hiro` 2017-03-02 00:46:41
Hey all, just noticed that in my global config the resolver is set to 3.16. I guess that's kinda old now. Is it worth setting it to something more up to date? Does this need to be done manually or is there way to make stack just pick the most recent one?
jle` 2017-03-02 00:49:26
hiro`: i'd just edit it in ~/.stack/global-project/stack.yaml
Theophane 2017-03-02 00:49:28
hiro`: have you tried editing ~/.stack/config.yaml ?
jle` 2017-03-02 00:49:41
i'm not sure if there's a way to make it update automatically, but i'm also not sure if that'd be a good idea
jle` 2017-03-02 00:49:55
but updaitng it is pretty simple
jle` 2017-03-02 00:50:09
and stack keeps track of things so that you can re-use the same installed packages across different snapshots/resolvers
jle` 2017-03-02 00:50:17
so you don't have to re-install everything every time you update your resolver
jle` 2017-03-02 00:50:35
(unless the resolvers use a different version of ghc of course)
hiro` 2017-03-02 00:56:09
I haven't edited yet, just wondered if it was worth doing / not doing. Updating to 8.3 now ...
iainhallam 2017-03-02 00:56:31
So the platform installer can't work with a path other than /usr/local/ghc..., but I would like to install to /opt/haskell.
iainhallam 2017-03-02 00:56:48
Anyone know a way to get it into the right place, or am I going to be compiling from scratch?
russellw 2017-03-02 00:57:16
When installing Haskell on Windows based on the instructions here https://www.haskell.org/platform/ step three, it says paste three lines into Cabal config. The default config file already contains lines assigning at least some of those variables to different values. Is that still intentional? Should I paste the lines from step three at the end of the file or what?
hiro` 2017-03-02 00:58:26
Are there any other steps that need to be followed when updating the global config?
hiro` 2017-03-02 00:59:11
Eg, stack setup, init, etc?
Onemorenickname_ 2017-03-02 01:07:40
i'm a newbie
Onemorenickname_ 2017-03-02 01:07:50
please dont kill me
Onemorenickname_ 2017-03-02 01:07:53
but
brynser 2017-03-02 01:08:03
hiro`: I think you'd need to run stack setup to download and install the appropriate GHC version
Onemorenickname_ 2017-03-02 01:08:21
is there a way to tell "Data.Set" : "I don't really mind the Ord of my keys, so just use the pointer address" ?
Tuplanolla 2017-03-02 01:09:05
What address is this, Onemorenickname_?
Onemorenickname_ 2017-03-02 01:09:27
Tuplanolla, I come from C
Tuplanolla 2017-03-02 01:10:02
You know the garbage collector moves things around, so even if you were to find the address of an object, it could suddenly stop working.
Onemorenickname_ 2017-03-02 01:10:17
Tuplanolla, I was just thinking about that
Onemorenickname_ 2017-03-02 01:10:24
I was typing "haskell garbage collection" in google :D
Tuplanolla 2017-03-02 01:10:51
We have mechanisms to pin objects, but those are only used with ffi.
Onemorenickname_ 2017-03-02 01:11:24
in haskell, how would you compare (x = 1:x) and (y = 1:y) ?
brynser 2017-03-02 01:11:40
The unordered-containers package has HashSet which doesn't use Ord
Onemorenickname_ 2017-03-02 01:11:44
(they were not made from the same expressions, just two results of functions calls happen to have these value)
Tuplanolla 2017-03-02 01:12:04
The comparison of infinite lists is semidecidable, so ideally you wouldn't compare them at all.
Onemorenickname_ 2017-03-02 01:12:16
"Deprecated: HashSet is deprecated. Please use Set instead.
Onemorenickname_ 2017-03-02 01:12:17
The HashSet is a type synonym for Set for backward compatibility. It is deprecated and will be removed in furture releases."
brynser 2017-03-02 01:12:35
Oh, I didn't know that, sorry
Onemorenickname_ 2017-03-02 01:12:55
Tuplanolla, I don't think comparing recursive lists (different from infinite lists) is only semidecidable
Onemorenickname_ 2017-03-02 01:13:06
you can do some graph thingy if you can compare reference
Onemorenickname_ 2017-03-02 01:13:17
oh
Onemorenickname_ 2017-03-02 01:13:21
is there physical equality in haskell ?
Onemorenickname_ 2017-03-02 01:13:40
still, this won't give me an ord, even if it exists
brynser 2017-03-02 01:13:43
Onemorenickname_: I think that deprecation warning is for a different HashSet
brynser 2017-03-02 01:13:58
I don't see it here https://hackage.haskell.org/package/unordered-containers-0.2.7.2/docs/Data-HashSet.html
Onemorenickname_ 2017-03-02 01:15:10
brynser, i just typed HashSet haskell in google and got https://hackage.haskell.org/package/hashmap-1.3.2/docs/Data-HashSet.html
Onemorenickname_ 2017-03-02 01:15:21
so i will pick your package then :)
brynser 2017-03-02 01:15:32
:D
Onemorenickname_ 2017-03-02 01:15:49
ddo you know if it can hash recursive lists ?
