I learned about the curry language server in the most recent episode of the Haskell Interlude podcast - thank you for developing this!
I managed to set it up with the ecode editor, but ran into an issue with the paths.json file. My understanding is as follows: By default the curry-language-server assumes that pakcs is used as a compiler and calls pakcs cypm deps --language-server to populate the .json file with the relevant paths. I'm using kics2 almost exclusively, though, so this failed. After some digging I realized that I can just call cypm deps --language-server manually, using the cypm executable that comes with kics2 for (seemingly) the same effect.
I was wondering now, why the curry-language-server does not just call cypm here without assuming a specific curry compiler? Wouldn't that be more robust?
I learned about the curry language server in the most recent episode of the Haskell Interlude podcast - thank you for developing this!
I managed to set it up with the ecode editor, but ran into an issue with the
paths.jsonfile. My understanding is as follows: By default the curry-language-server assumes that pakcs is used as a compiler and callspakcs cypm deps --language-serverto populate the .json file with the relevant paths. I'm using kics2 almost exclusively, though, so this failed. After some digging I realized that I can just callcypm deps --language-servermanually, using thecypmexecutable that comes with kics2 for (seemingly) the same effect.I was wondering now, why the curry-language-server does not just call
cypmhere without assuming a specific curry compiler? Wouldn't that be more robust?