Closed treeowl closed 3 years ago
Disclaimer: this is my personal opinion, not opinion as of bound
maintainer.
Really, with stack
and cabal new-build
out there, is it an issue? I, for example, use ~/.stack/global/stack.yaml
with the same resolver as in most projects (actually single, but huge one), so i can just say stack ghci
and play with over 300 packages from Hackage (to answer reddit or SO questions etc).
Let's solve the dependency problem by having tools to handle dependencies, not by removing dependencies. I beg you.
-1 on splitting. On Tue, Jun 6, 2017 at 5:39 PM Oleg Grenrus notifications@github.com wrote:
Disclaimer: this is my personal opinion, not opinion as of bound maintainer.
Really, with stack and cabal new-build out there, is it an issue? I, for example, use ~/.stack/global/stack.yaml with the same resolver as in most projects (actually single, but huge one), so i can just say stack ghci and play with ~300 packages from Hackage (to answer reddit or SO questions etc).
Let's solve the dependency problem by having tools to handle dependencies, not by removing dependencies. I beg you.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ekmett/bound/issues/61#issuecomment-306625457, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAQwmKX7wvhTf54JLSEN-9LfQSWn4CFks5sBccqgaJpZM4Nx8fX .
@treeowl I've just found myself having this problem and split out bound-simple
: https://hackage.haskell.org/package/bound-simple-0.1.0.0/candidate . It only requires 'transformers', and doesn't even impose template-haskell on users. Only works with GHCs that provide DerivingVia though.
That package has
-- |
-- Copyright : (C) 2013 Edward Kmett
-- (C) 2021 Marco Zocca (ocramz)
but
copyright: 2021 Marco Zocca
and no reference that it's split out of bound
.
It's ok to fork, but hiding the copyright is not nice (and deleting it completely would be the violation of the license).
Also, please add other-extensions: DerivingVia
to indicate that users need GHC which supports that. The tested-with: GHC ==7.10.3
is clearly not true.
I'm also curious what will happen with future GHCs, as then GHC.Generics
will have Generically
(https://gitlab.haskell.org/ghc/ghc/-/merge_requests/5726), I'd expected to see a compat package, rather then all packages defining their own Generically
.
Anyway, seems there is a solution on the way, so I'm closing this issue.
Thank you @phadej for the thorough review; it wasn't my intention to hide the copyright notice, just I didn't update the defaults in the Cabal file. Good spot regarding Generically; I didn't even know it's being moved into base.
bound
depends onprofunctors
to allow it to offer a few trivial bits of lensy glue. Would it be possible to get abound-core
package or similar that just doesn't bother? Even playing around withbound
is annoying because of the dependency.