Newest at the top
| 2025-01-08 06:50:22 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
| 2025-01-08 06:42:26 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2025-01-08 06:39:34 +0100 | <monochrom> | If you want Nothing ??? pure foo = pure foo, consider using <|> instead: Nothing <|> pure foo. In other words, you are looking for { throw Exception; } catch Exception { return foo; } |
| 2025-01-08 06:37:49 +0100 | <monochrom> | But if you prefer pragmatics to theory, every exception system, in every language, says that { throw Exception; return (div 5 2); } is going to throw that exception rather than return normally. |
| 2025-01-08 06:36:40 +0100 | <monochrom> | Belated: Nothing >> pure foo = Nothing is pretty much forced by the monad laws. Several laws are involved, but the most damning one is k >>= pure = k, now let k be Nothing. |
| 2025-01-08 06:31:39 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 245 seconds) |
| 2025-01-08 06:27:03 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2025-01-08 06:15:36 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
| 2025-01-08 06:12:17 +0100 | forell | (~forell@user/forell) (Ping timeout: 248 seconds) |
| 2025-01-08 06:09:00 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2025-01-08 06:08:04 +0100 | <jackdk> | I like https://hackage.haskell.org/package/directory-tree for this sort of thing, unless you really need to handle large inputs (I haven't tested its lazy `readDirectoryWithL` function). Otherwise, yeah, streaming libraries would be the way to go. I tend to use package `streaming`. |
| 2025-01-08 06:03:12 +0100 | eL_Bart0 | (eL_Bart0@dietunichtguten.org) (Ping timeout: 246 seconds) |
| 2025-01-08 06:01:12 +0100 | <hololeap> | I'm ok with using conduit here |
| 2025-01-08 06:00:36 +0100 | <int-e> | I'm not arguing against that, it's a separate question |
| 2025-01-08 06:00:35 +0100 | <c_wraith> | that's kind of the whole point of it. |
| 2025-01-08 05:59:48 +0100 | <c_wraith> | which actually won't construct the whole list in memory before moving to the next step |
| 2025-01-08 05:59:36 +0100 | <c_wraith> | you have tools like https://hackage.haskell.org/package/conduit-1.3.6/docs/Data-Conduit-Combinators.html#v:sourceDirec… |
| 2025-01-08 05:59:25 +0100 | <c_wraith> | I mean, conduit really is better. |
| 2025-01-08 05:58:54 +0100 | Square2 | (~Square@user/square) (Ping timeout: 265 seconds) |
| 2025-01-08 05:58:09 +0100 | <int-e> | oh it's \m -> lift m >>= fromList |
| 2025-01-08 05:57:57 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 248 seconds) |
| 2025-01-08 05:57:00 +0100 | <hololeap> | this was the one use-case where ListT was actually the most elegant way to do it |
| 2025-01-08 05:56:54 +0100 | Feuermagier | (~Feuermagi@user/feuermagier) Feuermagier |
| 2025-01-08 05:54:44 +0100 | <lambdabot> | LogicT . (. foldr) . (.) . (>>=) |
| 2025-01-08 05:54:44 +0100 | <int-e> | @pl \x -> LogicT (\f c -> x >>= foldr f c) |
| 2025-01-08 05:53:53 +0100 | <c_wraith> | conduit *is* overkill, but it's at least overkill designed for the job. |
| 2025-01-08 05:53:30 +0100 | <c_wraith> | oh, I guess https://hackage.haskell.org/package/conduit-1.3.6/docs/Data-Conduit-Combinators.html is the better option |
| 2025-01-08 05:53:22 +0100 | <hololeap> | but maybe you're right |
| 2025-01-08 05:53:17 +0100 | <hololeap> | I felt like conduit would be overkill |
| 2025-01-08 05:53:10 +0100 | <c_wraith> | like https://hackage.haskell.org/package/conduit-1.3.6/docs/Data-Conduit-List.html exists specifically to replace ListT |
| 2025-01-08 05:52:42 +0100 | <c_wraith> | you really should be using conduit or something for that |
| 2025-01-08 05:52:27 +0100 | <hololeap> | https://github.com/hololeap/cabal-portage/blob/main/portage-hs/test/Emerge/ParserTests.hs#L66-L76 |
| 2025-01-08 05:52:22 +0100 | <hololeap> | c_wraith: I have some old code that uses ListT, and I'm trying to replace it so that it builds on newer base (since they removed ListT) |
| 2025-01-08 05:51:00 +0100 | <c_wraith> | LogicT really wants to break that apart |
| 2025-01-08 05:50:43 +0100 | <c_wraith> | the issue is that m [a] is a *single* action that produces an entire list. |
| 2025-01-08 05:50:39 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2025-01-08 05:49:59 +0100 | Square2 | (~Square@user/square) Square |
| 2025-01-08 05:49:50 +0100 | <c_wraith> | based on the type, it's not going to be as nice as you'd want |
| 2025-01-08 05:49:38 +0100 | peterbecich | (~Thunderbi@syn-047-229-123-186.res.spectrum.com) peterbecich |
| 2025-01-08 05:49:35 +0100 | Square2 | (~Square@user/square) (Remote host closed the connection) |
| 2025-01-08 05:49:31 +0100 | <hololeap> | blah :: IO [FilePath] -> LogicT IO FilePath |
| 2025-01-08 05:49:01 +0100 | <hololeap> | d <- blah $ listDirectory dataDir |
| 2025-01-08 05:48:39 +0100 | pavonia | (~user@user/siracusa) (Quit: Bye!) |
| 2025-01-08 05:47:56 +0100 | <hololeap> | err LogicT m a |
| 2025-01-08 05:47:45 +0100 | <hololeap> | anyone know if there is a canonical way to "lift" m [a] to ListT m a |
| 2025-01-08 05:42:57 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
| 2025-01-08 05:41:00 +0100 | Smiles | (uid551636@id-551636.lymington.irccloud.com) (Quit: Connection closed for inactivity) |
| 2025-01-08 05:38:34 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2025-01-08 05:36:29 +0100 | JuanDaugherty | (~juan@user/JuanDaugherty) (Quit: JuanDaugherty) |
| 2025-01-08 05:32:12 +0100 | Square2 | (~Square@user/square) Square |