Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 655 656 657 658 659 660 661 662 663 664 665 .. 5022
— 502,152 events total
2020-10-15 18:23:07 → plast1k_ joins (~plast1k_@196.207.181.246)
2020-10-15 18:23:20 → knupfer1 joins (~Thunderbi@mue-88-130-61-082.dsl.tropolys.de)
2020-10-15 18:24:36 → cr3 joins (~cr3@192-222-143-195.qc.cable.ebox.net)
2020-10-15 18:24:48 × dyeplexer quits (~lol@unaffiliated/terpin) (Remote host closed the connection)
2020-10-15 18:24:49 × wroathe quits (~wroathe@c-73-24-27-54.hsd1.mn.comcast.net) (Ping timeout: 258 seconds)
2020-10-15 18:24:49 → chaosmasttter joins (~chaosmast@p200300c4a711b001c4478125cc0e8bf4.dip0.t-ipconnect.de)
2020-10-15 18:25:05 → deviantfero joins (~deviantfe@190.99.48.231)
2020-10-15 18:25:47 → shailangsa joins (~shailangs@host86-186-133-58.range86-186.btcentralplus.com)
2020-10-15 18:26:17 <jsynacek> Hi! I'm having a problem understanding State and >>= together.
2020-10-15 18:26:22 <jsynacek> I have this: foo :: State Int [Int]
2020-10-15 18:26:27 <jsynacek> foo = get >>= \s -> pure [s+10] >>= \s' -> pure $ s' ++ [7,8,9]
2020-10-15 18:26:54 <jsynacek> Can anyone please explain to me how the s' suddenly changes from Int to [Int]? I can't figure it out.
2020-10-15 18:27:11 → isovector1 joins (~isovector@172.103.216.166)
2020-10-15 18:27:13 <jsynacek> I mean it's still supposed to be of the same type as the state, right?
2020-10-15 18:27:25 × knupfer quits (~Thunderbi@200116b82c60f7000c4226fffe43b3db.dip.versatel-1u1.de) (Ping timeout: 240 seconds)
2020-10-15 18:27:25 knupfer1 is now known as knupfer
2020-10-15 18:28:03 → xff0x joins (~fox@2001:1a81:5350:d900:f51f:ced4:4c46:54ba)
2020-10-15 18:28:13 × dansho quits (~dansho@ip68-108-167-185.lv.lv.cox.net) (Quit: Leaving)
2020-10-15 18:28:13 <jsynacek> So I would suppose that s and s' should both be Int. But s is Int and s' is [Int].
2020-10-15 18:28:14 <verement> :t \s -> pure [s+10]
2020-10-15 18:28:15 <lambdabot> (Applicative f, Num a) => a -> f [a]
2020-10-15 18:28:40 → acidjnk_new joins (~acidjnk@p200300d0c723781658597087ca157dd5.dip0.t-ipconnect.de)
2020-10-15 18:29:23 <monochrom> I don't understand the "change" mental model. There is no change.
2020-10-15 18:29:59 <monochrom> "pure x >>= \y -> ..." necessarily satisfies y=x therefore y has the same type as x.
2020-10-15 18:30:18 <larou> johnw: it would be some arbitrary label, probably a Symbol, to accompany the instance - that way allowing many such instances, that otherwise would clash
2020-10-15 18:30:20 <monochrom> Oh, x is a list this time? Then it is a list this time.
