Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 294 295 296 297 298 299 300 301 302 303 304 .. 5022
— 502,152 events total
2020-09-29 15:18:22 <turion> ski: Yes
2020-09-29 15:18:51 <turion> For your example it's not needed, right
2020-09-29 15:18:54 → voyons_osti joins (~dan@107-190-41-58.cpe.teksavvy.com)
2020-09-29 15:18:58 <ski> right
2020-09-29 15:19:08 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-29 15:19:12 <turion> I was implicitly thinking about my use case, which is a kind of free arrow
2020-09-29 15:19:16 × filwishe1 quits (~filwisher@cpc76738-dals23-2-0-cust186.20-2.cable.virginm.net) (Ping timeout: 246 seconds)
2020-09-29 15:20:51 → zpao joins (~zpao@185.163.110.116)
2020-09-29 15:21:43 <turion> This kind of problem sounds like something that someone has written a functional pearl about 10 years ago
2020-09-29 15:22:52 × jespada quits (~jespada@90.254.241.6) (Ping timeout: 265 seconds)
2020-09-29 15:23:43 × heatsink quits (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net) (Ping timeout: 260 seconds)
2020-09-29 15:24:20 × takuan quits (~takuan@178-116-218-225.access.telenet.be) (Remote host closed the connection)
2020-09-29 15:25:02 <Cale> Ah, writing the instance, you end up wanting to know not only that there are instances of Arbitrary for (cat a String) and (cat String a) for example, but also for (cat String String) and all combinations of intermediate types you plan to use
2020-09-29 15:25:07 × luke quits (~luke@bitnomial/staff/luke) (Quit: sleep)
2020-09-29 15:25:13 → jespada joins (~jespada@90.254.241.6)
2020-09-29 15:25:37 → knupfer joins (~Thunderbi@i5E86B4C1.versanet.de)
2020-09-29 15:27:01 <Cale> So the instance head has O(n^2) constraints on it if you want n intermediate types :)
2020-09-29 15:27:13 × koz quits (~koz@121.99.240.58) (Remote host closed the connection)
2020-09-29 15:27:21 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-09-29 15:27:29 → koz joins (~koz@121.99.240.58)
2020-09-29 15:27:33 <ski> yes, hence i wanted `someArbitrary :: Gen (exists b. SomeArbitrary (cat b) *> FreeCat cat a b)', for `instance SomeArbitrary (cat a) => SomeArbitrary (FreeCat cat a)'
2020-09-29 15:27:48 × elliott_ quits (~elliott_@pool-100-36-54-163.washdc.fios.verizon.net) (Ping timeout: 272 seconds)
2020-09-29 15:28:14 <Cale> tsk, wildcards not allowed in instance declarations
2020-09-29 15:28:17 <Cale> annoying
2020-09-29 15:28:33 <ski> so, when it's picking a new type for you, it should also give you an instance for how you could continue from that type
2020-09-29 15:28:43 <ski> (this is very specific to `FreeCat', obviously)
2020-09-29 15:29:01 <Cale> yeah
2020-09-29 15:29:08 → bloodstalker joins (~bloodstal@46.166.187.188)
2020-09-29 15:29:15 <Cale> You could also just demand everything you'll end up needing
2020-09-29 15:29:20 × rslima____ quits (sid26145@gateway/web/irccloud.com/x-zjgqdcvtyayglqgm) (Ping timeout: 260 seconds)
2020-09-29 15:29:35 <Cale> It's also probably sufficient just to have a single intermediate type, in practice. :P
2020-09-29 15:29:48 × Herasmus quits (~mb@91-160-105-150.subs.proxad.net) (Quit: leaving)
2020-09-29 15:29:58 <ski> you mean for other cases than `FreeCat' ?