Tuplanolla 2017-03-02 01:17:28
Hashes or not, what you want is impossible.
Onemorenickname_ 2017-03-02 01:18:45
Tuplanolla, I can not have a structure holding recursive lists without duplicates ?
Onemorenickname_ 2017-03-02 01:19:07
it's a bit problematic for my program termination
JuanDaugherty 2017-03-02 01:20:33
working the FFI will assuage ur Cness
Onemorenickname_ 2017-03-02 01:21:38
hm, there has to be an efficient way to do it
brynser 2017-03-02 01:21:51
russellw: I just installed Haskell Platform on Windows and didn't see the lines in step 3 in my cabal config
Onemorenickname_ 2017-03-02 01:22:04
JuanDaugherty, the lists are built in haskell
JuanDaugherty 2017-03-02 01:22:06
interface with some hoary thing, the default xml tools like xerces usw maybe
Tuplanolla 2017-03-02 01:23:18
Hey, are you the Brick man, JuanDaugherty?
JuanDaugherty 2017-03-02 01:24:05
um, i don't think so
Tuplanolla 2017-03-02 01:24:49
There is another J. Daugherty then.
Tuplanolla 2017-03-02 01:25:02
What a coincidence.
JuanDaugherty 2017-03-02 01:25:06
there are indeed
Rodenbach 2017-03-02 01:25:14
Do lenses ship with ghc? Or are they an external lib?
JuanDaugherty 2017-03-02 01:25:19
even other Juan Daughertys
cocreature 2017-03-02 01:25:33
Rodenbach: external libs
Rodenbach 2017-03-02 01:25:36
k
Tuplanolla 2017-03-02 01:26:48
Are people with that name naturally drawn to Haskell or something?
JuanDaugherty 2017-03-02 01:27:17
Arthur Koestler, Roots of Coincidence
JuanDaugherty 2017-03-02 01:27:40
Jung on synchronicity
absence 2017-03-02 01:32:17
i'm using stack with lts-8.3 and get "GHC runtime linker: fatal error: I found a duplicate definition for symbol _js_decode_string" because it's defined in both aeson and json-stream, and then "ghc.EXE: panic! (the 'impossible' happened)". anyone seen this?
ertes 2017-03-02 01:37:52
Onemorenickname_: when programming haskell, you shouldn't think in terms of "stuff in memory"… you should think in terms of haskell's semantics
ertes 2017-03-02 01:38:12
x = 1:x; y = 1:y -- these are the same list
Onemorenickname_ 2017-03-02 01:38:54
ertes : yep, but that makes some thing highly impractical
Onemorenickname_ 2017-03-02 01:39:05
if i have to reason in terms of semantics :
merijn 2017-03-02 01:39:06
Well, sometimes you should think about stuff in memory, but only after you've learned how to do things without doing that :p
ertes 2017-03-02 01:39:27
Onemorenickname_: which things?
Onemorenickname_ 2017-03-02 01:39:27
i have a big list
Onemorenickname_ 2017-03-02 01:39:39
i have some function accessing multiple times items of the big list, and I want to make a small sublist from these access (without adding several time the same item)
Onemorenickname_ 2017-03-02 01:39:52
one solution would be to have boolean marking on the big list for each function call
ertes 2017-03-02 01:40:05
Onemorenickname_: you want sharing?
Onemorenickname_ 2017-03-02 01:40:06
and i don't see any other, as I can not check uniqueness
Onemorenickname_ 2017-03-02 01:40:36
i don't know what is, so i will say no
Onemorenickname_ 2017-03-02 01:41:04
(the big thing is an array, the small thing are lists)
Onemorenickname_ 2017-03-02 01:41:17
(hence, I don't want something which require O(N) in space)
Onemorenickname_ 2017-03-02 01:42:01
and if i can't compare recursive lists between them (semi-decidable) and can not use physical equality
ertes 2017-03-02 01:42:02
Onemorenickname_: for example if you have a Map or a Set, and you create a new one based on it, the new one will not be an actual memory-copy of the old one
Onemorenickname_ 2017-03-02 01:42:15
ertes, ofc
hiro` 2017-03-02 01:42:17
brynser: Thanks. Actually stack ghci is still returning an error message, saying "Error parsing targets: The specified targets matched no packages."
ertes 2017-03-02 01:42:19
Onemorenickname_: the parts that didn't change will be shared, and that's fine in haskell, because values are immutable
Onemorenickname_ 2017-03-02 01:42:21
(I expect that from functional languages)
ertes 2017-03-02 01:42:34
Onemorenickname_: that's what i mean by sharing
Onemorenickname_ 2017-03-02 01:42:37
i see
Onemorenickname_ 2017-03-02 01:42:43
that's not my problem then
Onemorenickname_ 2017-03-02 01:43:23
I'm going to eat
Onemorenickname_ 2017-03-02 01:43:26
And then make a minimal example
Onemorenickname_ 2017-03-02 01:43:35
So it's easier to talk about it
ertes 2017-03-02 01:44:09
Onemorenickname_: yeah, that would be useful
