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
kmicklas has quit [Ping timeout: 256 seconds]
pie__ has quit [Ping timeout: 264 seconds]
<thoughtpolice> I'm surprisingly irritated by having to do that.
<simpson> thoughtpolice: Above and beyond, IMO. I would +1 splitting this, and I'd also +1 getting it merged.
<simpson> gchristensen: I am digging how ofborg shows me CLI-like incantations for each of its test.
<simpson> *of its tests, even.
<thoughtpolice> There isn't much to reasonably split out, IMO, except the version bumps. The rest is all fairly small and self contained.
<thoughtpolice> Plus the tests need to be fixed. In theory I could split the test migration out into another PR, but the current one is IMO basically about right.
<simpson> I was indeed thinking of those version bumps.
orivej has quit [Ping timeout: 260 seconds]
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
taktoa has quit [Remote host closed the connection]
taktoa has joined #nixos-dev
mbrgm has quit [Ping timeout: 264 seconds]
mbrgm has joined #nixos-dev
pie__ has joined #nixos-dev
zybell_ has quit [Ping timeout: 260 seconds]
pie__ has quit [Ping timeout: 240 seconds]
ma27 has quit [Ping timeout: 260 seconds]
pie__ has joined #nixos-dev
ma27 has joined #nixos-dev
pie__ has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-dev
ma27 has quit [Ping timeout: 245 seconds]
pie__ has joined #nixos-dev
ma27 has joined #nixos-dev
jtojnar_ has joined #nixos-dev
jtojnar has quit [Quit: jtojnar]
jtojnar_ has quit [Ping timeout: 260 seconds]
<clever> found a minor bug in https://nixos.org/nixos/options.html#programs.ssh
<clever> Extra configuration text appended to ssh_config. See ssh_config5 for help.
<clever> See <citerefentry><refentrytitle>ssh_config</refentrytitle><manvolnum>5</manvolnum></citerefentry>
<peti> Perl seems broken in current revisions of 'master'. Apparently, it can no longer deal with an UTF-8 locale? Beats my why that is, and it happened only after I updated a couple of days ago. Am I the only one who's seeing that (on NixOS)?
<adisbladis[m]> peti: I saw some other complaints on the issue tracker related to locales earlier today. Might be something bigger going on.
<adisbladis[m]> I'm too busy at work today to look into it further
<LnL> is there a way to not fail when a cache is unavailable? fallback doesn't seem to work
jtojnar has joined #nixos-dev
<LnL> :/
pie__ has quit [Ping timeout: 255 seconds]
<LnL> bloody dns caching...
zybell has joined #nixos-dev
<Mic92> We should have soname bumps for opengl drivers. The last driver update lead to incompatibilities between /run/opengl-drivers and the nix store.
ckauhaus has joined #nixos-dev
ma27 has quit [Ping timeout: 256 seconds]
orivej has quit [Ping timeout: 268 seconds]
jtojnar has quit [Read error: Connection reset by peer]
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 245 seconds]
orivej has joined #nixos-dev
pie__ has joined #nixos-dev
pie__ has quit [Ping timeout: 256 seconds]
pie__ has joined #nixos-dev
jtojnar has joined #nixos-dev
pie__ has quit [Ping timeout: 264 seconds]
pie__ has joined #nixos-dev
<peti> locale support is broken in master after the recent glibc 2.27 update.
<peti> Programs can no longer configure an UTF-8 environment.
<peti> That seems kinda bad.
Synthetica has joined #nixos-dev
pie_ has joined #nixos-dev
pie__ has quit [Read error: Connection reset by peer]
pie_ has quit [Read error: Connection reset by peer]
pie_ has joined #nixos-dev
<shlevy> Looks like the locale code successfully opens the archive then looks in $glibc/lib/locale, which doesn't exist
<shlevy> And 2.26 doesn't do that
<shlevy> Will look more this afternoon...
<copumpkin> builtins.tryEval doesn't like errors from builtins.fetchurl? :(
<shlevy> Nope. I'm not sure if that was intentional, but in our nix-fetchers project (soon to be open sourced) I return error codes instead of throwing
<ikwildrpepper> shlevy: what kind of stuff will we be able to fetch with nix-fechters?
<ikwildrpepper> fetchers.. still can't type...
<shlevy> ikwildrpepper: The only thing that currently exists is fetch-pypi-hash, which relies on pypi's "no reupload of the same file" policy and uses its API to get the relevant hash
<shlevy> ikwildrpepper: Planning on at least a variant of fetchgit that does caching a bit differently and has a few other tweaks... eventually to be upstreamed once validated I hope
<shlevy> ikwildrpepper: And may have something to make turning a stackage LTS number and a set of dependency names into a package set without needing to pre-determine hashes or anything
<copumpkin> boo :)
<shlevy> why? :)
<shlevy> Because stackage or because you don't trust the purity?
<LnL> I think there's a bug with ssh-ng + build progress indicator
<LnL> it streams the entire log instead of a single line
ma27 has joined #nixos-dev
<copumpkin> shlevy: boo because I can't catch fetchurl errors :)
pie_ has quit [Ping timeout: 264 seconds]
orivej has quit [Ping timeout: 256 seconds]
kmicklas has joined #nixos-dev
<shlevy> Ah :D
<thoughtpolice> Is there a hard limit on what should go to staging? Some postgresql bumps will affect 'postgresql', which will percolate a bit.
