Newest at the top
| 2026-01-05 02:48:33 +0100 | humasect | (~humasect@dyn-192-249-132-90.nexicom.net) (Ping timeout: 260 seconds) |
| 2026-01-05 02:47:19 +0100 | geekosaur | -o geekosaur |
| 2026-01-05 02:47:03 +0100 | geekosaur | -r |
| 2026-01-05 02:46:51 +0100 | <oats> | gotcha, thanks |
| 2026-01-05 02:46:43 +0100 | ChanServ | +o geekosaur |
| 2026-01-05 02:46:10 +0100 | <c_wraith> | and @undo doesn't give you what GHC does. It is an independent reimplementation of the rules from the Haskell report. |
| 2026-01-05 02:45:37 +0100 | <c_wraith> | technically they are, but not in a way that influences semantics. There are some performance optimizations. |
| 2026-01-05 02:45:00 +0100 | <oats> | oh I was just trying out desugaring list comprehensions, because I thought I remembered someone telling me they were special to ghc in some way |
| 2026-01-05 02:44:28 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 2026-01-05 02:44:16 +0100 | <c_wraith> | [ x | x <- foo ] as opposed to do { x <- foo ; pure x } |
| 2026-01-05 02:43:57 +0100 | humasect | (~humasect@dyn-192-249-132-90.nexicom.net) humasect |
| 2026-01-05 02:43:53 +0100 | <oats> | ? |
| 2026-01-05 02:43:22 +0100 | <c_wraith> | You don't have to write pure. :) |
| 2026-01-05 02:42:56 +0100 | <oats> | don't remember who told me that |
| 2026-01-05 02:42:53 +0100 | <oats> | interesting, I thought list comprehensions were special in some way |
| 2026-01-05 02:42:43 +0100 | <int-e> | oats: Int (machine word sized), yes. |
| 2026-01-05 02:42:34 +0100 | <lambdabot> | concatMap (\ a -> case a of { Just x -> [x]; _ -> []}) [Just 1, Nothing] |
| 2026-01-05 02:42:34 +0100 | <oats> | @undo [x | Just x <- [Just 1, Nothing] ] |
| 2026-01-05 02:42:25 +0100 | <Axman6> | and fail str = [] for [] |
| 2026-01-05 02:41:45 +0100 | <oats> | int-e: I'm not super clear on #, that's an unboxed integer right? |
| 2026-01-05 02:41:37 +0100 | <lambdabot> | [Just 1, Nothing] >>= \ a -> case a of { Just x -> pure x; _ -> fail ""} |
| 2026-01-05 02:41:37 +0100 | <int-e> | @undo do { Just x <- [Just 1, Nothing]; pure x } |
| 2026-01-05 02:41:31 +0100 | <c_wraith> | the report actually specifies that. |
| 2026-01-05 02:41:05 +0100 | <oats> | wonder what that desugars to |
| 2026-01-05 02:40:33 +0100 | <yahb2> | [1] |
| 2026-01-05 02:40:33 +0100 | <oats> | % do { Just x <- [Just 1, Nothing]; pure x } |
| 2026-01-05 02:39:13 +0100 | <yahb2> | [1,2,3,4,6,7,8,9,10] |
| 2026-01-05 02:39:13 +0100 | <oats> | % do { x <- [1..10]; if (x == 5) then fail "bleh" else pure (); pure x } |
| 2026-01-05 02:38:49 +0100 | <haskellbridge> | <loonycyborg> case is one of few constructs that aren't syntactic sugar for something else. |
| 2026-01-05 02:38:11 +0100 | <int-e> | So up to that point, `case` was the only way to do this. `case` is also clearer about the order of operations. |
| 2026-01-05 02:38:10 +0100 | <c_wraith> | there are also things like GADT matching, where GHC just won't compile code that matches a GADT constructor in a let/where |
| 2026-01-05 02:37:25 +0100 | merijn | (~merijn@host-cl.cgnat-g.v4.dfn.nl) merijn |
| 2026-01-05 02:37:12 +0100 | <int-e> | oats: as recently as GHC 8.8.4, ghc would choke on `let i = 1#` (with or without bang) |
| 2026-01-05 02:37:06 +0100 | <haskellbridge> | <loonycyborg> Those seem to be really different things though, so it smells like some theoretical insight here could be made.. |
| 2026-01-05 02:35:13 +0100 | <haskellbridge> | <loonycyborg> but ye, irrefutable and lazy are connected. |
| 2026-01-05 02:35:01 +0100 | <yahb2> | <interactive>:159:5: warning: [GHC-53633] [-Woverlapping-patterns] ; Pattern match is redundant ; In a pattern binding: Just x = ... ; ; *** Exception: <interactive>:159:5-35: Non-exhausti... |
| 2026-01-05 02:35:01 +0100 | <oats> | % let Just x = (Nothing :: Maybe Int) in x |
| 2026-01-05 02:34:47 +0100 | <oats> | %let Just x = (Nothing :: Maybe Int) in x |
| 2026-01-05 02:34:30 +0100 | <c_wraith> | If it's on the left side of <- in a do block, yes |
| 2026-01-05 02:34:16 +0100 | <haskellbridge> | <loonycyborg> Sometimes should be converted into "fail" |
| 2026-01-05 02:34:02 +0100 | <c_wraith> | yeah, just the usual error about a pattern match failure |
| 2026-01-05 02:33:45 +0100 | <haskellbridge> | <loonycyborg> Same thing as usual? |
| 2026-01-05 02:33:33 +0100 | <haskellbridge> | <loonycyborg> and what happens in case of non-exhaustive match at runtime? |
| 2026-01-05 02:33:17 +0100 | <haskellbridge> | <loonycyborg> Ah makes sense I guess |
| 2026-01-05 02:32:14 +0100 | <c_wraith> | loonycyborg: You need the BangPatterns extension, then you mark the match with ! |
| 2026-01-05 02:31:49 +0100 | <haskellbridge> | <loonycyborg> Like you said they're irrefutable "by default" |
| 2026-01-05 02:31:47 +0100 | <haskellbridge> | <loonycyborg> How can you make patterns in let and where refutable? |
| 2026-01-05 02:30:33 +0100 | <int-e> | c_wraith: yes |
| 2026-01-05 02:29:56 +0100 | <c_wraith> | Yeah. "irrefutable" is a slightly weird way of saying "trust the programmer, you don't need to check this immediately" |
| 2026-01-05 02:29:49 +0100 | <oats> | t'would appear |