Hi all, Here are my two cents. Disclaimer: I'm another Hackage Trustee and I care about this for the same reason Oleg does (specifically: the transitive closure of `diagrams`)

Would it be worth trying to consolidate similar data structures and have Chart + diagrams migrate to Only [1] (which is up to date)?

[1]: https://hackage.haskell.org/package/Only

I have suggested that here https://github.com/DavidM-D/indexed-list-literals/issues/1 Definitely in the longer term, I'd rather see a move to the Only package. But in the short term I'd find it much easier to get current packages working on GHC 8.4 if OneTuple supported 8.4. Oleg's changes look sensible to me, and adding a co-maintainer seems reasonable given that the last release was nine years ago and the sole maintainer seems uncontactable. My preferred course of action would be to add Oleg as a co-maintainer, and move downstream packages to `Only` over time. I intend to participate in the latter. Cheers, George

I ran into this issue in the taffybar built plan also and had to fix it locally since I didn't realize it was abandoned. So this isn't a fully localized issue. On Fri, Apr 13, 2018 at 9:14 AM, Oleg Grenrus wrote:

my Chart + diagrams usage depends on SVGFonts [1], which depend on tuple [2] which depend on OneTuple.

So if SVGFonts could drop a dependency on `tuple`, that would solve my usage problem too. CC:ing SVGFonts maintainers. In fact, dropping the dependency is easy in this case: https://github.com/diagrams/SVGFonts/pull/26

OTOH, Only's description says

NOTE: There is also the OneTuple package which by using a boxed data-type provides a 1-tuple type which has laziness properties which are more faithful to the ones of Haskell's native tuples; whereas the primary purpose of Only is to provide the traditionally so named type-wrapper for attaching typeclass instances.

So there's still some value in keeping OneTuple alive.

[1]: https://hackage.haskell.org/package/OneTuple [2]: http://hackage.haskell.org/package/tuple

On 13.04.2018 15:38, Ivan Lazar Miljenovic wrote:

...

On 13 April 2018 at 22:28, Oleg Grenrus wrote:

...

Hi all,

I tried to contact John Dorsey who is the sole maintainer of OneTuple [1], but the stated email address (haskell@colquitt.org) bounces back.

The last version of OneTuple was uploaded "by JohnDorsey at Tue Mar 24 17:35:51 UTC 2009" - 9 years ago. OneTuple was working great so far, but GHC-8.4.1 / base-4.11 introduced breaking change, which affects OneTuple too: Semigroup is now a superclass of Monoid. Therefore none of OneTuple versions is GHC-8.4.1/base-4.11 compatible.

I'm a Hackage Trustee [1] and also a user of OneTuple package (though not direct - it's in the transitive closure of Chart + diagrams), so I'd like to see GHC-8.4 compatible OneTuple version soon. Would it be worth trying to consolidate similar data structures and have Chart + diagrams migrate to Only [1] (which is up to date)?

[1]: https://hackage.haskell.org/package/Only

...

The required changes are very small, see https://github.com/phadej/OneTuple/pull/1 or more precisely this exact commit: https://github.com/phadej/OneTuple/pull/1/commits/ 1dbc4d3186f707034953e131995c590eb50cd9af

I'd like to become (co-)maintainer of OneTuple. I have no other plans than to keep it compiling with recent versions of GHC. Maybe I'll add NFData (from `deepseq`) instance though.

Best Regards, Oleg Grenrus

[1]: https://hackage.haskell.org/package/OneTuple

_______________________________________________ Libraries mailing list Libraries@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries

_______________________________________________ Libraries mailing list Libraries@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries