Newest at the top
| 2024-11-16 23:33:41 +0100 | <hellwolf> | not your garden variety of senior engineer |
| 2024-11-16 23:33:31 +0100 | <hellwolf> | oh, that's a nice CV title "Type-level programmer" |
| 2024-11-16 23:33:04 +0100 | <hellwolf> | I haven't applied to an actual type yet. let me keep coercing vim to generate code for me. |
| 2024-11-16 23:33:02 +0100 | <geekosaur> | there's probably some type level programmer who's hit it |
| 2024-11-16 23:32:39 +0100 | <geekosaur> | best not to ask that question, I suspect 🙂 |
| 2024-11-16 23:32:22 +0100 | <hellwolf> | in what the actual practical case that error message can be useful? |
| 2024-11-16 23:32:12 +0100 | <hellwolf> | I mean, why would you suggest that? |
| 2024-11-16 23:32:05 +0100 | <hellwolf> | I am just glad that the compiler error messages suggest me to go on that dark path |
| 2024-11-16 23:31:50 +0100 | <geekosaur> | did you actually apply that? |
| 2024-11-16 23:31:36 +0100 | <geekosaur> | surprised they didn't simply consider that some weird Lisp 🙂 |
| 2024-11-16 23:31:31 +0100 | <hellwolf> | so, we shall declare that, the Glorious Haskell Compile can support unlimited amount of constraints. |
| 2024-11-16 23:30:52 +0100 | <hellwolf> | type As :: Type -> Constraint |
| 2024-11-16 23:30:29 +0100 | <hellwolf> | https://paste.tomsmeding.com/fTQmMzsa |
| 2024-11-16 23:30:24 +0100 | <geekosaur> | huh. recent pastes shows a bunch of gift card ads but they censor you? |
| 2024-11-16 23:30:14 +0100 | <lambdabot> | Help us help you: please paste full code, input and/or output at e.g. https://paste.tomsmeding.com |
| 2024-11-16 23:30:14 +0100 | <int-e> | @where paste |
| 2024-11-16 23:30:11 +0100 | <hellwolf> | would i flood the screen if I do it here |
| 2024-11-16 23:30:00 +0100 | <hellwolf> | oh, being censorred |
| 2024-11-16 23:29:42 +0100 | <hellwolf> | 1000 of them |
| 2024-11-16 23:29:40 +0100 | <geekosaur> | 404 |
| 2024-11-16 23:29:33 +0100 | <hellwolf> | (a, (..)). A NP constraint. |
| 2024-11-16 23:29:16 +0100 | <hellwolf> | https://pastebin.com/BTCYzKLN |
| 2024-11-16 23:29:02 +0100 | geekosaur | wonders at what point the typechecker goes "wtf are you asking me to solve?!" |
| 2024-11-16 23:28:39 +0100 | <hellwolf> | I did it |
| 2024-11-16 23:28:38 +0100 | <hellwolf> | yep |
| 2024-11-16 23:26:43 +0100 | <hellwolf> | let me try... |
| 2024-11-16 23:26:38 +0100 | <hellwolf> | It encourages me to go a bit crazy, isn't it |
| 2024-11-16 23:26:12 +0100 | <hellwolf> | what the sorcery was the suggestino. |
| 2024-11-16 23:25:55 +0100 | <hellwolf> | "Instead, use a nested tuple) |
| 2024-11-16 23:25:33 +0100 | <hellwolf> | sorry, function declaration |
| 2024-11-16 23:25:24 +0100 | <hellwolf> | while the class declaration does not. |
| 2024-11-16 23:25:14 +0100 | <geekosaur> | interesting |
| 2024-11-16 23:25:01 +0100 | <hellwolf> | constraint kind subjects to the same limit |
| 2024-11-16 23:24:49 +0100 | <hellwolf> | Instead, use a nested tuple |
| 2024-11-16 23:24:49 +0100 | <hellwolf> | • Constraint tuple arity too large: 1000 (max arity = 64) |
| 2024-11-16 23:24:49 +0100 | <hellwolf> | constraintslim.hs:2008:13: error: [GHC-89347] |
| 2024-11-16 23:24:39 +0100 | peterbecich | (~Thunderbi@syn-047-229-123-186.res.spectrum.com) (Ping timeout: 260 seconds) |
| 2024-11-16 23:22:16 +0100 | <geekosaur> | at Core level the names are "f$…" iirc (the … is the z-encoded class name and type parameters) |
| 2024-11-16 23:21:30 +0100 | <geekosaur> | I guess program-wide uniqueness means they have to be named |
| 2024-11-16 23:21:17 +0100 | <hellwolf> | (max size is 64) |
| 2024-11-16 23:21:17 +0100 | <hellwolf> | • A 100-tuple is too large for GHC |
| 2024-11-16 23:21:17 +0100 | <hellwolf> | tuplelimits.hs:1:6: error: [GHC-94803] |
| 2024-11-16 23:21:15 +0100 | <hellwolf> | sorry, wrogn copy paste |
| 2024-11-16 23:21:06 +0100 | <hellwolf> | say no more, let me write a test case |
| 2024-11-16 23:20:11 +0100 | Square | (~Square@user/square) (Ping timeout: 244 seconds) |
| 2024-11-16 23:18:55 +0100 | <geekosaur> | actually (b) is wrong, instance dictionaries are value level but bypass the limit unless they're compiled internally to magic named records (I know they are at Core level, but not sure about before that) |
| 2024-11-16 23:11:28 +0100 | <hellwolf> | okay, made the classes less trivial, with one function only. still okay, but notice a tiny little delay when loading through ghci |
| 2024-11-16 23:11:11 +0100 | michalz | (~michalz@185.246.207.203) (Remote host closed the connection) |
| 2024-11-16 23:10:02 +0100 | <hellwolf> | 1000s is fine too, I hope the empty class did not mattered. |
| 2024-11-16 23:08:31 +0100 | <hellwolf> | 100 constraints was fine |