Ancestors

Toot

Written by Philippa Cowderoy on 2024-11-11 at 22:18

When testing a new parser in Haskell from ghci, make sure all the result types have Show instances...

(one day I won't relearn this lesson)

=> More informations about this toot | More toots from flippac@types.pl

Descendants

Written by Philippa Cowderoy on 2024-11-11 at 22:19

I also found a simple mistake in my parsing combinator library, but thankfully it was pretty quick to track down by narrowing down in ghci which part of the grammar was erroneously failing and then looking at the main parse function

Fucked up handling the "no matches" case for */many - had all the infrastructure for distinguishing "succeeded empty", then failed instead! 🤦‍♀️

=> More informations about this toot | More toots from flippac@types.pl

Written by Philippa Cowderoy on 2024-11-11 at 22:23

Trying to decide if I'm likely to stay up late enough anyway that I might as well write the technically-compiler from my AST to stuff in my abstract machine monad. There might be one or two decisions too many even doing it the naive way to not be asking for it, but if I pull this off and add the "primitives" map I've got a working interpreter (though far from a pleasant repl)

=> More informations about this toot | More toots from flippac@types.pl

Written by deech on 2024-11-11 at 22:36

@flippac if it makes you feel any better from personal experience this specific problem occurs in companies with budgets in the tens of millions

=> More informations about this toot | More toots from deech@mastodon.social