Logs: liberachat/#haskell
| 2021-08-03 18:35:27 | tomsmeding | meddles in stuff |
| 2021-08-03 18:35:29 | chris | is now known as Guest3788 |
| 2021-08-03 18:36:16 | <DigitalKiwi> | tom smeding is medling |
| 2021-08-03 18:36:40 | → | Guest38 joins (~Guest38@2a02:3035:807:4e8c:ec4d:e1f6:4f08:1e6e) |
| 2021-08-03 18:37:05 | <tomsmeding> | or middling? |
| 2021-08-03 18:37:29 | <DigitalKiwi> | now you're definitely medling |
| 2021-08-03 18:38:12 | × | Guest38 quits (~Guest38@2a02:3035:807:4e8c:ec4d:e1f6:4f08:1e6e) (Client Quit) |
| 2021-08-03 18:38:13 | × | forell quits (~forell@user/forell) (Quit: ZNC - https://znc.in) |
| 2021-08-03 18:38:16 | → | goepsilongo joins (~chacho@2603-7000-ab00-62ed-0000-0000-0000-0bd0.res6.spectrum.com) |
| 2021-08-03 18:38:23 | <DigitalKiwi> | https://www.thefreedictionary.com/meddling |
| 2021-08-03 18:38:30 | <DigitalKiwi> | :( |
| 2021-08-03 18:39:11 | × | burnsidesLlama quits (~burnsides@dhcp168-023.wadham.ox.ac.uk) (Ping timeout: 252 seconds) |
| 2021-08-03 18:39:12 | → | mastarija joins (~mastarija@78-3-212-215.adsl.net.t-com.hr) |
| 2021-08-03 18:39:33 | → | Pickchea joins (~private@user/pickchea) |
| 2021-08-03 18:39:44 | × | Guest3788 quits (~chris@81.96.113.213) (Ping timeout: 252 seconds) |
| 2021-08-03 18:40:01 | <tomsmeding> | mixing thoroughly? That I surely do |
| 2021-08-03 18:40:29 | <DigitalKiwi> | this is like the time i misspelled emigrate with 3 m's |
| 2021-08-03 18:40:51 | <DigitalKiwi> | because i thought it had 2 but typoed an extra one lol |
| 2021-08-03 18:43:28 | <DigitalKiwi> | tomsmeding is a mojito |
| 2021-08-03 18:44:20 | <tomsmeding> | a small magic charm, that's a nice thing to be :) |
| 2021-08-03 18:45:16 | <DigitalKiwi> | A mojito is a Cuban rum-based cocktail. |
| 2021-08-03 18:45:55 | × | MoC quits (~moc@user/moc) (Quit: Konversation terminated!) |
| 2021-08-03 18:45:59 | <DigitalKiwi> | you medle the mint |
| 2021-08-03 18:46:16 | <DigitalKiwi> | with a muddler |
| 2021-08-03 18:46:18 | <DigitalKiwi> | lol |
| 2021-08-03 18:46:59 | <tomsmeding> | "Origin: 1930s Cuban Spanish, from mojo + diminutive suffix -ito." -- mojo: "A magic charm, talisman, or spell." |
| 2021-08-03 18:47:06 | <DigitalKiwi> | and then your bartender beats you over the head with it; or maybe that's just me ;( |
| 2021-08-03 18:47:07 | <tomsmeding> | ok I'll stop |
| 2021-08-03 18:47:29 | <DigitalKiwi> | one of my favorite bars to order mojitos at was called mojo smokehouse and ales |
| 2021-08-03 18:49:41 | → | drd joins (~drd@93-39-151-19.ip76.fastwebnet.it) |
| 2021-08-03 18:52:19 | <DigitalKiwi> | anyone want to talk about haskell |
| 2021-08-03 18:53:05 | <davean> | DigitalKiwi: Does it come with mojitos? |
| 2021-08-03 18:54:13 | <DigitalKiwi> | fresh out of rum all i have is tea ;( |
| 2021-08-03 18:54:36 | <DigitalKiwi> | err out of tea now too |
| 2021-08-03 18:55:06 | → | thyriaen joins (~thyriaen@dynamic-078-055-141-033.78.55.pool.telefonica.de) |
| 2021-08-03 18:56:01 | <DigitalKiwi> | davean: are you in a meeting today lol drakonis and you and i were talking about guix... |
| 2021-08-03 18:56:12 | <drakonis> | i am clearly not in a meeting |
| 2021-08-03 18:56:53 | <davean> | I like tea, how about we talk about tea |
