Closed phadej closed 4 years ago
This seems plausible to me, although I'd prefer if @ekmett signed off on this before moving ahead.
I’m planning on putting a new generation of distributive into that code as well using the approach Aaron Vargo came up with. Once that is done, I’d be happier pulling the trigger on an adjunctions refactor, as it’ll rip about half of this package out, by making distributive as powerful as representable is today, but more efficient.
Sent from my iPhone
On Aug 29, 2019, at 7:22 AM, Oleg Grenrus notifications@github.com wrote:
@phadej commented on this pull request.
In hkd/hkd.cabal:
@@ -0,0 +1,61 @@ +cabal-version: 2.2 +name: hkd +version: 0 +synopsis: "higher-kinded data" +description:
- "Higher-kinded data" utilities. +homepage: https://github.com/ekmett/codex/tree/master/hkd#readme Yes. As implied, I didn't put too much work into human/documentation side of things yet.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
adjunctions
have been waiting for something to happen for 1,5 years (aaron’s stuff was merged in 2018).
I’d suggest this if there are no clear veto for this change, that we go forward with it now. And get current state of adjunctions released.
On 29 Aug 2019, at 21.49, Edward Kmett notifications@github.com wrote:
I’m planning on putting a new generation of distributive into that code as well using the approach Aaron Vargo came up with. Once that is done, I’d be happier pulling the trigger on an adjunctions refactor, as it’ll rip about half of this package out, by making distributive as powerful as representable is today, but more efficient.
Sent from my iPhone
On Aug 29, 2019, at 7:22 AM, Oleg Grenrus notifications@github.com wrote:
@phadej commented on this pull request.
In hkd/hkd.cabal:
@@ -0,0 +1,61 @@ +cabal-version: 2.2 +name: hkd +version: 0 +synopsis: "higher-kinded data" +description:
- "Higher-kinded data" utilities. +homepage: https://github.com/ekmett/codex/tree/master/hkd#readme Yes. As implied, I didn't put too much work into human/documentation side of things yet.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
The difference is I started actually writing the code in question recently and don’t particularly want to thrash on two big releases in rapid succession.
Sent from my iPhone
On Aug 29, 2019, at 12:32 PM, Oleg Grenrus notifications@github.com wrote:
adjunctions
have been waiting for something to happen for 1,5 years (aaron’s stuff was merged in 2018).I’d suggest this if there are no clear veto for this change, that we go forward with it now. And get current state of adjunctions released.
On 29 Aug 2019, at 21.49, Edward Kmett notifications@github.com wrote:
I’m planning on putting a new generation of distributive into that code as well using the approach Aaron Vargo came up with. Once that is done, I’d be happier pulling the trigger on an adjunctions refactor, as it’ll rip about half of this package out, by making distributive as powerful as representable is today, but more efficient.
Sent from my iPhone
On Aug 29, 2019, at 7:22 AM, Oleg Grenrus notifications@github.com wrote:
@phadej commented on this pull request.
In hkd/hkd.cabal:
@@ -0,0 +1,61 @@ +cabal-version: 2.2 +name: hkd +version: 0 +synopsis: "higher-kinded data" +description:
- "Higher-kinded data" utilities. +homepage: https://github.com/ekmett/codex/tree/master/hkd#readme Yes. As implied, I didn't put too much work into human/documentation side of things yet.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.
@ekmett does that involve rewriting parts of hkd
and/or including parts of it directly in distributive
(or adjunctions
?
Or to ask more directly, can hkd
be uploaded to Hackage soon?
@ekmett do you have a WIP of next generation distributive somewhere in public?
To be honest, after the feedback here, I got fairly frustrated and went off to do other things that I wouldn't have to fight my way uphill on.
@ekmett I apologize if my actions are perceived as negative, I only try to make adjunctions
master
be closer to the releasable state (e.g. if the release would be forced for external reasons).
If this is not the direction adjunctions
will be moving, just close this pull request.
That perhaps sounded way more negative than intended. I apologize.
Found another package in HKD space: https://hackage.haskell.org/package/czipwith-1.0.1.2/docs/Data-CZipWith.html
ping
As a one year anniversary of this PR, I'm dropping this initiative.
The only "drawback" is that we only support GHC-7.6 and above, also for
adjunctons
. But who can live withoutPolyKinds
nowadays? (The real reason is that we cannot exporttype (~>)
on older GHC).ping @ekmett @RyanGlScott
hkd
defines natural transformations stuff, which is overlap withnatural-transformation
package.Nat
inhkd
doesn't have$$
field accessor. I didn't change that.I might myself need
hkd
soon too, so having it uploaded to Hackage would be great. I can be a co-maintainer.If this direction is approved, I'd write haddocks for
hkd
before upload.