Profpatsch

Are you sure #800 has a solution for typed holes? It looks like it’s about custom builtins instead. But I like the `TODO` trick, maybe we should just put that...

Are the network requests parallelized? Every nth one taking 1–3s shouldn’t matter if we do 10–50 at once. Maybe there’s a way to take advantage of HTTP/2 here?

Personally, I *do* like the current dhall fmt behaviour, especially since dhall is by design a line-break and leading whitespace insensitive language. But I think this discussion is derailing the...

I found another one: ``` [ bins.emacsclient , "--create-frame" , "--eval" -- TODO: this obviously fails if the mail address contains " , "(url-mailto (url-generic-parse-url \"\${1}\"))" ] ``` loses the...

I’d just put `User-Agent: Dhall` and be done with it. `User-Agent` is a horrible misfeature in HTTP and should be purged.

I have to say that I pretty much implemented this blindly, I don’t know if there’s any golden tests that have to be corrected now, or if I missed any...

I kept as closely to the GHC version of z-encoding as possible, most of the code is copied verbatim (except I switched the function to `Text` and added a check...

Oh, another problem I found is that sum keys can be the same nearly arbitrary symbols in Dhall, but the nix representation of anonymous function keys can not: The dhall...

I added a second commit which should fix this problem. Now only function arguments and let-bindings should be z-encoded anymore, because those can’t really be reduced any further. Maybe there...

I haven’t found a way to make a `let` binding appear in the nix code though, maybe that is not even possible? They always normalize out in my examples and...