2020-10-15 18:30:20 <dolio> You named the variables to confuse yourself. :)
2020-10-15 18:31:05 <johnw> larou: I suppose if you used a Nat and yourself ensured monotonicity, you could impose a type-level ordering constraint
2020-10-15 18:31:11 <larou> then you would have the downstream reference constrained to reference the corresponding label
2020-10-15 18:31:27 <larou> johnw: hmm, good idea
2020-10-15 18:31:57 <larou> maybe that aswell as a label
2020-10-15 18:32:22 <larou> otherwise you would only be able to make one value of the datatype
2020-10-15 18:32:43 <larou> or, have to use arbitrarily large first Nat for all the subsequent values
2020-10-15 18:33:07 → cole-h joins (~cole-h@c-73-48-197-220.hsd1.ca.comcast.net)
2020-10-15 18:33:09 <geekosaur> jsynacek, note that the only thing that does anything with the state is the `get`; you then produce a list, and then consume a list, and produce another list, carrying the state along without touching it
2020-10-15 18:33:23 <larou> or maybe not a Nat
2020-10-15 18:33:28 <larou> but like f ()
2020-10-15 18:33:33 <larou> for the linked container f
2020-10-15 18:33:50 <larou> like, if it were a doubly linked list, a Nat would be ok - but not if it were a tree
2020-10-15 18:34:03 <larou> erm, a pointer on (f ())
2020-10-15 18:34:22 <jsynacek> geekosaur: Yes, that's it. I totally confused myself.
2020-10-15 18:34:28 <jsynacek> geekosaur++
2020-10-15 18:34:40 <larou> so the "name" label, has to be the same for every constructor used sequentially to construct the container
2020-10-15 18:34:55 <larou> they are all just different bits of the "same container"
2020-10-15 18:35:20 <larou> and the pointer to the position of the part being referenced would be like the Nat
2020-10-15 18:35:32 <larou> not sure how your monotonicity constraint works then..
2020-10-15 18:36:05 <larou> i think the doubly linked list example has "the immediately preceding value" expressed via the constraint
2020-10-15 18:36:16 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2020-10-15 18:36:21 <larou> not anything less than the current index
2020-10-15 18:36:39 <larou> or did you mean monotinic position, like successive "forwards" positions of a zipper?
2020-10-15 18:37:04 <larou> indexing at type level *each* of the positions, not just the cyclic references
2020-10-15 18:37:06 <hyperisco> so with the Haskell VSCode extension... 1) any way to insert a suitable import given an identifier? 2) any way to have ctrl+shift+b build the project?
2020-10-15 18:37:45 <hyperisco> so far the error underlining hasn't exploded so that is progress
2020-10-15 18:37:58 <larou> i guess then you would have something like a tree zipper, and the constraint would be that you would only be able to navigate down and forwards, not upwards and backwards... thats kind of like the ordering of a Nat...
2020-10-15 18:39:01 <hyperisco> I hate knowing a name like getArgs but not knowing what module it is in
2020-10-15 18:39:07 <hyperisco> if the IDE could help me with that it would be a godsend
2020-10-15 18:41:05 <monochrom> Tranditionally, people install hoogle locally for that. They probably also get their IDEs to call hoogle.
2020-10-15 18:41:27 <jsynacek> hyperisco: vscode has a feature called Tasks. You can use it to run 'cabal build' on ctrl+shift+b.
2020-10-15 18:41:34 → bitmagie joins (~Thunderbi@200116b806495000583b859276a8aad8.dip.versatel-1u1.de)
2020-10-15 18:41:36 <monochrom> Me, I just crack open a web browser tab and use the haddock index.
