Closed photm5 closed 9 years ago
@froozen I fixed the shell.nix problem you introduced in #36, does it still work for you?
EDIT: Oops, still not fixed... Did it really work for you?
Yeah, it worked perfectly fine.
Did you try forcing cabal to reconfigure using cabal configure
? This is where the problem happens, since it seems to forget the new hprotoc
version. (It errors since the old version doesn’t know the --lenses
flag.)
Both cabal configure
and cabal build
run fine for me.
Weird... Could you try if it still works at ca34250385ea2b81ad9803a7d9cb35c8c33c6e30?
Oh, wait a minute, I changed the network-anonymous-tor version by accident...
shell.nix
seems to be messed up in general.
Could you try out fa93becaac1a2f80e99ce795c26e96b77bd2c031 please?
It doesn’t help that GitHub spits out the whole commit log again whenever I change one single commit in the history.
fa93be
works fine for me. The expression looks disguting, though.
What is your nix-version? I'm on 1.10.
It does. When I first added this dirt to the codebase, I tried to refactor it myself (Using foldl, too), but failed horribly. You seem to be a bit more proficient with Nix, as you got it to work with your setup at least :)
I’m on 1.9. However, I think the problem is not caused by our Nix versions, but rather the NixOS channel (I didn’t upgrade for a while), since Nix itself tries to stay a lot more compatible than the nixpkgs are.
Could you try upgrading and see if it works, then?
Yes, it does work with the current Nix and Nix expressions. Finally removing that ugly Fix shell.nix
commit.
Great. Yay for higher-order abstractions!
Is there a plan to add lenses for those functions? It would make sense for a lot of them.
Do you miss any primitive or lens?
Looks good to me, you can merge. :+1:
Fixes #29. I think this is everything we need. Do you see something I missed?