Hi Michael, [fullquote for d-haskell@l.d.o] Am Samstag, den 24.11.2012, 20:10 +0200 schrieb Michael Snoyman: > Hi all, > I'm working on a new blog post: > https://github.com/yesodweb/yesodweb.com-content/blob/master/blog/2012/11/stable-vetted-hackage.md > > This is a follow up to my post two weeks ago regarding solving Cabal > hell with a stable set of packages selected from Hackage. The code I > mention in the article is still somewhat fresh, but seems to be > working. It's also based on the same code which I use for creating the > Yesod Platform and for running packdeps.haskellers.com, so it's not > completely untested. > As the blog post stands now, this is basically a single person's > effort, trying to get the community to join in. I think this project > has much more of a chance to succeed if the initial announcement has > more backing. By backing, I'm looking at two sides: > 1. Having more package maintainers signed up to include their > packages. > 2. Having some Linux distribution maintainers on board with (at least > trying to) use this new process for creating their set of packages. > For the first group: I've tested a build with the following additional > packages: > lens, aws, hjsmin, yesod-eventsource, base64-bytestring, > stylish-haskell, hlint, hoogle, acid-state, snap-server > Excluding some already existing test suite failures, everything builds > and tests correctly. So I'm asking for maintainers to sign their > packages up to be part of this list of stable packages. That > essentially means that you will keep your code working with the newest > versions of dependencies, and with the newest version of the Haskell > Platform. > This clearly isn't any form of binding agreement, and if the process > turns out to be too onerous, you can drop out at any time. However, > I'm hopeful that this will fit into the normal maintenance process > you're already performing, and having an automated system making sure > everything is running smoothly may in fact simplify your current > maintenance overhead. > To Linux distribution maintainers: I would imagine that a lot of > what's going on here is similar to your current processes. Being able > to all pool resources at a single point should ease everyone's > maintenance burdens, and having multiple people testing code should > mean that problems become apparent sooner rather than later. > Let me know if you have any questions, or if you'd like to sign up. If > you think there are other packages that should be included initially, > let me know as well. > Michael from the Debian POV, we would definitely benefit from such an approach, and are likely to look there for new versions to package, at least as long as your set of packages are reasonably up-to-date. For us, having complex set of packages (e.g. all yesod-related packages, or the various packages leading to algebra) included completely would ease the work on upgrading that, precisely because we would not get stuck half-way when we we find out that parts of the set are not compatible. Debian has a one-version-per-package policy, so naturally we would prefer if your hackage selected would also have the invariant. Thanks, Joachim -- Joachim "nomeata" Breitner Debian Developer nomeata@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nomeata@joachim-breitner.de | http://people.debian.org/~nomeata
Attachment:
signature.asc
Description: This is a digitally signed message part