Cale 2017-03-09 03:45:23
(seq is primitive and ensures that its first argument is evaluated before resulting in its second argument)
bodisiw 2017-03-09 03:45:48
Athas, i admit i usually code for the 'happy path'... because i am lazy
Cale 2017-03-09 03:45:59
Yeah, it screws up the denotational semantics a bunch
bodisiw 2017-03-09 03:46:00
but haskell and i can't both be lazy ;-)
Athas 2017-03-09 03:46:26
bodisiw: well, that's fine. I think good language (and program) design requires compromises for practical reasons.
Athas 2017-03-09 03:46:45
Haskell only compromises in a very few places. Look at OCaml for a language that perhaps compromises a bit too much.
dredozubov 2017-03-09 04:06:44
basically i have a multi-package repo that i usually build with stack/nix
dredozubov 2017-03-09 04:07:09
nix have an override to build with -O2 for production builds
dredozubov 2017-03-09 04:07:29
but i'm wondering if i should set it in ghc-options in .cabal files
bennofs 2017-03-09 04:07:51
dredozubov: i don't think you should set it in .cabal files
dredozubov 2017-03-09 04:08:38
any explicit pro/cons?
ertes 2017-03-09 04:20:39
dredozubov: why do you even compile? i just use GHCi during development (or rather my emacs + haskell-interactive-mode does)
ertes 2017-03-09 04:37:04
this is my wrapper script (h-i-m fires up this one instead of "ghci"): https://github.com/esoeylemez/config/blob/master/bin/nix-ghci
dredozubov 2017-03-09 04:37:50
i have to run now, but i'll definitely look at it later, thanks!
ertes 2017-03-09 04:37:51
hopefully h-i-m will get first-class support for nix some day, so it will no longer be needed
