Simon Marlow

> Do the tests exercise the different versions of the code that are run/produced for different Happy settings, such as "use (or not use) GHC-specific extensions"? Yes!

@kindaro Happy is non-standard because it is bootstrapped. See the README.md file for instructions on how to build: https://github.com/simonmar/happy/blob/master/README.md

> @simonmar What you are saying is that this package is not going to be handled by new-build? No, not at all! It works fine with new-build. But you have...

Oh, I got confused between Alex and Happy. Alex is the one with the build instructions in the readme, it looks like we haven't done this for Happy yet. But...

Have you tried it? If it works, it's a feature!

Ok, thanks for the suggestion then. I don't see any reason why it couldn't be done.

This relies on the `interruptible` extension to the FFI to allow the wait to be interrupted by the async exception. See https://downloads.haskell.org/~ghc/latest/docs/html/users_guide/ffi-chap.html#interruptible-foreign-calls It works on Linux (I just checked), but...

Richard, if I understand correctly, you are proposing that `let! x = undefined` would behave differently from `let !x = undefined`, right? That would be subtle and surprising, perhaps even...

But in the absence of the special case, `let !p = e in ...` is useless because it means exactly the same as `let p = e in ...`. So...

The `Pretty` library runs in two modes: the normal mode where it is nicely indenting things, and a fast mode where it doesn't indent anything. The fast mode is used...