Closed tcr closed 7 years ago
Will need to land after #22.
I've rebased and updated the script a little (no rm -rf stack-work
, since that is unnecessary I think, stack/cabal will find changed files after the update).
I've also pushed another change to the alex/happy repos to make the template output system independent, so that we don't have to commit spurious changes like the "/tmp/ghc..." file pragmas.
Great. I'll keep tabs on how well rm -rf ./stack-work
continues to work—some combination reinstalling ghc and running these had issues.
This adds the two alex-rust and happy-rust forks. The root
./regen.sh
script will pull those submodules, then regenerate the parser/lexer files. It's very aggressive with rebuilding but it leads to a reproducible outcome. I imagine in development, it might be easier to just generate the output directly from happy/alex forks and commit it rather than using this script.I also did a cursory look at moving parser-c-core into its own crate, but the dependencies are so intertwined that I don't think that is going to happen, unless we invert the dependencies (which seems like it won't avoid long compile times.)