release.nix to test
<Sonarpulse> and found a bunch of stuff
<Sonarpulse> shlevy: left two comments on the off-chance you started looking
<Sonarpulse> Dezgeg: thanks
<Sonarpulse> Dezgeg shlevy: anyone nice way to set mutually exclusive options?
<Sonarpulse> I can also make custom packages expression rule out the others
<shlevy> Sonarpulse: There's a deprecation mechanism in NixOS
<Sonarpulse> ok I'll start with that
<Sonarpulse> cool nixos/modules/config/krb5/default.nix has both
coconnor has quit [Remote host closed the connection]
mbrgm has quit [Ping timeout: 256 seconds]
mbrgm has joined #nixos-dev
<dtz> oh YAY better nixos platform support?! haha my localSystem kludges can't wait to be removed xD
<Sonarpulse> dtz: :)
<Sonarpulse> I made you a reviewer
<Sonarpulse> if you wanna be responsible for this one
<Sonarpulse> (I don't know about nixos modules)
<Sonarpulse> I'll go finish your fixed output names one :)
kmicklas has quit [Ping timeout: 256 seconds]
alp has quit [Ping timeout: 252 seconds]
alp has joined #nixos-dev
orivej has joined #nixos-dev
pie_ has joined #nixos-dev
orivej has quit [Ping timeout: 240 seconds]
phreedom has quit [Read error: Connection reset by peer]
phreedom has joined #nixos-dev
davidlt has joined #nixos-dev
MichaelRaskin has quit [Quit: MichaelRaskin]
orivej has joined #nixos-dev
goibhniu has joined #nixos-dev
ma271 has joined #nixos-dev
ma271 has quit [Ping timeout: 260 seconds]
Jackneill has quit [Read error: Connection reset by peer]
Jackneill has joined #nixos-dev
Synthetica has joined #nixos-dev
<Synthetica> Is there a reason that we are use the .deb for atom instead of building from source?
<tilpner> Probably because it's harder to build from source, and the packager chose to have the deb instead of not at all
ma271 has joined #nixos-dev
ma271 has quit [Client Quit]
ma27 has joined #nixos-dev
mbrock has quit [Disconnected by services]
<NinjaTrappeur> Hey, I'm diving in nix code these days. I'm fixing minor bugs as an excuse to read the code base. All the backlogged issues are already assigned to somebody. I did not found any document describing your issues workflow, is it ok for me to fix some of these?
<NinjaTrappeur> (the ones that seems abandonned)
<NinjaTrappeur> I don't wanna look rude by mistake :)
<Mic92> NinjaTrappeur: you don't need an excuse for fixing bugs. Everyone appriate this.
<Mic92> race condition are possible but too likely
<Mic92> *not too
<NinjaTrappeur> Alright
JosW has joined #nixos-dev
JosW has quit [Ping timeout: 245 seconds]
JosW has joined #nixos-dev
orivej has quit [Ping timeout: 260 seconds]
<Dezgeg> hm: warning: unable to download 'https://cache.nixos.org/y17nir3py2z065glzxh6wyp17xdkaddh.narinfo': Peer certificate cannot be authenticated with given CA certificates (60);
<copumpkin> LnL: checkPhase on all derivations? have a reference? I must've missed that :)
<LnL> copumpkin: no just pyhton stuff
<copumpkin> ah okay
Synthetica has quit [Quit: Connection closed for inactivity]
<Dezgeg> gchristensen: you never ran into this certificate stuff with the multi-user installer? never needed to set NIX_SSL_CERT_FILE for the daemon?
