I'd like to improve my release process to avoid problems like this. I could re-release reactive-glut & reactive-fieldtrip whenever there's a new reactive release, and re-release reactive-fieldtrip whenever there's a new reactive-glut. While manageable for me, other people who build on any of these libraries--or *any* library--will have the same problem. I wonder what's causing the fragility here. Any ideas? - Conal On Wed, Dec 17, 2008 at 8:56 AM, Patai Gergely wrote:

...

However, when I try to compile Tetris or load it into ghci, it says that it can't match 'UI' (expected) with 'reactive-glut-0.1.4:FRP.Reactive.GLUT.UI.UI' (inferred) in dropPiece, and I've no idea where it is getting the latter from. There are no leftover .hi files or anything else in reach. Is reactive-fieldtrip a possible culprit? Yep, that was it. I reinstalled reactive-fieldtrip, and it works again.

So this is happening when people do cabal-install? I'm rather surprised it doesn't run through the installed libraries that it knows about and checks if they have a dependency on an updated package, recompiling if they do. But the dependency says "reactive >= 0.10.2, reactive-glut >= 0.0.4", which certainly doesn't change due to the update. Isn't linking to a specific version a limitation imposed by ghc itself?

-- http://www.fastmail.fm - A fast, anti-spam email service.