Phil de Joux

Can we please add predicate functions to Float that test for +Infinity, -Infinity and NaN. Haskell's GHC does it with [FFI calls](https://gitlab.haskell.org/ghc/ghc/-/blob/master/libraries/base/GHC/Float.hs#L1416-1426).

These comments from [primFloat.c](https://gitlab.haskell.org/ghc/ghc/-/blob/master/libraries/base/cbits/primFloat.c) say how to test for these: ```c /* Floating point NaN iff exponent is all ones, mantissa is non-zero (but see below.) */ /* A float...

The `GHC/*` modules are themselves within the `Language.Haskell.Liquid.*` tree so the warning is impossible to follow, as written when `GHC/*` modules depend on others within that subtree. ``` src> tree...

Spinning off a `ghc-liquid` package (#1970) for modules in `Language.Haskell.Liquid.GHC` would prevent imports from going the wrong way.

From looking at the github action logs for the cabal runner and comparing to local testing, it looks like this failure happens on mac but not on ubuntu.

I see from the help that reload reloads modules. Here I only explicitly loaded one, didn't I (from the command line arguments)? ``` :l[oad] {modulename} load module(s) :r[eload] reload the...

I'm hitting this **unable to expand name** message too with reproduction files at https://github.com/flight-earth/coriolis-effect/tree/bug/unable-expand-name/bugs/unable-expand-name. The message is not logged first time compiling (after deleting `.koka` folder) but it is on...

With the minimal reproduction I linked to, this problem seems harmless but I do have a more serious problem associated with the code I first struck this problem with, at...

I built koka from source on branch dev and can reproduce the builds first time around but not again with the source being compiled in a peer folder checked out...

If I change the type to int then I can match on the zero literal pattern no problem: ```diff -- type dm { con DM{ deg: int; min: double }...