quchen 2017-02-27 00:15:33
runeks: Did you try hitting ^C lots of times?
quchen 2017-02-27 00:15:33
It's my usual solution.
quchen 2017-02-27 00:15:44
I think newer GHC versions give you a stack trace instead of <>, no?
quchen 2017-02-27 00:17:02
Hm, nevermind, it does not.
merijn 2017-02-27 00:17:52
quchen: Sounds like something that'd make a good feature request?
_sras_ 2017-02-27 01:22:33
MarcelineVQ: I just set the -ddump-splices option from ghci
oisdk 2017-02-27 01:23:37
I've got a restricted monad which has a more efficient liftA2, liftA3, etc than the bind counterparts, but it can't use (<*>) because it can't store function types. I'm wondering if there's a way to use applicativedo notation somehow.
_sras_ 2017-02-27 01:23:46
MarcelineVQ: Got it!
cocreature 2017-02-27 01:23:46
_sras_: there is -ddump-to-file but I'm not sure if that works from inside ghci
MarcelineVQ 2017-02-27 01:27:25
I think I had -ddump-to-file set in my .cabal file and when I ran ghci it made a bit of a spew into my src/ dir
MarcelineVQ 2017-02-27 01:27:53
Could be a solved thing by now.
lyxia 2017-02-27 01:34:33
oisdk: I think you can redefine <*> adequately with some typeclass shenanigans
lyxia 2017-02-27 01:34:41
oisdk: what does your type and liftAn look like
lyxia 2017-02-27 01:35:05
Must liftAn be individually defined by hand
fendor 2017-02-27 01:37:37
is there a way to cleanly uninstall all packages from cabal and start over with a fresh installation?
merijn 2017-02-27 01:37:59
fendor: Which OS?
fendor 2017-02-27 01:38:05
solus
dcoutts 2017-02-27 01:38:08
fendor: rm -fr ~/.ghc ~/.cabal
merijn 2017-02-27 01:38:12
^^
oisdk 2017-02-27 01:38:15
Yeah, each liftAn has to be written by hand currently. I can write a function which works on heterogeneous lists (https://gist.github.com/oisdk/a29fbbc0f5e94720d062f57389ff8c30) and define all of the applicative operations from that, but I'm wondering if there'a a better way
fendor 2017-02-27 01:38:16
ok, did that, did not work
fendor 2017-02-27 01:38:47
i mean, installing stack by cabal install stack did not work in Solus, but this seems to be a problem with the OS
dcoutts 2017-02-27 01:38:48
fendor: then you'll have to be more specific, since that does remove all the state
cocreature 2017-02-27 01:40:28
there can be packages in /usr/lib/ghc-$version
fendor 2017-02-27 01:40:48
some c - librarys seem to be faulty
dcoutts 2017-02-27 01:40:48
ok, so you're talking about stack rather than cabal
fendor 2017-02-27 01:40:48
well, maybe, i'm not sure, i'm trying to install stack via cabal
cocreature 2017-02-27 01:40:48
fendor: what is the problem you're seeing that makes you think that the problem is a non-clean package database?
fendor 2017-02-27 01:41:04
i can't show the output right now, wrong device, but the installation fails, sometimes with the hint, that a library is missing, for example #include cant be found
cocreature 2017-02-27 01:41:25
limits.h is a c header, that has nothing to do with cabal
lyxia 2017-02-27 01:41:27
oisdk: I think you can redefine <*> to wrap your arguments in a sort of free applicative, and then apply the right version of liftAn at the end.
fendor 2017-02-27 01:42:14
and such libraries
cocreature 2017-02-27 01:43:06
fendor: I think you really need to show us the exact output for us to be able to help you
fendor 2017-02-27 01:43:13
ok, gimme a sec
cocreature 2017-02-27 01:43:44
(please use something like http://lpaste.net/ and don't start pasting a wall of text here)
