Newest at the top
| 2025-04-06 19:51:29 +0200 | <milan> | mauke: Thanks |
| 2025-04-06 19:50:00 +0200 | tromp | (~textual@2001:1c00:3487:1b00:c873:d422:44c2:cc0c) (Quit: My iMac has gone to sleep. ZZZzzz…) |
| 2025-04-06 19:48:33 +0200 | <geekosaur> | …if that's the question |
| 2025-04-06 19:48:24 +0200 | <geekosaur> | TypeOperators is default these days, I think? (at least if you turn on NoStarIsType so they're available) |
| 2025-04-06 19:46:46 +0200 | <hellwolf> | *2010 |
| 2025-04-06 19:46:42 +0200 | <hellwolf> | https://www.haskell.org/onlinereport/haskell2010/haskellch10.html#x17-17700010.2 <-- is that even part of haskell2012 |
| 2025-04-06 19:40:46 +0200 | <mauke> | I think all symbols can be used at the start of type names |
| 2025-04-06 19:40:36 +0200 | econo_ | (uid147250@id-147250.tinside.irccloud.com) |
| 2025-04-06 19:39:31 +0200 | <mauke> | type names are weird now |
| 2025-04-06 19:39:00 +0200 | <lambdabot> | "0.0 + 1.4142135623730951i" |
| 2025-04-06 19:38:58 +0200 | <mauke> | > case sqrt (-2) of x :+ y -> show x <> " + " <> show y <> "i" |
| 2025-04-06 19:38:39 +0200 | <hellwolf> | type name can also start with ":", right? |
| 2025-04-06 19:37:23 +0200 | <mauke> | or an uppercase letter |
| 2025-04-06 19:37:17 +0200 | <mauke> | milan: if it starts with : it's a constructor |
| 2025-04-06 19:36:58 +0200 | <hellwolf> | yahb2 is just one regex away from AGI |
| 2025-04-06 19:36:12 +0200 | <geekosaur> | (whether the bridge lets that work or npot seems to depend on version, and they don't provide config for it) |
| 2025-04-06 19:35:59 +0200 | <yahb2> | https://paste.tomsmeding.com/AtPoarKA |
| 2025-04-06 19:35:58 +0200 | <hellwolf> | %% :i Maybe |
| 2025-04-06 19:35:21 +0200 | <haskellbridge> | <geekosaur> oh, crap, they removed that again |
| 2025-04-06 19:35:09 +0200 | <haskellbridge> | <geekosaur> %% :i Maybe |
| 2025-04-06 19:35:08 +0200 | tromp | (~textual@2001:1c00:3487:1b00:c873:d422:44c2:cc0c) |
| 2025-04-06 19:34:40 +0200 | <geekosaur> | leave a blank line. but lambdabot doesn't understand :i anyway |
| 2025-04-06 19:34:32 +0200 | XZDX | XZDX_afk |
| 2025-04-06 19:34:18 +0200 | <haskellbridge> | <hellwolf> :i Maybe |
| 2025-04-06 19:34:07 +0200 | <haskellbridge> | <hellwolf> > :! Maybe |
| 2025-04-06 19:32:58 +0200 | <milan> | Hello I just found I can pattern match on cons (:) and I can't on add (+). I managed to find that cons is constructor and add is just function. How can I in ghci find what functions are also constructors and thus can be used in pattern matching? |
| 2025-04-06 19:30:00 +0200 | weary-traveler | (~user@user/user363627) user363627 |
| 2025-04-06 19:24:16 +0200 | <haskellbridge> | <hellwolf> with vdq, it's part of the regular currying order. |
| 2025-04-06 19:24:05 +0200 | <haskellbridge> | <hellwolf> with typeapplication, the annoying thing is that you need to know what's the order of the type variables, with type class's type variables come first. you can also use @_ to skip ones that you don't care |
| 2025-04-06 19:23:23 +0200 | <haskellbridge> | <hellwolf> with 9.10, you have an alternative syntax (which will be used for even more complicated use case) that is RequiredTypeArguments (or vdq, mouthfully, visible dependent quantification) |
| 2025-04-06 19:21:43 +0200 | <haskellbridge> | <hellwolf> unless it's really polymorphic to any m1 m2, then for a good reason, you will need to use type applications |
| 2025-04-06 19:21:15 +0200 | <haskellbridge> | ... long message truncated: https://kf8nh.com/_heisenbridge/media/kf8nh.com/yxUTAVDLAhEYmWfAmyqvumAm/BKPqvW_Oq7A (4 lines) |
| 2025-04-06 19:21:14 +0200 | <haskellbridge> | <hellwolf> okay. https://hackage.haskell.org/package/monad-control-1.0.3.1/docs/src/Control.Monad.Trans.Control.htm… |
| 2025-04-06 19:18:57 +0200 | <emojelly> | ok, really gotta run now, but I'll read if there's anything more! Thanks again for the help so far! |
| 2025-04-06 19:18:41 +0200 | <emojelly> | (StT is actually coming from MonadTransControl, but shouldn't matter) |
| 2025-04-06 19:18:11 +0200 | <haskellbridge> | <hellwolf> check the doc, it has a list of rules when you can or cannot define dependencies |
| 2025-04-06 19:18:05 +0200 | <emojelly> | type StT t a :: * |
| 2025-04-06 19:18:03 +0200 | <emojelly> | for illustration, pretend the same class would define: |
| 2025-04-06 19:17:58 +0200 | <haskellbridge> | <hellwolf> but there is a lot of limitations, usually it won't work for you |
| 2025-04-06 19:17:46 +0200 | <haskellbridge> | <hellwolf> then you may use typefamilydependencies |
| 2025-04-06 19:17:35 +0200 | <emojelly> | It's a type family |
| 2025-04-06 19:17:22 +0200 | <haskellbridge> | <hellwolf> I don't know if it applies to your case. But I do see "StT t a1", I wonder what is StT. |
| 2025-04-06 19:17:00 +0200 | <c_wraith> | But it needs to know what t is to select an instance |
| 2025-04-06 19:16:50 +0200 | <emojelly> | yeah but I'm not sure how to use a fundep here when there's only a single type? |
| 2025-04-06 19:16:47 +0200 | <haskellbridge> | <hellwolf> type family has similar mechanism; which has its own limitation. |
| 2025-04-06 19:16:45 +0200 | <c_wraith> | err, in the type of |
| 2025-04-06 19:16:34 +0200 | <haskellbridge> | <hellwolf> GHC will check that |
| 2025-04-06 19:16:26 +0200 | <emojelly> | (have to run, will read later, thanks for the help so far!) |
| 2025-04-06 19:16:22 +0200 | <haskellbridge> | This means that knowing "a" uniquely determins "b" |
| 2025-04-06 19:16:22 +0200 | <haskellbridge> | <hellwolf> class A a b | a -> b |