Slimmed down the nix distribution of booster #342
Conversation
goodlyrottenapple
changed the title
goodlyrottenapple
changed the title
| QuickCheck -old-random +templatehaskell, | ||
| any.StateVar ==1.2.2, | ||
| any.adjunctions ==4.4.2, | ||
| any.aeson ==2.0.3.0, | ||
| aeson -cffi +ordered-keymap, | ||
| any.aeson-pretty ==0.8.9, | ||
| aeson-pretty -lib-only, | ||
| any.ansi-terminal ==0.11.4, | ||
| ansi-terminal -example +win32-2-13-1, |
There was a problem hiding this comment.
The script scripts/freeze-cabal-to-stack-resolver.sh generates these flags so maybe we should not remove them?
Or is the nix-ified version not generating them?
There was a problem hiding this comment.
My local version of cabal removed those and i dont really have a way to run the exact version as on CI. Since we don't use the freeze file for anything anyway (as far as i know) and solely depend on stack.yaml in CI, it doesn't matter much IMO and in the next haskell backend ci update job it will likely put those back anyway...
There was a problem hiding this comment.
OK fair enough, if we run into problems with the script we'll notice it soon enough.
I thought we were using the freeze file in the nix build (which uses cabal)...
There was a problem hiding this comment.
We were before, but I have switched us back to using stack.yaml/stackage as source of truth for nix.
| @@ -28,5 +28,5 @@ done | |||
| echo "Running a request which gets stuck, with logTiming enabled" | |||
| ${client} \ | |||
| execute $dir/state-c.execute ${client_args} -O log-timing=true | \ | |||
| $JQ 'del(.result.logs[].time)' | \ | |||
| jq 'del(.result.logs[].time)' | \ | |||
There was a problem hiding this comment.
since the script can now be invoked locally, I've removed $JQ and assume jq is on local path
Copy of runtimeverification/haskell-backend#3677