[ghc-steering-committee] Type annotated quoters (#125) --- recommendation: accept

[Simon Peyton Jones]

My current POV is that I think this feature is far far more limited than I originally thought -- and hence unattractive.  (I’ve posted on the discussion thread to that effect.)

I’m willing to be educated!

Simon

From: ghc-steering-committee <ghc-steering-committee-bounces at haskell.org> On Behalf Of Sandy Maguire
Sent: 11 July 2019 20:10
To: ghc-steering-committee at haskell.org
Subject: [ghc-steering-committee] Type annotated quoters (#125) --- recommendation: accept

Hi all,

Proposal 125 is on the topic of making quasiquoters which annotate the type of expression they produce. For example:

  qq :: TQuasiQuoter Int

which would work when used via

  foo :: Int
  foo = [qq|| blah ||]

but would be a type error (even before expanding the splice) in the context of

  bar :: String
  bar = [qq|| blah ||]

I'm in favor of this proposal. It's small, a natural extension to typed expressions+quasiquoters, and solves the very real problem of statically verifying literals at compile time.

The proposal is here: https://github.com/ghc-proposals/ghc-proposals/pull/125
and my recommendation is: https://github.com/ghc-proposals/ghc-proposals/pull/125#issuecomment-510612026

I have recommend a syntactic change that is *not* present in the original proposal text. It seems unfair to make the proposer jump through more hoops after a year of inactivity on this proposal.

As usual, silence will be considered assent!

Best,
Sandy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-steering-committee/attachments/20190719/2e620161/attachment-0001.html>