Newest at the top
| 2025-05-12 06:15:09 +0200 | jmcantrell | (~weechat@user/jmcantrell) (Ping timeout: 252 seconds) |
| 2025-05-12 06:14:33 +0200 | <ski> | (percent-encoded in this case. however, <https://en.wikipedia.org/wiki/Curry's_paradox> also works) |
| 2025-05-12 06:13:55 +0200 | <ski> | "o" with diaresis |
| 2025-05-12 06:13:47 +0200 | <ski> | it's an "รถ" |
| 2025-05-12 06:13:31 +0200 | <EvanR> | what is the character between L and b my client exploded |
| 2025-05-12 06:13:06 +0200 | <ski> | using type-level recursion, rather than value-level recursion, yes |
| 2025-05-12 06:12:51 +0200 | ski | . o O ( <https://en.wikipedia.org/wiki/L%C3%B6b's_paradox> ) |
| 2025-05-12 06:12:27 +0200 | <EvanR> | oh it's the Y combinator smuggled into haskell |
| 2025-05-12 06:11:57 +0200 | <ski> | by-name |
| 2025-05-12 06:11:30 +0200 | <EvanR> | is this by value santa |
| 2025-05-12 06:11:11 +0200 | <ski> | er, s/t = Santa/t = MkSanta/ |
| 2025-05-12 06:10:41 +0200 | <ski> | t = Santa (\t -> f (t `runSanta` t)) |
| 2025-05-12 06:10:25 +0200 | <ski> | where |
| 2025-05-12 06:10:20 +0200 | <ski> | fix f = t `runSanta` t |
| 2025-05-12 06:10:06 +0200 | <ski> | fix :: (a -> a) -> a |
| 2025-05-12 06:09:49 +0200 | <ski> | newtype Santa a = MkSanta {runSanta :: Santa a -> a} |
| 2025-05-12 06:07:40 +0200 | <EvanR> | it's turtles all the way down |
| 2025-05-12 06:06:51 +0200 | <ski> | indeed |
| 2025-05-12 06:06:04 +0200 | <EvanR> | ...) -> a) -> a) -> a) -> a |
| 2025-05-12 06:04:51 +0200 | <ski> | `omega :: o where o = o -> a' in Haskellish terms |
| 2025-05-12 06:03:57 +0200 | <ski> | ^ shorter way to express the same type |
| 2025-05-12 06:03:47 +0200 | <ski> | val omega : 'b -> 'a as 'b = <fun> |
| 2025-05-12 06:03:44 +0200 | <ski> | # let omega : 'o -> 'a as 'o = fun x -> x x;; |
| 2025-05-12 06:01:09 +0200 | <ski> | (this is useful for "binary methods" and "clone methods / functional update". see <https://ocaml.org/manual/5.3/objectexamples.html#s:functional-objects> and <https://ocaml.org/manual/5.3/objectexamples.html#s:binary-methods>) |
| 2025-05-12 05:58:15 +0200 | <ski> | without `-rectypes', it only allows such cycles, if they go through at least one object type |
| 2025-05-12 05:58:04 +0200 | <monochrom> | OK that's deep magic. |
| 2025-05-12 05:57:50 +0200 | <ski> | `ocaml -rectypes' does, yes |
| 2025-05-12 05:57:41 +0200 | <monochrom> | Oh God it does equi-recursive types?! |
| 2025-05-12 05:57:33 +0200 | <EvanR> | it's an infinite type |
| 2025-05-12 05:57:10 +0200 | <ski> | val omega : ('a -> 'b as 'a) -> 'b = <fun> |
| 2025-05-12 05:57:08 +0200 | <ski> | # let omega x = x x;; |
| 2025-05-12 05:57:06 +0200 | <EvanR> | this is some fun t business |
| 2025-05-12 05:56:29 +0200 | <monochrom> | Wait, how does that accept that "t t" is typeable? |
| 2025-05-12 05:56:11 +0200 | <EvanR> | alright let me think |
| 2025-05-12 05:54:46 +0200 | <ski> | - : int = 144 |
| 2025-05-12 05:54:43 +0200 | <ski> | # fix (fun fib n -> match n with 0 -> 0 | 1 -> 1 | n -> fib (n - 1) + fib (n - 2)) 12;; |
| 2025-05-12 05:54:33 +0200 | <ski> | val fix : (('a -> 'b) -> 'a -> 'b) -> 'a -> 'b = <fun> |
| 2025-05-12 05:54:28 +0200 | <ski> | # let fix f = (fun t -> t t) (fun t x -> f (t t) x);; |
| 2025-05-12 05:53:53 +0200 | <EvanR> | uhg |
| 2025-05-12 05:53:41 +0200 | <monochrom> | No, it's designed to be compatible with by-value evaluation. |
| 2025-05-12 05:53:03 +0200 | <EvanR> | does the by value version need a different evaluation strategy |
| 2025-05-12 05:52:27 +0200 | <EvanR> | ok but the point of a fix construct is so you don't have to munge these nuts combinators, and so they can have a type |
| 2025-05-12 05:50:41 +0200 | <monochrom> | It is correct. :) |
| 2025-05-12 05:50:26 +0200 | <EvanR> | I probably messed up |
| 2025-05-12 05:50:16 +0200 | <EvanR> | \f -> (\t -> t t) (\x -> f (\v -> (x x v))) |
| 2025-05-12 05:49:03 +0200 | <ski> | yea, i prefer adding in `(\t -> t t)' for the first part, no need to repeat something longer |
| 2025-05-12 05:48:04 +0200 | <EvanR> | but yours is more explainable |
| 2025-05-12 05:47:42 +0200 | <EvanR> | I found this on some page Z = \f -> (\x -> (f (\v -> (x x v)))) (\x -> (f (\v -> (x x v)))) |
| 2025-05-12 05:47:33 +0200 | michalz | (~michalz@185.246.207.203) |
| 2025-05-12 05:47:01 +0200 | <monochrom> | So it is missing out on fix (0 :) = 0 : 0 : ... but it's OK, call-by-value languages don't have lazy lists in the first place. :) |