| 2021-08-03 18:57:05 | <davean> | DigitalKiwi: Haha, I should be but I'm not! |
| 2021-08-03 18:57:14 | <davean> | Whats the Haskell story on guix? |
| 2021-08-03 18:57:34 | <drakonis> | hmm |
| 2021-08-03 18:57:56 | <delYsid> | hmm, plain old [] vs Unboxed Vector and ST. I know the answer is probably "you have to try", but maybe you guys have some tipps before I start. I want to get the memory usage (GC) and runtime down of my chess legal move generator. Currently, it generates a [Ply] where newtype Ply = Ply Word16. I am suspecting the list is adding quite a lot to the memory usage, so I was hoping Unboxed Vector could be better? My plan: "new" a vector |
| 2021-08-03 18:57:57 | <drakonis> | the haskell story seems to be pretty fine right now? |
| 2021-08-03 18:57:57 | <delYsid> | with sufficient size, generate all moves "write"ing the entries to the vector and inc'ing an STRef, and slice the used part of the MVector off at the end. Can this be faster/better then plain "genMove : ms" way of just adding to head? |
| 2021-08-03 18:58:01 | <drakonis> | they recently upgraded ghc |
| 2021-08-03 18:58:02 | <DigitalKiwi> | can i do a guix command ot pull in a ghci with an hackage packges |
| 2021-08-03 18:58:06 | <drakonis> | yes |
| 2021-08-03 18:58:12 | <drakonis> | ish |
| 2021-08-03 18:58:21 | <davean> | drakonis: Of course you don't have to try, this can be directly computed |
| 2021-08-03 18:58:41 | <drakonis> | there isnt a single file containing all of hackage |
| 2021-08-03 18:58:51 | <davean> | drakonis: and yes, a vector would be far smaller for any significant sized list |
| 2021-08-03 18:58:57 | <drakonis> | wrong ping btw |
| 2021-08-03 18:59:13 | <davean> | bah yah |
| 2021-08-03 18:59:17 | <DigitalKiwi> | if i learn racket will that help with guix heh |
| 2021-08-03 18:59:21 | <drakonis> | probably? |
| 2021-08-03 18:59:32 | <drakonis> | at least with getting acclimated to a lispy lang |
| 2021-08-03 18:59:45 | <davean> | drakonis: so whats this "ish"ness? |
| 2021-08-03 18:59:51 | <DigitalKiwi> | https://www.coursera.org/learn/programming-languages has three parts and racket is one of them lol |
| 2021-08-03 18:59:57 | <delYsid> | davean: I guess you were talking to me? |
| 2021-08-03 19:00:02 | <davean> | delYsid: I was |
| 2021-08-03 19:00:09 | → | Testtesttesz joins (~Testtestt@p5b05e91c.dip0.t-ipconnect.de) |
| 2021-08-03 19:00:18 | <drakonis> | the ishness is that there isnt a kitchen sink containing every hackage package |
| 2021-08-03 19:00:28 | <sm> | srid: that looks nice! |
| 2021-08-03 19:00:43 | <drakonis> | you'll definitely have a sizeable amount of packages but not everything |
| 2021-08-03 19:00:47 | <monochrom> | If list laziness is important to your algorithm, vector will kill it. |
| 2021-08-03 19:00:57 | <drakonis> | you can invoke a hackage importer if there's anything you dont have available |
| 2021-08-03 19:01:02 | <drakonis> | you want that isnt packaged yet |
| 2021-08-03 19:01:26 | <davean> | drakonis: I use to program scheme, back around when the Brown group was renaming stuff, its a nice, small language, but I moved on because I thought it ended up too messy with too much local semantics - how varies are the semantics in guix's implimentation and how much do they have 1000 options with a horrible combinatorial number of interactions for functions? |
