Felix Yan <felixonmars@archlinux.org> writes:
On 01/09/2017 07:26 AM, Magnus Therning wrote:
Well, the subject line says it all really. Does ArchHaskell still have a role in the Arch world?
Actually I am planning to make everything dynamic-linked in next GHC release, and kill all static libraries in the packages. This way haskell software will be one step closer to follow Arch packaging guidelines. The ArchHaskell packages, IMHO, will be more suited for Haskell development after the change.
But as you mentioned, if system-wide packages are no longer a need for development, the need for two set of packages is also gone.
Indeed! :)
On 01/09/2017 06:23 PM, Nicola Squartini via arch-general wrote:
1. How do you manage dependencies and rebuilds of Haskell packages without using cblrepo?
I still try to use cblrepo to track dependency versions, I just don't use it to generate PKGBUILDs directly.
I *love* hearing that it's being used. I'm just throwing this out there... if ArchHaskell goes the way of the dodo I'd be more than happy to continue working on cblrepo, and make it more suitable for your use case.
2. Shouldn't you track also Cabal's metadata revisions?
The metadata updates are mostly about dependency versions and bounds, so I usually update that directly in the PKGBUILD instead. Please correct me if I'm wrong here, though.
That is true, I think there is a document somewhere describing what kind of edits to metadata that is allowed via Hackage. Since cblrepo keeps track of ranges of dependencies itself I've found cases where it's really useful to pull xrevs into its database (cblrepo.db). However, this doesn't really mean that one has to release a new package on a new xrev (essentially tell pacman about the change to dependencies). If one is willing to move this knowledge from cblrepo to pacman in some other way (I'm guessing you're doing it manually, Felix) then that's fine. I'm just too lazy to do that ;) /M -- Magnus Therning OpenPGP: 0x927912051716CE39 email: magnus@therning.org jabber: magnus@therning.org twitter: magthe http://therning.org/magnus As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality. — Albert Einstein