<{^_^}>
attribute 'runtimeShell' missing, at (string):494:1
<qyliss>
^ makeWrapper doesn't evaluate for that reason. curious.
justanotheruser has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos-dev
copumpkin has quit [Quit: Hmmm]
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos-dev
justanotheruser has quit [Quit: WeeChat 2.9]
rajivr has joined #nixos-dev
justanotheruser has joined #nixos-dev
lukegb has quit [Quit: ~~lukegb out~~]
lukegb has joined #nixos-dev
orivej has quit [Ping timeout: 240 seconds]
<Graypup_>
so i have a haskell executable that's having locale issues when i run it from outside a nix-shell eg with ./result/bin/klfc. anyone know how I can get the right env vars or such to get it to work properly?
<Graypup_>
i'm running nix from another distro, in case that complicates issues or such
copumpkin has quit [Read error: Connection reset by peer]
pmy has quit [Quit: WeeChat 3.1]
<siraben>
Will users get the new macOS installer?
jonringer has quit [Ping timeout: 250 seconds]
orivej has joined #nixos-dev
cole-h has quit [Ping timeout: 260 seconds]
b42 has left #nixos-dev ["WeeChat 2.9"]
orivej has quit [Ping timeout: 240 seconds]
maralorn has joined #nixos-dev
pmy has joined #nixos-dev
<maralorn>
Does anyone have opinions on creating subdirectories in maintainers/scripts? We are about to add multiple Haskell scripts and it would seem cleaner to me to meaningfully group them.
<maralorn>
* Does anyone have opinions on creating subdirectories in maintainers/scripts? We are about to add multiple Haskell related shell scripts and it would seem cleaner to me to meaningfully group them.
<maralorn>
One thumbs up on matrix. Good enough for me. ;-)
<maralorn>
Second question: Is anyone working on changing the rendering of the nixpkgs manual? I personally would very much like to move the haskellPackages documentation back into the nixpkgs manual, but at this point that would mean a not very small downgrade in usability.
pmy has quit [Read error: Connection reset by peer]
orivej has joined #nixos-dev
pmy has joined #nixos-dev
<DamienCassou>
hi
<DamienCassou>
I have a problem with a python package. When building it, I get "Collecting ofxparse@ https://github.com/jseutter/ofxparse/tarball/..." followed by network errors. I understand that nix won't give access to the network but I don't understand why building the package is fetching from network. Is there anything I have to do on the package definition to prevent network accesses?
<sterni>
maralorn: yeah ryantm is working on that, but depends on what you consider the downgrade specifically
<sterni>
maralorn: also I think creating a subdirectory makes sense
<sterni>
DamienCassou: most likely you'll need to check the source or the setup.py and see what it is doing
<sterni>
this is nothing that usually happens with python packages
<maralorn>
sterni: The primary downgrade is not to have any table of content for subsections of the haskell part.
<sterni>
maralorn: well I mean it is kind of a reimplementation of mmdoc, but with a reduced closure for bootstrapping purposes
<sterni>
maralorn: currently building the manual depends on pandoc and docbook which is quite a hefty closure you need to bootstrap just to build the manual
<sterni>
maralorn: with mdbook it's also pretty bad as you need the rust compiler
<sterni>
maralorn: would be happy to have a call of some sorts :)
<maralorn>
mmdoc ist a reimplementation of mdbook you mean?
<maralorn>
sterni: I meant for chatting in general.
<sterni>
well not a reimplementation, but it does fulfill the same purpose
<maralorn>
gchristensen: Is there an endpoint on hydra where I can get a list of all failed builds of an evaluation?
<thblt>
Sorry if it's a bit early, but I believe this recent and trivial PR of mine is ready to be merged, if anyone wants to have a look? https://github.com/NixOS/nixpkgs/pull/121246 Thanks!
<domenkozar[m]>
That's Nix that uses mdbook and this section is 4 levels deep
<maralorn>
How many layers of subsections will it support?
orivej has joined #nixos-dev
<domenkozar[m]>
I don't know, but we can probably change it?
<maralorn>
Ah, I didn‘t see all your replies on the phone client .