2020-10-15 18:42:14 <jsynacek> hyperisco: Check out https://code.visualstudio.com/docs/editor/tasks#vscode
2020-10-15 18:42:37 <hyperisco> monochrom, when you have to import hundreds of things a day that adds up hugely
2020-10-15 18:43:18 → fendor joins (~fendor@178.115.131.201.wireless.dyn.drei.com)
2020-10-15 18:43:21 <hyperisco> jsynacek, okay thanks
2020-10-15 18:43:40 <hyperisco> how about tabs… I have VSCode set to width 2 but for whatever reason in .hs files it is back to 4
2020-10-15 18:44:45 × sleblanc quits (~sleblanc@unaffiliated/sebleblanc) (Ping timeout: 258 seconds)
2020-10-15 18:45:31 × jneira_ quits (~jneira@80.30.100.250) (Ping timeout: 258 seconds)
2020-10-15 18:46:22 × bitmagie quits (~Thunderbi@200116b806495000583b859276a8aad8.dip.versatel-1u1.de) (Client Quit)
2020-10-15 18:46:45 → jneira_ joins (~jneira@80.30.100.250)
2020-10-15 18:47:04 <hyperisco> just need a file-extension-specific override... okay
2020-10-15 18:47:28 hackage implicit-hie 0.1.2.0 - Auto generate hie-bios cradles & hie.yaml https://hackage.haskell.org/package/implicit-hie-0.1.2.0 (Avi_Dessauer)
2020-10-15 18:49:40 → bitmagie joins (~Thunderbi@200116b806495000583b859276a8aad8.dip.versatel-1u1.de)
2020-10-15 18:50:32 × thir quits (~thir@p200300f27f025800e13a3eca7002ede4.dip0.t-ipconnect.de) (Ping timeout: 260 seconds)
2020-10-15 18:51:09 <hyperisco> would be nice if you could get the type of a term
2020-10-15 18:51:24 <hyperisco> seems you can only get types for imported names
2020-10-15 18:51:53 × bitmagie quits (~Thunderbi@200116b806495000583b859276a8aad8.dip.versatel-1u1.de) (Client Quit)
2020-10-15 18:51:55 <hyperisco> oh nvm it is working now... now I just wish there was a hotkey instead of hovering
2020-10-15 18:52:02 <hyperisco> because going to the mouse is no good
2020-10-15 18:53:27 <hyperisco> apparently is a vscode thing if you do ctrl+k+i ... hope this helps someone
2020-10-15 18:54:32 × conal_ quits (~conal@64.71.133.70) (Quit: Computer has gone to sleep.)
2020-10-15 18:54:52 <hyperisco> oh damnnnnnn it will auto import names
2020-10-15 18:56:14 <hyperisco> ctrl+. while cursored on the name... gives several import options
2020-10-15 18:56:41 <hyperisco> Haskell officially has good IDE support in my eyes… well done. where do I insert my money?
2020-10-15 18:57:09 <larou> i guess the problem would then be that every version of the container would still have this label
2020-10-15 18:57:29 <larou> maybe a synonym with a quantification would allow it to be polymorphic in this label
2020-10-15 18:58:07 → heatsink joins (~heatsink@107-136-5-69.lightspeed.sntcca.sbcglobal.net)
2020-10-15 18:58:11 <larou> well, this is why it was called a "Let" class anyway, so that it binds the "name of the value" as a type parameter
2020-10-15 18:58:36 <larou> then you can at least constrain the cyclic references to be to the same container
2020-10-15 18:58:45 <larou> but im not sure what would happen when modifying it
2020-10-15 18:59:01 <larou> like, i guess you would have to give it a new type label and Let instance
2020-10-15 18:59:07 <larou> this seems impossible
2020-10-15 18:59:29 <larou> its these kind of considerations that makes it seem like it would be better as a language feature than as a library
2020-10-15 18:59:32 × kuribas quits (~user@ptr-25vy0i8gfnrq7cvfnez.18120a2.ip6.access.telenet.be) (Remote host closed the connection)
2020-10-15 19:00:00 <larou> you would have all the knot tying stuff being done under the hood then too
2020-10-15 19:00:22 × xff0x quits (~fox@2001:1a81:5350:d900:f51f:ced4:4c46:54ba) (Remote host closed the connection)
2020-10-15 19:00:42 → xff0x joins (~fox@2001:1a81:5350:d900:f127:8532:42d4:579b)
2020-10-15 19:01:25 × dhil quits (~dhil@195.213.192.122) (Ping timeout: 240 seconds)
2020-10-15 19:02:14 → conal joins (~conal@64.71.133.70)
2020-10-15 19:03:31 → berberman_ joins (~berberman@unaffiliated/berberman)
2020-10-15 19:03:46 × taurux quits (~taurux@net-93-144-64-237.cust.dsl.teletu.it) (Ping timeout: 256 seconds)
←Prev  Next→
Page 1 .. 655 656 657 658 659 660 661 662 663 664 665 .. 5022

All times are in UTC.