Ivan Lazar Miljenovic
Ivan Lazar Miljenovic
@begriffs: @hvr has just merged a few PRs, so he'd probably be happy to accept one that fixes this issue.
I'd like this as well, working on projects with others were the formatting differs wildly; as such except for greenfield projects I'd like to disable it in some projects.
The problem with this is that it depends on what the underlying instances have for complexities. But yes, this would be a good thing to have.
So I've started doing so [here](https://github.com/haskell/fgl/blob/complexities/Data/Graph/Inductive/Graph.hs); what do you think? I'm not sure how feasible this will be though, as a lot of it starts to become "this is whatever...
Not that I know of. They also have similar complexities (IIRC, apart from `noNodes`, the only difference is that PatriciaTree has better constants albeit with a possible pathological case that...
I thought this was something I had fixed already (or at least had a valid reason not to because it broke some other associativity). If we're going to change this,...
Well, if you do a log of FGL work then you tend to remember its precedence along ;-) But yes, I mean code that assumes it's left associative (though I...
This is going to be a braking change, so it will need to wait until a major version release.
This will have to be a major version bump in case anyone is actually using/relying upon them. But I agree; the `unsafePerformIO` squicks me out.
Yes, a lot of FGL was written assuming no multiple edges. Unless you want to duplicate the function space, we can't change this until the next major version bump.