1 patch for repository http://darcs.net:
patch 8978a9326874ed792b0d3c8c4ee1c95936963cb2
Author: Guillaume Hoffmann <guillaumh@gmail.com>
Date: Tue Jan 26 11:44:31 ART 2016
* bump second html dependency
I wonder if we can also require cabal 1.20 [1] and then rollback the
following patch to simplify Setup.lhs:
patch 850bd63d03fb36f10a0c999f35eb2ea8b44c5192
Author: Ganesh Sittampalam <ganesh@earth.li>
Date: Tue Jun 3 03:16:39 ART 2014
* conditionalise replHook using Template Haskell
[1] https://packages.debian.org/jessie/cabal-install
I'm using 1.18 in my GHC 7.6 environment, but there probably isn't any
good reason for that. But I'm also not sure it's worth removing this
too aggressively, unless someone needs to touch anything else in that
area.
In this case, more than just bumping lower bounds to reduce the possible
combinations of libraries to build darcs, I'm more concerned about
having code that's easy to understand.
I would say the change is worth it if we can assume that the typical
user of ghc 7.6 can get cabal 1.20 easily. Is it the case?
This is what the rollback patch would look like:
1 patch for repository http://darcs.net:
patch 5838a1d222e0979b5cf256d1406d133fa96a2f21
Author: Guillaume Hoffmann <guillaumh@gmail.com>
Date: Tue Jan 26 12:26:21 ART 2016
* require cabal 1.20 and rollback template haskell trick in Setup.lhs
I've checked and I can personally use Cabal 1.20 fine.
But the HP that came with GHC 7.6.2 had Cabal 1.16:
https://github.com/haskell/haskell-
platform/blob/master/hptool/src/Releases2013.hs
I think it'd be more logical to wait till our minimum GHC version is
7.10, unless it causes any actual problems. If anyone wants to
refactor Setup.hs, it'd definitely be a good candidate for dropping.