Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 319 320 321 322 323 324 325 326 327 328 329 .. 5022
— 502,152 events total
2020-09-30 07:59:03 <kindaro> No, only those in a specified variety. If a type is not supported, I should respectfully decline.
2020-09-30 07:59:51 × hekkaidekapus quits (~tchouri@gateway/tor-sasl/hekkaidekapus) (Quit: hekkaidekapus)
2020-09-30 08:00:20 → coot joins (~coot@37.30.59.210.nat.umts.dynamic.t-mobile.pl)
2020-09-30 08:00:36 <tomsmeding> then I guess if you're already making a list of the supported ones, well, you already have to write the list :)
2020-09-30 08:01:06 <tomsmeding> doing it with Enum/Read/Show can help reducing the number of times you have to write that list, I guess
2020-09-30 08:01:20 → dhil joins (~dhil@11.29.39.217.dyn.plus.net)
2020-09-30 08:03:19 <lortabac> kindaro: maybe you can try to generate a big GADT with TemplateHaskell and a mechanical rule for parsing, but at some point you necessarily need to list all the types at least once
2020-09-30 08:03:39 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 08:04:21 <lortabac> something like data CliType :: * -> * where Int :: CliType Int; Char :: CliType Char...
2020-09-30 08:05:17 <kindaro> Suppose I want to support all types that are instances of `Show` that are in scope at the time of compilation.
2020-09-30 08:05:37 <kindaro> I then need to get a list of such types and build a case expression.
2020-09-30 08:06:50 <tomsmeding> "all types that are instances of `Show`" sounds nice but is not something that Haskell lets you query easily
2020-09-30 08:07:30 → notzmv` joins (~user@177.45.26.174)
2020-09-30 08:07:33 <tomsmeding> not even sure if TemplateHaskell will get you that
2020-09-30 08:08:09 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 256 seconds)
2020-09-30 08:08:39 <tomsmeding> oh it does
2020-09-30 08:08:54 <tomsmeding> ooooooh that is decidedly ugly
2020-09-30 08:09:15 memxor is now known as kkd
2020-09-30 08:09:22 <kindaro> Can you give me some links?
2020-09-30 08:09:47 × howdoi quits (uid224@gateway/web/irccloud.com/x-zozvjyzrnsmbzmrc) (Quit: Connection closed for inactivity)
2020-09-30 08:11:02 × notzmv quits (~user@unaffiliated/zmv) (Ping timeout: 258 seconds)
2020-09-30 08:12:20 × tzh quits (~tzh@2601:448:c500:5300::82b3) (Quit: zzz)
2020-09-30 08:12:58 <tomsmeding> just checking myself if it will work at all, but https://hackage.haskell.org/package/template-haskell-2.16.0.0/docs/Language-Haskell-TH.html#v:reifyInstances
2020-09-30 08:13:37 <lortabac> kindaro: there are a couple of ugly solutions for that, though I would never use them for real
2020-09-30 08:13:49 <lortabac> @hackage constraints-emerge
2020-09-30 08:13:49 <lambdabot> https://hackage.haskell.org/package/constraints-emerge
2020-09-30 08:14:05 × xcmw quits (~textual@dyn-72-33-0-89.uwnet.wisc.edu) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-09-30 08:14:29 <kindaro> Oh, this is cute!
2020-09-30 08:14:31 <lortabac> https://github.com/rampion/constraint-unions
2020-09-30 08:15:27 × bahamas quits (~lucian@unaffiliated/bahamas) (Ping timeout: 240 seconds)
2020-09-30 08:16:27 × lucid_0x80 quits (~lucid_0x8@188.253.231.113) (Ping timeout: 240 seconds)
2020-09-30 08:16:27 <kindaro> I recall passing explicit dictionaries is being considered for GHC: https://github.com/ghc-proposals/ghc-proposals/pull/324
2020-09-30 08:18:01 → knupfer joins (~Thunderbi@mue-88-130-61-186.dsl.tropolys.de)
2020-09-30 08:18:29 <tomsmeding> kindaro: import Language.Haskell.TH ; import Language.Haskell.TH.Ppr ; putStrLn $( LitE . StringL . pprint <$> reifyInstances ''Show [ VarT (mkName "a") ] )
2020-09-30 08:18:34 <tomsmeding> that prints stuff :p
2020-09-30 08:18:54 <tomsmeding> presumably if you know a bit more about templatehaskell you can make that work somehow?
2020-09-30 08:22:05 → bahamas joins (~lucian@unaffiliated/bahamas)
2020-09-30 08:22:21 → ces joins (~ces@52d3ce3c.dynamic-ip.k-net.dk)
2020-09-30 08:23:27 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 08:24:23 → raichoo joins (~raichoo@213.240.178.58)
2020-09-30 08:26:14 → cpressey joins (~cpressey@79-72-201-29.dynamic.dsl.as9105.com)
2020-09-30 08:26:15 → kuribas joins (~user@ptr-25vy0i8if8811202m3x.18120a2.ip6.access.telenet.be)
2020-09-30 08:28:44 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 272 seconds)
2020-09-30 08:32:04 → __monty__ joins (~toonn@unaffiliated/toonn)
2020-09-30 08:33:32 <kuribas> what is wrong with clojure guys absolutely having to bash haskell to show their language? https://lispcast.com/user-wizard-scenario/
2020-09-30 08:33:49 × carldd1 quits (~carldd@90-224-49-113-no56.tbcn.telia.com) (Read error: Connection reset by peer)
2020-09-30 08:33:52 <merijn> kuribas: Insecurity
2020-09-30 08:33:53 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 08:34:25 <merijn> Nothing illustrates your point as well as a strawman the size of the wicker man...