<thoughtpolice> I mean in terms of # of rebuilds, sorry. Borg tagged it at ~500 rebuilds for grahamc's original update to postgresql patch levels: https://github.com/NixOS/nixpkgs/pull/36386
<thoughtpolice> I don't think that's excessive but I didn't really think about it until recently.
<thoughtpolice> Mmmm, actually, maybe I should keep these updates in my PR. ;_;
<thoughtpolice> I have another change that will imply a rebuild of all postgresql derivations anyway (because they lacked a 'version' attribute previously), so it's probably better to do this only once rather than twice.
Sonarpulse has joined #nixos-dev
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
Synthetica has quit [Quit: Connection closed for inactivity]
Lisanna has joined #nixos-dev
ma27 has quit [Quit: WeeChat 2.0]
ma27 has joined #nixos-dev
ckauhaus_ has joined #nixos-dev
ckauhaus_ has quit [Remote host closed the connection]
ckauhaus_ has joined #nixos-dev
ckauhaus_ has quit [Client Quit]
jtojnar has quit [Quit: jtojnar]
jtojnar has joined #nixos-dev
orivej has joined #nixos-dev
ma27 has quit [Remote host closed the connection]
ma27 has joined #nixos-dev
ma27 has quit [Ping timeout: 245 seconds]
ckauhaus has quit [Quit: Leaving.]
Sonarpulse has quit [Ping timeout: 260 seconds]
goibhniu has joined #nixos-dev
<copumpkin> has anyone thought about a builtins.readGitIgnore to be fed into filterSource? :P
<copumpkin> I think someone started doing that in pure nix
<copumpkin> but parsing and string manipulation in nix isn't beautiful
<Dezgeg> using builtins.fetchGit should suffice (or if not, it should be fixed)
<dtz> if Nix required 'git' to be available (or used it as a library) perhaps that would help? O:)
<LnL> yeah fetchGit solves a lot of the problems with ./.
<copumpkin> Dezgeg: well, I also want a filterSource-like logic
<copumpkin> I don't want the full git tree
<copumpkin> also, doesn't fetchGit ignore the working tree? or am I misremembering?
<shlevy> copumpkin: In practice I've found it sufficient to whitelist certain directories (e.g. 'src'), but this is a great thing to do with extraBuiltins ;)
<copumpkin> :)
<LnL> no fetchGit ./. works
<Dezgeg> if there's no option to have the working tree then it should be fixed to optionally have the working tree
<shlevy> Dezgeg: you can just use a normal path...
<copumpkin> shlevy: in my case, there are directories I want, but other legacy build tools also act like they own the folders
<Dezgeg> that gives you .gitignored files as well
Sonarpulse has joined #nixos-dev
<copumpkin> ah
<copumpkin> so if I could easily filter the output of fetchGit then it would be perfect
ma27 has joined #nixos-dev
jtojnar has quit [Read error: Connection reset by peer]
jtojnar has joined #nixos-dev
ma27 has quit [Ping timeout: 265 seconds]
<shlevy> Think I found the locale issue...
<shlevy> copumpkin: The fetch-git I'm planning on adding to nix-fetchers will return a path, not a store path, so you can use filterSource on it
xeji has joined #nixos-dev
<Sonarpulse> shlevy: say one wants requireFile but actually requireDirectory
<Sonarpulse> how does one pick flat vs recursive hashing?
jtojnar has quit [Remote host closed the connection]
<shlevy> Sonarpulse: One modifies nixpkgs with a recursive argument :P
<shlevy> Or you can override the resulting derivation
<Sonarpulse> shlevy: ...hmm?
<Sonarpulse> what would I override?
<Sonarpulse> I mean i'd just make requireDirectory if I need to
<Sonarpulse> but I don't know how
<Sonarpulse> what the magic key-value pair would be
<shlevy> outputHashMode = "recursive"
<Sonarpulse> oh right
<Sonarpulse> thanks!
jtojnar has joined #nixos-dev
<shlevy> Ugh
<shlevy> localedef is just broken in 2.27 it seems
mbrgm_ has joined #nixos-dev
mbrgm has quit [Ping timeout: 268 seconds]
mbrgm_ is now known as mbrgm
<LnL> wait, I thought fetchurl { sha256 = null; } was removed
<jtojnar> why do we use /run/current-system/sw/lib/locale/locale-archive instead of one from the nix store?
<jtojnar> oh, right i18n.nix
<LnL> what the, I think my journald got stuck or something :o
<zybell> shlevy:Why arent locales treated the same as binaries:Every incompatible change gets a new dir?
<zybell> incompatible change="version"
zybell has quit [Ping timeout: 265 seconds]
zybell has joined #nixos-dev
xeji has quit [Quit: WeeChat 2.0]
Lisanna has quit [Quit: Lisanna]
Lisanna has joined #nixos-dev
<jtojnar> zybell: we want to be able to modify the locales using NixOS module without rebuilding the whole system
<zybell> please note that I meant the file *format* version, not the content.
<zybell> essentially a hashsum over the *loader*
jtojnar has quit [Ping timeout: 264 seconds]