Holes in GHC

Chris Dornan chris at chrisdornan.com
Fri Feb 10 14:49:35 CET 2012

Hi Alkenade,

This is the way I do all of my Haskell programming, without any need for
language support. I never use undefined for anything else -- or head, any
other partial functions or partial case expressions. It may be because of
wiring or habituation, but I have assumed that this is the way folks program
Haskell. I suspect hardened Haskell hacks may not need it because they are
so habituated to doing it anyway.

If that is the case then it must be a great initiative! -- who cares about
the has-beens, the future is the important thing! 

I have too much on to help out much but I will try and provide feedback when
I can.


-----Original Message-----
From: glasgow-haskell-users-bounces at haskell.org
[mailto:glasgow-haskell-users-bounces at haskell.org] On Behalf Of Thijs
Sent: 10 February 2012 13:07
To: glasgow-haskell-users at haskell.org
Subject: Re: Holes in GHC

Hello all,

We've started a wiki-page discussing how this idea can be applied to GHC


There have already been a number of people who indicated they'd want to use
this, so feel free to use the page to leave your comments about how you'd
want to use it.

Any comments on the best way to implement it are welcome too. We are
currently working with the idea of named holes[1], for which we are not
using the method mentioned in our earlier mails, but we're working on an
implementation more similar to how implicit parameters work.
I.e. generating constrains when a hole is encountered, so the same name used
multiple times will be inferred correctly. However, these should propagate
differently from implicit parameters and should not create an error when the
constraint is not in the type signature.

Thijs Alkemade

[1] http://hackage.haskell.org/trac/ghc/wiki/Holes#Namedholes

Glasgow-haskell-users mailing list
Glasgow-haskell-users at haskell.org

More information about the Glasgow-haskell-users mailing list