Hmm, where in neuron/rib is the $(nix-build)/bin.. pattern being suggested? (Given that our use of gitignore doesn't always behave correctly, that pattern should not be recommended).
Oh okay. It was just an one-off thingy (to be done once, or in scripts). Normally you would just run nix-build and use whatever is in the ./result/bin directory
I meant it also made the --version argument implementation more complicated. Thus I'm not sure what is the actual benefit of using the gitignore expression filter.
I found the gitignore nix expression rather confusing. It seems like this could be avoided by not using the
$(nix-build)/bin/...pattern, as suggested in https://github.com/Gabriel439/haskell-nix/blob/master/project0/README.mdHmm, where in neuron/rib is the
$(nix-build)/bin..pattern being suggested? (Given that our use of gitignore doesn't always behave correctly, that pattern should not be recommended).Oh well I've been using it since your comment on https://github.com/srid/neuron/pull/55#issuecomment-609045769 :).
Oh okay. It was just an one-off thingy (to be done once, or in scripts). Normally you would just run
nix-buildand use whatever is in the ./result/bin directorycf. https://github.com/srid/rib-sample#building-the-executable
I meant it also made the
--versionargument implementation more complicated. Thus I'm not sure what is the actual benefit of using the gitignore expression filter.To avoid unnecessary rebuilds, mainly. Also see: https://blog.hercules-ci.com/nix/2019/05/15/nix-gitignore/