Sam A. Horvath-Hunt
Sam A. Horvath-Hunt
I think the above would actually compile to `number` via intersection of arguments, making it safe. ```ts string & (string | null) = string ```
If #156 happens this mightn't be necessary.
Could also consider an eDSL via free monads? More ergonomic with do notation than patching strings together.
We stopped taking stdin [here](https://github.com/unsplash/intlc/commit/bac9a042ecca15ae35fd743e59ac6c6650206e96): > This is closer to the likely eventual interface of receiving a glob of files. Which is referencing #25. I think we could add stdin/stdout...
This relates a bit to #129.
This'd also make golden test output diffs more legible.
I guess one way to implement this would be to convert the JS `Compiler` from `Reader Cfg` to `RWS Cfg () State` where `State` tracks indent level. Then wherever there's...
This fellow thinks there's no point to supporting offsets: https://stackoverflow.com/a/58243789/3369753 I am struggling to see the use case.
For anyone on NixOS, you can patch the binary as a workaround with `patchelf`. The only missing shared library is `gmp`. ```sh patchelf \ --set-interpreter "$(cat $NIX_CC/nix-support/dynamic-linker)" \ --set-rpath /path/to/gmp/lib...
For anyone using Nix here's an example derivation including a fix: ```nix # This fetches the intlc binary and, if on Linux, patches its dynamic links. See: # https://github.com/unsplash/intlc/issues/99 {...