2020-09-29 15:30:07 <Cale> No
2020-09-29 15:30:19 <Squarism> Cheery, thanks for the link. Thats exactly what I'm after
2020-09-29 15:30:46 → rslima____ joins (sid26145@gateway/web/irccloud.com/x-nijwoaxwcetixacu)
2020-09-29 15:31:38 <ski> hm, sorry. for some reason, i misread you as saying "insufficient", rather than "sufficient"
2020-09-29 15:32:22 <Cale> https://dpaste.com/5DUQARMCK -- anyway, something like this works
2020-09-29 15:32:49 → Tops2 joins (~Tobias@dyndsl-091-249-082-222.ewe-ip-backbone.de)
2020-09-29 15:33:50 → bitmapper joins (uid464869@gateway/web/irccloud.com/x-jerhumioweuciiqb)
2020-09-29 15:34:07 <Cale> Probably () would be a more useful choice than either Integer or String as a second type. Pick something with lots of values and something with few values :)
2020-09-29 15:34:17 <Cale> Maybe throw in Void
2020-09-29 15:34:27 × turion quits (~turion@ip1f10fe5a.dynamic.kabel-deutschland.de) (Remote host closed the connection)
2020-09-29 15:35:04 <z0> /quit
2020-09-29 15:35:14 × z0 quits (~z0@bl17-251-27.dsl.telepac.pt) (Quit: leaving)
2020-09-29 15:35:40 <monochrom> Nice, GHC 9.0.1 alpha1 was released yesterday
2020-09-29 15:37:07 <monochrom> onoes, the GHC user's guide will use the readthedocs format :)
2020-09-29 15:38:00 → turion joins (~turion@ip1f10fe5a.dynamic.kabel-deutschland.de)
2020-09-29 15:38:35 × lagothrix quits (~lagothrix@unaffiliated/lagothrix) (Ping timeout: 258 seconds)
2020-09-29 15:38:50 → lagothrix joins (~lagothrix@unaffiliated/lagothrix)
2020-09-29 15:39:29 <turion> Ideally, have lots of different types, and ideally tweak the intermediate types depending over what we want the free category
2020-09-29 15:40:28 × jgt quits (~jgt@46.250.27.223.pool.breezein.net) (Ping timeout: 272 seconds)
2020-09-29 15:40:37 <turion> E.g. in FreeCat (->) a b we basically want everything we can think of that has (CoArbitrary a, Arbitrary b). In ski's example, we want just the types that occur in Fun
2020-09-29 15:41:09 × jedws quits (~jedws@121.209.139.222) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-09-29 15:41:18 → solonarv joins (~solonarv@astrasbourg-653-1-252-112.w92-161.abo.wanadoo.fr)
2020-09-29 15:43:31 × xerox_ quits (~xerox@unaffiliated/xerox) (Quit: leaving)
2020-09-29 15:43:50 <ski> turion : could also allow other types to be used, with `Id'
2020-09-29 15:44:16 × kritzefitz quits (~kritzefit@fw-front.credativ.com) (Ping timeout: 256 seconds)
2020-09-29 15:44:47 <Cale> With newer GHCs, you could use QuantifiedConstraints to get the thing you want in the (->) case
2020-09-29 15:45:24 <Cale> It's also probably possible to construct a type family that lets you turn a list of types into a list of all the needed Arbitrary constraints between them
2020-09-29 15:45:37 <turion> ski: Hmm true. So maybe we always want as many types as possible. But sometimes we'll have to restrict to the ones that are specified in a constructor
2020-09-29 15:45:42 <Cale> (rather, into the combination of those constraints directly
2020-09-29 15:45:43 <Cale> )
2020-09-29 15:46:57 <ski> turion : perhaps there could be some way to "inject" which types one'd like it to consider for existentials .. not sure how one'd want that to work, though
2020-09-29 15:47:17 <turion> ski: Like a list of proxies hidden in an existential type
2020-09-29 15:47:26 <Cale> Also, note that you don't have to stick to a single Arbitrary instance.
2020-09-29 15:48:03 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Ping timeout: 260 seconds)
2020-09-29 15:48:04 <turion> Cale: What do you mean?
