Newest at the top
| 2025-03-12 16:22:47 +0100 | pavonia | (~user@user/siracusa) (Quit: Bye!) |
| 2025-03-12 16:14:08 +0100 | infinity0 | (~infinity0@pwned.gg) (Ping timeout: 245 seconds) |
| 2025-03-12 16:07:05 +0100 | infinity0 | (~infinity0@pwned.gg) infinity0 |
| 2025-03-12 16:00:39 +0100 | <[exa]> | tomsmeding: iiiiiiiinteresting thanks! |
| 2025-03-12 15:54:03 +0100 | T0NN | (~T0NN@193.42.62.209) |
| 2025-03-12 15:50:33 +0100 | alfiee | (~alfiee@user/alfiee) (Ping timeout: 268 seconds) |
| 2025-03-12 15:47:40 +0100 | lottaquestions | (~nick@2607:fa49:5040:b100:5b52:ac7b:b1b7:f59f) lottaquestions |
| 2025-03-12 15:45:39 +0100 | alfiee | (~alfiee@user/alfiee) alfiee |
| 2025-03-12 15:34:53 +0100 | TheCoffeMaker | (~TheCoffeM@user/thecoffemaker) TheCoffeMaker |
| 2025-03-12 15:31:11 +0100 | <carbolymer> | but in my quick test `sortBy (\_ _ -> GT)` results in just reversed list |
| 2025-03-12 15:30:39 +0100 | <carbolymer> | I guess it's not true for all comparison functions in sortBy |
| 2025-03-12 15:30:39 +0100 | <carbolymer> | >otherwise, e. g., for _ _ -> GT, the ordered list simply does not exist |
| 2025-03-12 15:29:21 +0100 | <tomsmeding> | it would be nice if this was also mentioned in 'sortBy', but I think we can extrapolate |
| 2025-03-12 15:28:59 +0100 | <tomsmeding> | carbolymer: from the haddocks of 'sort': The sort function implements a stable sorting algorithm. |
| 2025-03-12 15:28:57 +0100 | <carbolymer> | that was my intuition, thanks tomsmeding for confirming ;] |
| 2025-03-12 15:28:10 +0100 | T0NN | (~T0NN@104.28.196.52) (Client Quit) |
| 2025-03-12 15:27:39 +0100 | <tomsmeding> | for longer lists, it splits the list into runs of alternatingly ascending/descending elements; in your case there's just one run, so the algorithm terminates there |
| 2025-03-12 15:27:14 +0100 | <tomsmeding> | looking at the implementation, for lists of length <=4 it has a manually written decision tree that, if I am to believe the comments, should only reorder if a preceding element is strictly greater than a following element |
| 2025-03-12 15:26:51 +0100 | T0NN | (~T0NN@104.28.196.52) |
| 2025-03-12 15:26:35 +0100 | T0NN | (~T0NN@104.28.196.52) (Client Quit) |
| 2025-03-12 15:26:14 +0100 | <tomsmeding> | I'm fairly sure yes |
| 2025-03-12 15:25:55 +0100 | <carbolymer> | even simpler: `sortBy (\_ _ -> LT)` |
| 2025-03-12 15:25:23 +0100 | <carbolymer> | `sortBy compare` I mean |
| 2025-03-12 15:25:15 +0100 | T0NN | (~T0NN@104.28.196.52) |
| 2025-03-12 15:24:54 +0100 | <carbolymer> | is Data.List.sortBy guaranteed to be stable over `[Foo]` if I have a trivial `instance Ord Foo where compare _ _ = LT` ? In other words, I expect it to not change the order of elements. |
| 2025-03-12 15:24:48 +0100 | TheCoffeMaker | (~TheCoffeM@user/thecoffemaker) (Ping timeout: 252 seconds) |
| 2025-03-12 15:23:35 +0100 | T0NN | (~T0NN@104.28.196.52) (Client Quit) |
| 2025-03-12 15:23:18 +0100 | son0p | (~ff@2800:e6:4000:d723:c181:4205:f2b1:437a) son0p |
| 2025-03-12 15:22:55 +0100 | T0NN | (~T0NN@104.28.196.52) |
| 2025-03-12 15:19:23 +0100 | alexherbo2 | (~alexherbo@2a02-8440-3504-140e-55c2-d7c2-899b-0789.rev.sfr.net) alexherbo2 |
| 2025-03-12 15:07:17 +0100 | acidjnk_new | (~acidjnk@p200300d6e7283f52210926b325fe4262.dip0.t-ipconnect.de) |
| 2025-03-12 15:03:59 +0100 | alfiee | (~alfiee@user/alfiee) (Ping timeout: 260 seconds) |
| 2025-03-12 15:03:11 +0100 | son0p | (~ff@2800:e6:4000:d723:c181:4205:f2b1:437a) (Read error: Connection reset by peer) |
| 2025-03-12 14:59:14 +0100 | alfiee | (~alfiee@user/alfiee) alfiee |
| 2025-03-12 14:59:00 +0100 | CiaoSen | (~Jura@2a02:8071:64e1:7180::ac59) (Ping timeout: 265 seconds) |
| 2025-03-12 14:54:37 +0100 | forell | (~forell@user/forell) forell |
| 2025-03-12 14:52:26 +0100 | <Inst> | btw probie thanks for teaching me how to abuse foldr on lists; after playing around for optimization, foldr is almost literally a for loop due to foldr triggering list fusion |
| 2025-03-12 14:52:15 +0100 | forell_ | (~forell@host-178-216-90-220.sta.tvknaszapraca.pl) (Quit: ZNC - https://znc.in) |
| 2025-03-12 14:51:08 +0100 | Inst | (~Inst@user/Inst) Inst |
| 2025-03-12 14:41:47 +0100 | weary-traveler | (~user@user/user363627) user363627 |
| 2025-03-12 14:38:49 +0100 | weary-traveler | (~user@user/user363627) (Remote host closed the connection) |
| 2025-03-12 14:35:57 +0100 | <tomsmeding> | with git master versions of stuff, you can use LLVM 15, still static _and_ dynamic libraries, which is easy on current ubuntu and macOS |
| 2025-03-12 14:35:21 +0100 | <tomsmeding> | this was easy on ubuntu a few years ago |
| 2025-03-12 14:35:13 +0100 | <tomsmeding> | with the hackage-released versions of packages, you also need to install LLVM 9, both static and dynamic libraries |
| 2025-03-12 14:34:54 +0100 | <tomsmeding> | yes |
| 2025-03-12 14:34:50 +0100 | <kuribas> | Shouldn't it be faster than massiv/hmatrix then? |
| 2025-03-12 14:33:32 +0100 | <tomsmeding> | we hope to get it in somewhere this spring, but no guarantees |
| 2025-03-12 14:33:22 +0100 | <tomsmeding> | I have an unmerged patch to accelerate to not link to LLVM but pass LLVM IR to clang instead, which makes the whole thing a lot easier to install |
| 2025-03-12 14:33:15 +0100 | <merijn> | Storable Vectors are great |
| 2025-03-12 14:32:51 +0100 | <tomsmeding> | yes |