robertkennedy 2017-02-27 14:45:13
It's just that I have two versions of every function, IE `fMaybe :: A -> Maybe B` and `fEither :: A -> Either String B`, and I'd thought to unify them through fails in lieu of Nothing and returns instead of Rights
Squarism 2017-02-27 14:45:59
So does this cabal file look legit? http://lpaste.net/353043
glguy 2017-02-27 14:47:39
Squarism: I wouldn't have expected you to need the -main-is flags in ghc-options in general, maybe something's going on there. And you're missing most of your package version bounds. The directory structure seems fine
Squarism 2017-02-27 14:48:20
glguy, ok. Ill look into that
robertkennedy 2017-02-27 14:50:04
glguy: do you encounter that idiom a lot, and does using an mtl really not add too much overhead?
robertkennedy 2017-02-27 15:05:59
Got it. The library uses the fail return, and I just instantiate a type downstream to accept the function
hboats 2017-02-27 15:34:43
is it reasonable to use STM with parallel strategies? Or is STM really for concurrency?
peddie 2017-02-27 15:35:29
hboats: the point of parallel strategies is you don't need to use STM to evaluate things simultaneously, isn't it?
peddie 2017-02-27 15:35:39
or really, explicit threads at all
hboats_ 2017-02-27 15:38:49
weird disconnection =/ In any case I was hoping that I could use STM to control writes to an array. I have a number of things that would write to the same array. Technically the issue is that the program is asynchronous, but I'm trying to implement some amount of multiprocessing in it
peddie 2017-02-27 15:39:53
hboats_: if you just want to calculate a bunch of array members in parallel, you can probably use parallel strategies. if you want to do some concurrent I/O or whatnot, you might need something else
fryguybob 2017-02-27 15:43:24
hboats_: STM doesn't really support transactional reads and writes to mutable arrays right now.
hboats_ 2017-02-27 15:44:19
Yeah, I was hoping there might be some simple solution I missed. I'm trying to gain some parallelism on an asynchronous cellular automata, and have seen very little success
Sornaensis 2017-02-27 15:44:30
is STM lock-free
Sornaensis 2017-02-27 15:44:54
I'm playing around with implementing a synchronous queue in haskell for a parallel computing course
