Transpilation is currently pretty slow, especially because of the many round-trips necessary when making roundtrips from ghcid. I should investigate ways to speed up this process, if possible. (It's worth noting that, as far as I can tell, profiling is particularly difficult in this case due to ghcid's multithreadedness.)
Would using the GHC API directly (e.g. via ghc-lib) improve performance? (That is, how much overhead does GHCi(d) have?)
Transpilation is currently pretty slow, especially because of the many round-trips necessary when making roundtrips from ghcid. I should investigate ways to speed up this process, if possible. (It's worth noting that, as far as I can tell, profiling is particularly difficult in this case due to ghcid's multithreadedness.)
Would using the GHC API directly (e.g. via
ghc-lib) improve performance? (That is, how much overhead does GHCi(d) have?)