Logs: liberachat/#haskell
| 2021-08-09 17:41:42 | <lechner> | i was going to say |
| 2021-08-09 17:41:44 | <dminuoso> | should probably read 'relude ^>= 0.7.0.0' |
| 2021-08-09 17:41:52 | → | lavaman joins (~lavaman@98.38.249.169) |
| 2021-08-09 17:42:17 | <lechner> | thanks for catching that! |
| 2021-08-09 17:42:26 | <maerwald[m]> | it isn't easy to get a language into the tight spot which is ease of refactoring, because too much expressivity can easily kill that property too |
| 2021-08-09 17:42:48 | Guest7145 | is now known as random-jellyfish |
| 2021-08-09 17:42:53 | × | random-jellyfish quits (~Guest71@217.156.31.64) (Changing host) |
| 2021-08-09 17:42:53 | → | random-jellyfish joins (~Guest71@user/random-jellyfish) |
| 2021-08-09 17:44:41 | → | merijn joins (~merijn@83-160-49-249.ip.xs4all.nl) |
| 2021-08-09 17:45:07 | <lechner> | dminuoso: i actually copied and pasted that from a constraint i no longer needed but something got lost, sorry my head is spinning https://github.com/mrkkrp/megaparsec/issues/461 |
| 2021-08-09 17:46:24 | → | myShoggoth joins (~myShoggot@97-120-70-214.ptld.qwest.net) |
| 2021-08-09 17:46:27 | → | MorrowM joins (~Morrow@176.12.187.199) |
| 2021-08-09 17:46:31 | × | lavaman quits (~lavaman@98.38.249.169) (Ping timeout: 258 seconds) |
| 2021-08-09 17:46:38 | <lechner> | dminuoso: thanks so much! i am in business. is this a static executable i can copy to anoter machine like Golang? |
| 2021-08-09 17:46:47 | <dminuoso> | Perhaps? |
| 2021-08-09 17:46:54 | <myShoggoth> | Episode 1 of The Haskell Interlude podcast is liiiiiiiiive: https://haskell.foundation/podcast/ |
| 2021-08-09 17:46:57 | <dminuoso> | It likely isnt if you dont know. |
| 2021-08-09 17:47:08 | <dminuoso> | It takes effort to make static executables with GHC. |
| 2021-08-09 17:47:23 | <MorrowM> | myShoggoth: Oooh |
| 2021-08-09 17:47:29 | <tomsmeding> | lechner: run ldd on the executable |
| 2021-08-09 17:48:18 | <tomsmeding> | myShoggoth: 🎉 |
| 2021-08-09 17:48:30 | <srid[m]> | <lechner> "srid: I am now also on Matrix..." <- If you are on matrix, join #neuron:matrix.org |
| 2021-08-09 17:48:30 | <lechner> | well, at 64 megs it better bring some goodies |
| 2021-08-09 17:48:42 | <dminuoso> | lechner: This is due to some unfortunate defaults. |
| 2021-08-09 17:48:53 | <lechner> | srid[m]: i did---with two accounts |
| 2021-08-09 17:49:03 | <dminuoso> | If you configure your cabal globally to enable split sections and stripping, this usually gets down to acceptable sizes. |
| 2021-08-09 17:49:22 | <tomsmeding> | was there a downside to that on linux again? |
| 2021-08-09 17:49:43 | <srid[m]> | lechner: also see https://github.com/srid/ema-template/issues/3 |
| 2021-08-09 17:49:51 | <tomsmeding> | ah docs say compile time and compile memory usage |
| 2021-08-09 17:50:00 | <dminuoso> | Set `split-sections: True` in your ~/.cabal/config, and ensure you have stripping enabled in your cabal project. Note that this will recompile *all* dependencies, but it's a setting that arugably should be a cabal default. |
| 2021-08-09 17:50:45 | <lechner> | can it create separate debug symbols? |
| 2021-08-09 17:51:29 | <maerwald[m]> | dminuoso: probably isn't because it doesn't work on all platforms or sth |
| 2021-08-09 17:51:44 | <geekosaur> | I don't think it creates debug symbols at all without specific configuration? |
| 2021-08-09 17:51:57 | <dminuoso> | maerwald[m]: Oh yeah, I recall that argument and I think its flawed. Splitting sections when it's possible seems very sensible. |
