How to improve loading performance ?

[PierreR]

I am using dhall with dhall-openshift as such:

let oc = env:OC

let project =
      oc.Project::{
      , name = "cicd-doc-dev"
      , displayName = "cicd doc site"
      , requester = "cicd_cicd"
      }

in  project

env:OC is a freeze (remote) import to a package.dhall that looks like this:

let schemas = ./schemas/schemas.dhall

let openshift = { core = ./packages/openshift.dhall }

let func = ./func/func.dhall

in  openshift ∧ schemas ∧ func

and packages/openshift.dhall looks like this:

let Openshift =
      https://raw.githubusercontent.com/TristanCacqueray/dhall-openshift/v0.2.0/package.dhall sha256:5d6efe2ed15b3e4e263bafce9752c45c671a42bbefd4d483e87d40d41d089640

in    Openshift
    ∧ { List =
        { Type =
            { apiVersion : Text, kind : Text, items : List Openshift.Resource }
        , default =
          { apiVersion = "v1"
          , kind = "List"
          , items = [] : List Openshift.Resource
          }
        }
      }

Loading this file would take more than 20 minutes. Is there anything I can try to improve the situation ?

I am using

 → dhall-lsp-server --version
1.0.7

[Gabriella439]

@PierreR: This issue is probably more appropriate for the dhall-haskell repository because that i where the dhall-lsp-server implementation resides. This repository is for the VSCode client plugin, so I'll relocate this issue.

[Gabriella439]

I believe the simplest solution to this problem would be for the language server to maintain the cache in memory (already decoded) so that decoding costs only need to be paid once