2020-09-30 08:34:43 <kuribas> merijn: if they are happy with their language, why not just use it?
2020-09-30 08:34:55 → carldd1 joins (~carldd@90-224-49-113-no56.tbcn.telia.com)
2020-09-30 08:35:12 <merijn> kuribas: Because they feel insecure about using it and need to justify the superiority :p
2020-09-30 08:35:27 <merijn> That's like 90% of all "language X sucks, look at language Y" advocacy
2020-09-30 08:35:33 <kuribas> it's not like haskell is perfect
2020-09-30 08:35:38 <merijn> Unless is C or C++, then it's just truth
2020-09-30 08:35:44 <kuribas> lol
2020-09-30 08:35:49 <merijn> s/Unless/Unless X/
2020-09-30 08:35:50 tomsmeding X doubt
2020-09-30 08:36:30 <kuribas> I am perfectly fine with bashing haskell, but their criticism makes no sense.
2020-09-30 08:36:44 × hnOsmium0001 quits (uid453710@gateway/web/irccloud.com/x-arhqczabjmakjtmj) (Quit: Connection closed for inactivity)
2020-09-30 08:36:50 <merijn> kuribas: Of course, because they don't know haskell enough to make a coherent criticism
2020-09-30 08:37:07 <merijn> Just look at their silly 8 alternatives for partial validation
2020-09-30 08:37:17 <kuribas> I don't see other lisp guys do this, like common lisp or scheme.
2020-09-30 08:37:25 <kuribas> merijn: you mean spec?
2020-09-30 08:37:30 <merijn> There's much more elegant approaches using higher kinded types
2020-09-30 08:37:42 <merijn> kuribas: Now, the literaly example in your link
2020-09-30 08:37:43 × knupfer quits (~Thunderbi@mue-88-130-61-186.dsl.tropolys.de) (Ping timeout: 246 seconds)
2020-09-30 08:38:30 <merijn> kuribas: https://paste.tomsmeding.com/yy3PqZUb
2020-09-30 08:38:35 <tomsmeding> kuribas: that article is hilarious
2020-09-30 08:38:38 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 258 seconds)
2020-09-30 08:38:40 <merijn> As if any self-respecting Haskell programmer would write that
2020-09-30 08:38:49 <kuribas> yeah :)
2020-09-30 08:38:54 <merijn> That shit is begging for a digestive-functor/applicative validation so badly
2020-09-30 08:39:03 → chaosmasttter joins (~chaosmast@p200300c4a70aba018412238bea948280.dip0.t-ipconnect.de)
2020-09-30 08:39:10 <tomsmeding> it's basically saying: clojure is cool because it doesn't have the tools to give you type safety
2020-09-30 08:39:15 <tomsmeding> I mean, you can get the exact same in haskell
2020-09-30 08:39:18 <tomsmeding> they even did that in the article!
2020-09-30 08:39:20 <merijn> Like, Validation obsoletes their entire argument
2020-09-30 08:39:31 <tomsmeding> it's literally not an argument against haskell or for clojure
2020-09-30 08:39:47 <tomsmeding> only argument being made is that haskell is more flexible, apparently? (I don't know clojure so can't confirm/deny)
2020-09-30 08:39:59 <merijn> kuribas: Lisp/scheme programmers don't feel compelled to "prove" themselves because they've got history/legacy
2020-09-30 08:40:21 <tomsmeding> merijn: yes that example cracked me up too
2020-09-30 08:40:44 <merijn> kuribas: Clojure is "the new kid" of lisps and so they feel compelled to justify it being better than Scheme/Lisp by picking silly fights like thise
2020-09-30 08:41:20 <kuribas> merijn: or just a record with Maybes?
2020-09-30 08:41:40 → thc202 joins (~thc202@unaffiliated/thc202)
2020-09-30 08:41:49 <tomsmeding> well they discarded that because it wasn't flexible enough for the frontenders, right?
2020-09-30 08:42:03 <merijn> kuribas: Really just anything but what they actually did
2020-09-30 08:42:12 <kuribas> tomsmeding: yeah, most stuff with hashmaps can be done in haskell with records. Except for merging, but that's dodge anyway.
2020-09-30 08:42:39 <tomsmeding> except adding new fields at runtime
2020-09-30 08:42:43 <tomsmeding> which they wanted apparently
2020-09-30 08:43:11 <merijn> which is a dumb thing to want anyway :p
2020-09-30 08:43:26 <kuribas> https://gist.github.com/kuribas/8afd49cf87fe5b1c32b1009ab59c4b1d
2020-09-30 08:43:35 <tomsmeding> article makes the case that it isn't, but that isn't even the problem
2020-09-30 08:44:01 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-30 08:44:54 <kuribas> tomsmeding: this does exactly what they want, you can add fields easily, and change the order of processing.
2020-09-30 08:45:02 <kuribas> just swap the guards
2020-09-30 08:45:05 → vicfred_ joins (~vicfred@unaffiliated/vicfred)
2020-09-30 08:45:10 <tomsmeding> not as easily as they want, right?
2020-09-30 08:45:24 <kuribas> what do they want then?
2020-09-30 08:45:29 <tomsmeding> they didn't want to change the backend code to add fields, or didn't I read the article closely enough (likely)
←Prev  Next→
Page 1 .. 319 320 321 322 323 324 325 326 327 328 329 .. 5022

All times are in UTC.