| 2021-08-03 19:01:51 | <drakonis> | clean, really. |
| 2021-08-03 19:01:58 | <delYsid> | monochrom: Thats another argument against [], I definitely have a finite result, and it typically only has aroun 20 elements. |
| 2021-08-03 19:02:10 | <drakonis> | guix runs on top of guile |
| 2021-08-03 19:02:23 | <davean> | drakonis: well I'm less worried about containing, and more about solving |
| 2021-08-03 19:02:33 | <drakonis> | hmm |
| 2021-08-03 19:02:35 | <drakonis> | i see |
| 2021-08-03 19:02:43 | <delYsid> | So my thinking was, 20*2 = 40 bytes for an Unboxed Vector of Ply is probably *a lot* less then [Ply] with 20 elements memorywise? |
| 2021-08-03 19:02:58 | <drakonis> | 1000 options with a horrible combinatorial explosion? |
| 2021-08-03 19:03:01 | <DigitalKiwi> | (in that course the SML exercises were easy to covnert to haskell too i did them in both lol) |
| 2021-08-03 19:03:08 | <drakonis> | i'm not aware of that |
| 2021-08-03 19:03:12 | <davean> | I'm happy to add the packages I care about in some local overlay - if they have a good story for that - but I need it to get consistent sets, and I'm interested in how it manages said sets |
| 2021-08-03 19:03:29 | <drakonis> | oh they have a really good story for that |
| 2021-08-03 19:03:31 | <drakonis> | btw |
| 2021-08-03 19:03:34 | <drakonis> | the best one |
| 2021-08-03 19:03:53 | <drakonis> | by far and wide the best one right now |
| 2021-08-03 19:03:59 | <davean> | drakonis: the 1000 options thing was an exageration, just that instead of composing functions, functions tended to take options to change their behavior, due to the nature of how scheme composes |
| 2021-08-03 19:04:06 | <davean> | drakonis: so what is the story? |
| 2021-08-03 19:04:21 | <drakonis> | i'll refer to the docs to this |
| 2021-08-03 19:04:22 | <monochrom> | Or rather, how scheme doesn't compose. |
| 2021-08-03 19:04:32 | <drakonis> | but it has channels |
| 2021-08-03 19:04:48 | <monochrom> | Scheme has paid only lips service to function composition and currying. |
| 2021-08-03 19:04:49 | <drakonis> | so, you want to roll out your own repo? you just add it to the channels list and pull |
| 2021-08-03 19:05:11 | <drakonis> | oh, has it? |
| 2021-08-03 19:05:13 | → | jippiedoe joins (~david@77-171-152-62.fixed.kpn.net) |
| 2021-08-03 19:05:21 | × | jippiedoe quits (~david@77-171-152-62.fixed.kpn.net) (Client Quit) |
| 2021-08-03 19:05:26 | <drakonis> | function composition is a weird call out |
| 2021-08-03 19:05:32 | <drakonis> | because scheme does that really well |
| 2021-08-03 19:05:58 | <monochrom> | No, function composition is unusable without currying. |
| 2021-08-03 19:06:00 | <drakonis> | https://guix.gnu.org/manual/en/html_node/Channels.html davean |
| 2021-08-03 19:06:34 | <monochrom> | Look at how we in Haskell write "takeWhile (\x -> x<5) . filter p . map f" all the time. |
| 2021-08-03 19:06:40 | <monochrom> | Try it in Scheme. |
| 2021-08-03 19:06:56 | <drakonis> | idgi |
All times are in UTC.