2020-09-29 15:48:29 <Cale> You can write a bunch of these Gen values for various circumstances, and use them via QuickCheck's forAll
2020-09-29 15:48:35 <Cale> forAll :: (Show a, Testable prop) => Gen a -> (a -> prop) -> Property
2020-09-29 15:48:49 <turion> Ah yes :) that will sometimes make it easier
2020-09-29 15:48:52 <monochrom> Yikes, new extension LexicalNegation. If you turn it on, then "x -y" means "x (negate y)", and "(- b)" means the section "subtract b".
2020-09-29 15:49:11 × DirefulSalt quits (DirefulSal@gateway/vpn/privateinternetaccess/direfulsalt) (Remote host closed the connection)
2020-09-29 15:50:04 → DirefulSalt joins (DirefulSal@gateway/vpn/privateinternetaccess/direfulsalt)
2020-09-29 15:50:18 → elliott_ joins (~elliott_@pool-100-36-54-163.washdc.fios.verizon.net)
2020-09-29 15:50:19 → jb55 joins (~jb55@gateway/tor-sasl/jb55)
2020-09-29 15:50:37 <dolio> Oh, it turns off unary minus with spaces?
2020-09-29 15:50:39 <Cale> monochrom: I hate the direction that GHC's been going lately
2020-09-29 15:51:03 <ski> Cale : syntactically ?
2020-09-29 15:51:07 <monochrom> I have a conspiracy theory! LinearTypes was a smoke screen to distract people from noticing the really breaking changes like LexicalNegation and simplified subsumption. :)
2020-09-29 15:51:18 <Cale> ski: Syntactically, semantically...
2020-09-29 15:51:19 × nbloomf quits (~nbloomf@2600:1700:83e0:1f40:6d38:c363:743b:50e9) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2020-09-29 15:51:36 <monochrom> Yeah dolio, basically that.
2020-09-29 15:52:25 <Cale> LinearTypes is a disaster unfolding in slow motion, then there's all these little stupid syntactic extensions that everyone's going to have to deal with.
2020-09-29 15:52:36 <dolio> Simplified subsumption I can kind of get behind, if it's what I'm thinking of.
2020-09-29 15:52:43 <monochrom> Well, also, "x -y" meaning "x (negate y)" is turning on unary minus with absence of spaces. :)
2020-09-29 15:53:25 <monochrom> "\x -> sin x == sin -x" is now a valid property for quickcheck :)
2020-09-29 15:53:40 <monochrom> err, s/sin/cos/
2020-09-29 15:53:58 <dolio> Right. I just thought it was adding the rule about -x, not taking away - x.
2020-09-29 15:54:27 <Cale> I feel like RecordDotSyntax was the kind of thing that newbies would always ask for like a decade ago, but there were always good reasons not to do it. Now we've just gone ahead and ignored all those good reasons, and implemented something which makes it easy to write terms that nobody knows how to parse
2020-09-29 15:55:07 × acidjnk_new quits (~acidjnk@p200300d0c72378745c7b1a0a573f1493.dip0.t-ipconnect.de) (Ping timeout: 240 seconds)
2020-09-29 15:55:11 <Cale> (there's a rule for disambiguation, but it's not obvious)
2020-09-29 15:55:37 <monochrom> (basically using spaces again, yeah)
2020-09-29 15:55:59 <maerwald> magical spaces
2020-09-29 15:56:07 <Cale> It honestly makes me want to fork the entire language
2020-09-29 15:56:31 <maerwald> reminds me of the `return foo;` vs `foo` syntax in rust, which is nonsense imo
2020-09-29 15:56:34 → wroathe joins (~wroathe@c-73-24-27-54.hsd1.mn.comcast.net)
2020-09-29 15:56:40 <dolio> Taking away spaced unary minus might actually be good in that situation, I guess, because it's one less complication.
2020-09-29 15:56:51 <Cale> GHC doesn't need new language extensions, it needs to be simpler and produce better code in less time
2020-09-29 15:56:52 <dolio> Like "-x `mod` y" vs "- x `mod` y".
←Prev  Next→
Page 1 .. 294 295 296 297 298 299 300 301 302 303 304 .. 5022

All times are in UTC.