angerman changed the topic of #haskell.nix to: https://input-output-hk.github.io/haskell.nix - alternative haskell infrastructure for nix; logs at https://logs.nix.samueldr.com/haskell.nix
sorki has joined #haskell.nix
srk has quit [Ping timeout: 268 seconds]
sorki is now known as srk
puffnfre- has joined #haskell.nix
puffnfresh_ has quit [Ping timeout: 256 seconds]
angerman has quit [Ping timeout: 272 seconds]
angerman has joined #haskell.nix
pie_ has quit [Remote host closed the connection]
pie_ has joined #haskell.nix
hekkaidekapus has joined #haskell.nix
hekkaidekapus} has quit [Ping timeout: 268 seconds]
__monty__ has joined #haskell.nix
fendor has joined #haskell.nix
fendor_ has joined #haskell.nix
fendor has quit [Ping timeout: 256 seconds]
fendor has joined #haskell.nix
fendor_ has quit [Ping timeout: 264 seconds]
aveltras has joined #haskell.nix
<ocharles> https://input-output-hk.github.io/haskell.nix/tutorials/materialization/ Says I should see something like "Get `plan-sha256` with `nix-hash --base32...", but when I run with an `index-state`, I get:
<ocharles> "To make project.plan-nix for pandoc a fixed-output derivation but not materialized, set `plan-sha256` to the output of the 'calculateMaterializedSha' script in 'passthru'."
<ocharles> This is haskell.nix at 505130bfdaa9ecff5f50c4c415d4cb7db7089716
<ocharles> (That's basically HEAD)
<ocharles> are the docs out of date?
<__monty__> That's pretty likely. Maybe check when the doc was written : )
<ocharles> Hum. I can't even find a `calculateMaterializedSha` attribute on the result of a `hackage-package` call
<ocharles> The message talks about `passthru`, but that also doesn't exist
<__monty__> Yeah, I believe there's open issues about this. It should exist after a successful build but this has definitely come up before. (Just reporting on chat history.)
<ocharles> I don't mean it doesn't exist, I mean I can't even find the attribute :)
<ocharles> nix-build -E '(import ./nix/pkgs {}).pandoc.passthru.calculateMaterializedSha'
<ocharles> error: attribute 'passthru' missing, at (string):1:1
<ocharles> But yea, every time I look at this there seem to be issues around materialization so I never know if it's meant to work or not
<__monty__> I'm not familiar with the passthru mechanism. Not sure it should exist before a dependency is built that will end up being passed through.
<__monty__> IOHK is definitely using materialization. I haven't looked into getting it to work with my projects yet though.
<ocharles> Building pandoc with haskell.nix seems to crash our Hydra, so I'm just trying to see if materializing it would fix that
<ocharles> Hydra is useless with its error messages, I just get "unexpected EOF reading a line"
<__monty__> And building locally doesn't show any errors?
<ocharles> Nope
<ocharles> It's hydra-eval-jobs that's crashing, so something at evaluation time
<__monty__> Hmm, yeah your theory that too much IFD might be causing this does sound plausible.
<ocharles> I think I'm gonna step away from this for a bit, but I'll check back later
<ocharles> Right now I just want to throw this computer out the window
<ocharles> Not the best state to debug in ;)
craige has quit [Quit: ZNC - http://znc.in]
<michaelpj> ocharles: yeah those docs are out of date. It's a perpetual hassle telling people how to get the right attribute... calculateMaterializedSha is an attribute on the result of a `cabalProject` call, which everything is eventually. So if you use `hackage-project` you should get it, but some of the wrappers don't propagate it, which is one of the reasons I don't like `tools` :p
<ocharles> Ah, there's hackage-*project* too!
<ocharles> I do thinkk I agree with you and these "helpers" are actually just bloating the api and making things harder
<__monty__> Maybe a set of projection functions like "exes" would be a good alternative? Problem is I think that's exactly what was migrated away from in the first place?
<__monty__> Or was the problem that these projections were actual attributes of components?
<ocharles> hackage-project also doesn't have calculateMaterializedSha though. I'm doing:
<ocharles> self.haskell-nix.hackage-project {compiler-nix-name = "ghc8101"; name = "pandoc"; version = "2.11.3.2"; index-state = "2021-01-19T00:00:00Z";};
<ocharles> > builtins.attrNames pkgs.pandoc
<ocharles> [ "ghcWithHoogle" "ghcWithPackages" "hsPkgs" "index-state" "makeConfigFiles" "pkg-set" "plan-nix" "projectArgs" "projectCoverageReport" "projectCross" "projectFunction" "roots" "shellFor" "tool" "tools" ]
<ocharles> Ok, I think it's pandoc.plan-nix.passthru.calculateMaterializedSha
<ocharles> thanks michaelpj
<michaelpj> if you squint, the message does kind of say that ;)
<michaelpj> but yeah, the ergonomics around here are not good still :(
<ocharles> well it doesn't say to go into `plan-nix` anywhere, unless I missed that
<ocharles> Hurrah, I have successfully materialized the plan
<michaelpj> "To make project.plan-nix..."
<ocharles> Aha!
<__monty__> Hmm, any better way to run the updateMaterialized than trying to build and copying the path?
<__monty__> calculateMaterializedSha builds fine and then you can use result but by setting materialized you prevent building updateMaterialized because there's no materialized files yet.
<ocharles> Happy days, that has restored Hydra
aveltras has quit [Quit: Connection closed for inactivity]
pjb has quit [Read error: Connection reset by peer]
pjb has joined #haskell.nix
NinjaTrappeur has quit [Quit: WeeChat 2.9]
NinjaTrappeur has joined #haskell.nix
__monty__ has quit [Quit: leaving]
hekkaidekapus has quit [Remote host closed the connection]
hekkaidekapus has joined #haskell.nix