in my vim setup, i can hit '_i' (import) or '_im' (import module) on an Great. To late ;( I've now implemented this as well. My setup is getting the information
On Sun, Jun 17, 2007 at 03:39:25PM +0100, Claus Reinke wrote: directly from the installed packages (thus ghc-pkg describe package and then ghc -show-iface eachmodule.hi) Because I use caching it should be reasonable fast ? (I hope so, I still have to do some testing) The used packages are taken from the cabal file (and / or buffer setting)
identifier to get a menu of modules from which it could be imported; selecting an entry adds an import to the current module, either for the identifier only,
import Foo (a) add identifier b results in import Foo (a,b) ? or for the whole module it comes from (this feature
gets its information from haddock's indices I'm using tags and the plain source file most of the time. So I don't have trouble generating haddock documentation. I think its easy but I haven't spent much time on this. It contains all, the documentation and the implementation.
combined with quickfix mode, which will jump to the identifier not in scope, this seems to be quite close to what you ask for? I'll let you know when having finished the quickfix integration. Perhaps this can be done all automatically? http://www.cs.kent.ac.uk/people/staff/cr3/toolbox/haskell/Vim/ there's also '_.' which will use the same data to suggest possible fully qualified identifiers for the id under cursor. This might be useful, too
HaRe also had some transformations for manipulating imports/exports. http://www.cs.kent.ac.uk/projects/refactor-fp/hare.html I don't know HaRe at all. Would this even work when working with #ifdef etc?
I'd like to have some features, such as _T .. But this only works if the file can be parsed by ghc, right? Thanks for your suggestion. Marc Weber