Agreed on both points: LGTM; lets keep the CLC in the loop.

On 29 Jun 2023, at 08:16, Simon Peyton Jones wrote:

Looks good to me.

I have suggested that @monoidal opens a CLC issue. (This change the `base` API.)

Simon

On Thu, 29 Jun 2023 at 07:21, Joachim Breitner mailto:mail@joachim-breitner.de> wrote:

...

Dear committee,

Krzysztof Gogolewski in https://github.com/ghc-proposals/ghc-proposals/pull/596 amends the sized literals proposal so that GHC uses the syntax 42#Int8 instead of (intToInt8# 42#) in the stock derived Show instances for data types with sized unboxed numbers.

Implementation is also ready.

Now that we have a syntax for these literals it’s clear that we want to use them in the show instances.

There is no risk of breaking Show-Read-roundtripping, as Read instances cannot be derived (at the moment).

This is a change in behavior of Show, so we should think for a moment how this affects stability. But none of the libraries in libraries/, in particularly not base, are affected by this change, according to Krzysztof.

So I recommend to accept this.

Cheers, Joachim

-- Joachim Breitner mail@joachim-breitner.de mailto:mail@joachim-breitner.de http://www.joachim-breitner.de/

_______________________________________________ ghc-steering-committee mailing list ghc-steering-committee@haskell.org mailto:ghc-steering-committee@haskell.org https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-steering-committee

---

ghc-steering-committee mailing list ghc-steering-committee@haskell.org https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-steering-committee

This proposal could break a few things, in particular if someone made a Read instance for a type with sized integer fields (though considering how the Show instance looks, I'd probably not have gone that path personally, so I expect it to be quite rare). Probably, just in case, this should be advertised pretty loudly in the release note. Apart from that, this is a clear improvement. The author's patch even removes some net lines from GHC, which is always welcome. I'm in favour. On Sat, 1 Jul 2023 at 12:38, Joachim Breitner wrote:

Hi,

CLC appreciates the ping but says it’s out of scope:

https://github.com/haskell/core-libraries-committee/issues/183#issuecomment-...

Cheers, Joachim

Am Donnerstag, dem 29.06.2023 um 08:16 +0100 schrieb Simon Peyton Jones:

...

Looks good to me.

I have suggested that @monoidal opens a CLC issue. (This change the `base` API.)

Simon

On Thu, 29 Jun 2023 at 07:21, Joachim Breitner wrote:

...

Dear committee,

Krzysztof Gogolewski in https://github.com/ghc-proposals/ghc-proposals/pull/596 amends the sized literals proposal so that GHC uses the syntax 42#Int8 instead of (intToInt8# 42#) in the stock derived Show instances for data types with sized unboxed numbers.

Implementation is also ready.

Now that we have a syntax for these literals it’s clear that we want to use them in the show instances.

There is no risk of breaking Show-Read-roundtripping, as Read instances cannot be derived (at the moment).

This is a change in behavior of Show, so we should think for a moment how this affects stability. But none of the libraries in libraries/, in particularly not base, are affected by this change, according to Krzysztof.

So I recommend to accept this.

Cheers, Joachim

_______________________________________________ ghc-steering-committee mailing list ghc-steering-committee@haskell.org

https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-steering-committee

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/

_______________________________________________ ghc-steering-committee mailing list ghc-steering-committee@haskell.org https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-steering-committee

-- Arnaud Spiwack Director, Research at https://moduscreate.com and https://tweag.io.

Merged! Am Sonntag, dem 09.07.2023 um 21:17 +0200 schrieb Joachim Breitner:

Hi,

I only got positive feedback, so I’ll merge this in a few days.

Cheers, Joachim

Am Donnerstag, dem 29.06.2023 um 08:20 +0200 schrieb Joachim Breitner:

...

Dear committee,

Krzysztof Gogolewski in https://github.com/ghc-proposals/ghc-proposals/pull/596 amends the sized literals proposal so that GHC uses the syntax 42#Int8 instead of (intToInt8# 42#) in the stock derived Show instances for data types with sized unboxed numbers.

Implementation is also ready.

Now that we have a syntax for these literals it’s clear that we want to use them in the show instances.

There is no risk of breaking Show-Read-roundtripping, as Read instances cannot be derived (at the moment).

This is a change in behavior of Show, so we should think for a moment how this affects stability. But none of the libraries in libraries/, in particularly not base, are affected by this change, according to Krzysztof.

So I recommend to accept this.

Cheers, Joachim

-- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/