<maralorn>
Well, 4 levels might be enough. Although haskell4nix.readthedocs.io currently actually uses 6. (If you add the layers together with the fact that it would reside under Builders > Languages > Haskell > ...)
<ryantm>
I haven't looked lately but I think we were getting pretty close to converting everything in the nixpkgs manual to Markdown. There is still that big chunk of auto generated lib functions to figure out.
<ryantm>
mmdoc doesn't have a concept of nested sections. The document order is just the order the links appear in the manually made toc.md file.
<clever>
QQmlApplicationEngine failed to load component
<clever>
qrc:/main.qml:108:21: Type Button unavailable
<clever>
file:///nix/store/fvnw1b9sacbbcxar4zsxpxjlk2b4a3x3-qtquickcontrols2-5.15.2-bin/lib/qt-5.15.2/qml/QtQuick/Controls.2/Material/qmldir: plugin cannot be loaded for module ".nix.store.fvnw1b9sacbbcxar4zsxpxjlk2b4a3x3-qtquickcontrols2-5.15.2-bin.lib.qt-5.15.2.qml.QtQuick.Controls.Material": Module namespace 'QtQuick.Controls.Material' does not match import URI '.nix.store.fvnw1b9sacbbcxar4zsxpxjlk2b4a3x3-qtquickcontrols2-5.15.2-bin.lib.qt-5.15.2.qml.QtQuick
<clever>
i'm getting errors like this when running a QT app i just compiled in nix-shell, anybody know what the cause is?
<maralorn>
ryantm: That sounds perfectly flexible. Sound's great!
<maralorn>
Is there a builder which takes a tarball from the store and returns it unpacked?
<clever>
maralorn: it might be possible with fetchzip { url = "file://${./file.zip}"; }
<symphorien[m]>
sounds like ifd
<sterni>
stdenv has the unpackFile function which should do what you want
<clever>
otherwise, just use something custom like: runCommand "name" {} "unpackPhase ; mv $sourceRoot $out"
<clever>
sterni: that requires knowing the name, but unpackPhase will auto-detect the name, by comparing the files present before&after, and pop it into sourceRoot
<sterni>
yeah that's right :)
clerie has joined #nixos-dev
<lukegb>
clever: sounds like the QML2_IMPORT_PATH has ended up wrong, probably
<lukegb>
although we have a QtDeclarative patch to derive the path from the PATH so hm
<clever>
after brute-forcing my way thru it, i found that this fixes the problem for the app in question
supersandro2000 has joined #nixos-dev
<abathur>
repeat: Anyone want some glory? :) I think the merge of my installer PR lights a path to full uninstall/reinstall. I can't pursue it any time soon, but happy to discuss as needed. I wrote a proof-of-concept in Oct. that should be reheatable. Validation is also much easier since domen added installer tests.
<eyJhb>
abathur: I can't do anything, but is there a link to the PR?
<abathur>
there is (nix#4289), but most of it is out of the scope of the comment above, so I omitted because I don't think it'd be immediately obvious what's relevant about it without talking someone through it anyways
<gchristensen>
whew, just got the darwin builder up yet again. What happened is I'd manually bodged it to work the other day, and then on a daily basis it is reset to scratch ... and the bodge was undone :x
<maralorn>
<gchristensen "whew, just got the darwin builde"> That sounds exactly like the experience that nixos users love.^^
<gchristensen>
yup :D
<gchristensen>
the good news is the bodges were still safely in the nix store, waiting for me to apply them properly
cole-h has joined #nixos-dev
cole-h has quit [Quit: Goodbye]
jonringer has joined #nixos-dev
cole-h has joined #nixos-dev
<sterni>
you need to be a member of the whole hydra jobset to edit the configuration of a single jobset?
<gchristensen>
you get permission by owning the project
<sterni>
yeah that is unfortunate
<gchristensen>
oh?
AlwaysLivid has quit [Ping timeout: 240 seconds]
<sterni>
well I mean I'd be interested in changing the configuration for the haskell-updates jobset in the near future
<sterni>
but that would require me to be a member of the nixpkgs project it seems
<sterni>
but I'll guess I'll just poke someone when the time comes
<lukegb>
hmm
<lukegb>
firefox-esr segfaults when playing media at head
<lukegb>
qyliss: where's your PR status thingy again?