<LnL> pretty sure you need to set that
<Dezgeg> yes, that's why I'm asking :P
<Dezgeg> ah I think I missed this scripts/nix-profile-daemon.sh.in
zybell_ has quit [Quit: Leaving]
zybell_ has joined #nixos-dev
ma27 has quit [Ping timeout: 260 seconds]
ma27 has joined #nixos-dev
kmicklas has joined #nixos-dev
<Sonarpulse> niksnut: you there?
zybell_ has quit [Ping timeout: 260 seconds]
zybell_ has joined #nixos-dev
ma27 has quit [Ping timeout: 255 seconds]
peti has joined #nixos-dev
goibhniu has quit [Ping timeout: 264 seconds]
MichaelRaskin has joined #nixos-dev
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 240 seconds]
vcunat has joined #nixos-dev
vcunat has quit [Client Quit]
<disasm> Anyone have any problems with me cherry picking these two commits to 18.03? 80582c600da vimb-unwrapped: depend on generic webkitgtk instead of 218x AND 0b8d7fb16e1 wxwidgets: depend on generic webkitgtk instead of 218x
<disasm> currently both are overriding webkitgtk with a non-existing webkitgtk218x attribute
Lisanna has joined #nixos-dev
<LnL> sounds fine
<disasm> pushed, thanks for the approval LnL
<infinisil> Just opened https://github.com/NixOS/nixpkgs/pull/38522, which makes lib extensible, really useful for the module system outside of nixpkgs
<LnL> yeah looked at it, but I'd have to read through the source again to understand the implications
<infinisil> Basically the only difference is that lib now has the extend attribute which you can pass an overlay-like function to extend it
<infinisil> And these changes get properly reflected in evalModules
Lisanna has quit [Ping timeout: 240 seconds]
<taktoa> would anyone be interested in doing a group buy of a custom litographs shirt for Nix? e.g.: https://www.litographs.com/collections/t-shirts/genre_source_code
<taktoa> they're shirts that are completely covered in some kind of text, and they sell a few that are covered in the source code of various projects
<clever> now i need to remember what part of the shirt primops.cc is on!
<taktoa> (you can easily get a custom shirt made with whatever text you want, but I don't see any option for doing a custom design with different colors, so I'd probably have to email them about it, hence my interest in a group buy)
<infinisil> taktoa: No syntax highlighting, mehhh
<taktoa> lol
JosW has quit [Quit: Konversation terminated!]
<copumpkin> niksnut: if I'm seeing "acquiring/releasing lock: Resource deadlock avoided", any idea what that might be about?
<copumpkin> maybe false alarm
coconnor has joined #nixos-dev
<copumpkin> I don't suppose builtins.functionArgs can be taught to operate on functors
<copumpkin> also, why do functors take a self argument?
<copumpkin> oh I guess so if you modify part of the attrset, the function can see it
Lisanna has joined #nixos-dev
Lisanna has quit [Quit: Lisanna]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 240 seconds]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 265 seconds]
<aminechikhaoui> gchristensen: which Packet.net server types are being used for hydra.nixos.org ?
<gchristensen> Type 2 and Type 2A
<gchristensen> they have new names ....
<aminechikhaoui> hm I guess m1.xlarge.x86 from https://www.packet.net/bare-metal/servers/memory/
<aminechikhaoui> thanks !
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 264 seconds]
<Mic92> fpletz: https://github.com/zfsonlinux/zfs/pull/7329 that one is interesting for nixos
<Mic92> makes mountpoint=legacy for zfs obsolete
<clever> nice
<clever> i sometimes forget to add fileSystems. entries for new datasets
<clever> and then things break in fun ways
<infinisil> clever: Gotta use nix to manage your partitioning and fileSystems :P
<clever> infinisil: that gives me an idea, a zfs module so that just setting fileSystems."/foo" = { type = "zfs"; device = "pool/foo"; }; will automatically do `zfs create` on bootup
davidlt has quit [Ping timeout: 260 seconds]
<infinisil> Yeah
<ekleog> wouldn't it be nice to have for all filesystems, and not only for zfs?
<ekleog> ISTR someone started work in that direction, can't remember where it went
Sonarpulse has quit [Quit: Leaving]
<infinisil> ekleog: Ah it was aszlig: https://github.com/NixOS/nixpart
<ekleog> thanks! I wasn't finding it again :)