<qyliss>
it doesn't always update its Nixpkgs state correctly, because of permissions issues
<qyliss>
I have very limited disk space, so use a shared Nixpkgs on the server with multiple services accessing it, and getting the permissions right for that has been a nightmare
<qyliss>
I wish zfs had a bsdgroups option like ext4
<gchristensen>
would you like me to host it qyliss?
<puck>
qyliss: hrm, isn't this done by setting setgid on the directory?
<qyliss>
puck: it doesn't seem to be inherited in all situations
<qyliss>
gchristensen: interesting offer, let me think about it for a couple of days
<gchristensen>
sure
<puck>
qyliss: mm; i checked the zfs code and if you have setgid it seems to set it to the new directory
<gchristensen>
you could zfs clone the latest snapshot and chmod -r :)
<gchristensen>
I use this for doing pg dumps in CI without giving my CI runners loads of permissions
<qyliss>
puck: every time I think I've fixed it I come back a few days later and it has somehow lost group ownerships
<puck>
hrm
<qyliss>
gchristensen: pr-tracker will git fetch if it can't find the merge commit github tells it about
<qyliss>
it needs to be able to do this so it works for PRs that just got merged
<gchristensen>
yeah, it would be able to do that in its own clone
<gchristensen>
anyway, i don't need to explain it -- you've got a handleon it I'm sure
thblt has left #nixos-dev ["ERC (IRC client for Emacs 28.0.50)"]
<qyliss>
oh no I don't know any of this fancy ZFS stuff
<qyliss>
I barely know how to take a snapshot
<gchristensen>
ah so you can CoW a dataset to a new place with a new name and separate history
<qyliss>
but I feel like I shouldn't _need_ it, for this at least
<gchristensen>
probably shouldn't
<qyliss>
like this should all work
<qyliss>
git even has special setgid support
<qyliss>
... I wonder if maybe libgit2 does not though
<qyliss>
I don't remember if any of my services are using libgit2
<gchristensen>
you could create a zvol and put ext4 on it :D
<qyliss>
tempting
<qyliss>
clearly the best way forward is to just press ahead with Nix on NetBSD until I can switch out the kernel and reboot into that, where directory groups work sensibly :P
<endocrimes>
to make setgid work on zfs we had to setup things weirdly on our nas
<endocrimes>
(needed to enable some features/flags)
<qyliss>
endocrimes: ooh, could you find out what those were?
<endocrimes>
posixacls + recursively set facl
<qyliss>
acls were the next thing on my list to try
<endocrimes>
maxine set it up and is much better with filesystems than me lol
<puck>
huh, TIL aclinherit
<maxine>
I swear I know nothing about filesystems, but looking at my history, something like `setfacl -R -d -m g:users:rwx /home/systems; setfacl -R -m g:users:rwx /home/systems` is what did it for me, but tbh looking at it I'm unsure why just the first one wasn't enough
<symphorien[m]>
Iirc bsdgroups is the same as setgid on a directory
<endocrimes>
computers cursed
<qyliss>
symphorien[m]: supposedly, but that doesn't seem to be reflected in what's actually happening :(
<maxine>
Anyway, hope that helps. Of course the pool also has `acltype=posixacl` enabled, but I can try to actually remember if that was enough if it doesn't work
<qyliss>
I'll try that if I can't figure out what's going on with just setgid
<qyliss>
(and if I don't just get fed up and let gchristensen take over)
<symphorien[m]>
<qyliss "symphorien: supposedly, but that"> I think you need to set it on empty directories, it only affects newly created children
<qyliss>
I've set it recursively many times now
<qyliss>
hmm, actually it all seems to be working fine now
<qyliss>
lukegb: did you put a PR into it? Did it give you the right result?
<qyliss>
maybe whatever I did last time actually did fix it
<lukegb>
I did and I think so
<qyliss>
huh, neat
<qyliss>
well never mind then
Sigyn has left #nixos-dev ["Leaving the channel. /invite Sigyn #nixos-dev again if needed"]
rajivr has quit [Quit: Connection closed for inactivity]
<sterni>
anybody familiar with hydra jobsets up for reviewing #121407?