| 2021-08-09 17:52:17 | <geekosaur> | debug symbols are a recent addition to ghc |
| 2021-08-09 17:52:28 | <dminuoso> | Forcibly blowing up executable sizes by a factor of up to 25 just for the sake of some stubbornness is weird. |
| 2021-08-09 17:53:16 | <maerwald[m]> | I don't think any GHC dev is particularly stubborn... there are just bigger fish to fry :p |
| 2021-08-09 17:53:22 | <dminuoso> | heh |
| 2021-08-09 17:53:41 | <geekosaur> | this would be cabal, not ghc? |
| 2021-08-09 17:54:10 | <lechner> | Haskell executables may not produce much unexpected runtime behavior compared to other languages |
| 2021-08-09 17:54:56 | <dminuoso> | What do you mean exactly by unexpected runtime behavior? |
| 2021-08-09 17:55:47 | <fresheyeball> | I am working on a new project in plutus |
| 2021-08-09 17:55:57 | <fresheyeball> | and for some reason `cabal build` clones git repos |
| 2021-08-09 17:56:03 | <fresheyeball> | I have never seen cabal build do this |
| 2021-08-09 17:56:13 | <fresheyeball> | what's that all about? Is this new 3.0 stuff? |
| 2021-08-09 17:56:36 | <lechner> | dminuoso: the need use a debugger, although maybe you do have your fair share of OOMs |
| 2021-08-09 17:56:43 | <davean> | fresheyeball: thats cabal.project stuff, it just depends on a specific remote |
| 2021-08-09 17:56:47 | <sclv> | yep, cabal project files let you pin against git repos and their tags or hashes |
| 2021-08-09 17:57:11 | <sclv> | some bigger projects are making use of that a fair amount these days |
| 2021-08-09 17:57:36 | <dminuoso> | lechner: In principle you also need to use debugging tools in Haskell. |
| 2021-08-09 17:57:56 | <geekosaur> | there's at least one ghc bug reoort that begs to differ :þ https://gitlab.haskell.org/ghc/ghc/-/issues/163 |
| 2021-08-09 17:57:58 | <fresheyeball> | davean: is there a way to get cabal to use the deps provided by the nix-shell instead of downloading things? |
| 2021-08-09 17:58:01 | <maerwald[m]> | fresheyeball: cardano is full of source-repository packages. There are many libraries, but almost none or on hackage, so that's how they are pulled in |
| 2021-08-09 17:58:07 | × | ec quits (~ec@gateway/tor-sasl/ec) (Ping timeout: 244 seconds) |
| 2021-08-09 17:58:20 | <davean> | fresheyeball: Not and get a working build I expect |
| 2021-08-09 17:58:22 | <lechner> | dminuoso: i was trying to support your argument that debugging symbols may not be needed all the time |
| 2021-08-09 17:58:33 | <dminuoso> | It's hard to say whether you need it "less" or "more". From what I gather, the wide masses rarely use debugging tools, sadly. |
| 2021-08-09 17:58:35 | <sclv> | fresheyeball: only by changing the project file to not explicitly require those repos! |
| 2021-08-09 17:58:46 | <lechner> | geekosaur: ouch! |
| 2021-08-09 17:59:22 | <sclv> | lmao the famous "best bug ever" |
| 2021-08-09 17:59:24 | <lechner> | geekosaur: i know, it happened on Windoze! |
| 2021-08-09 17:59:57 | → | notzmv joins (~zmv@user/notzmv) |
| 2021-08-09 18:00:01 | × | cheater quits (~Username@user/cheater) (Ping timeout: 248 seconds) |
| 2021-08-09 18:00:05 | → | ec joins (~ec@gateway/tor-sasl/ec) |
| 2021-08-09 18:00:14 | <lechner> | sorry, deafening silence |
| 2021-08-09 18:01:02 | <tomsmeding> | love how there is no conversation on that ticket other than "will be fixed in a patch version" |
| 2021-08-09 18:01:26 | × | random-jellyfish quits (~Guest71@user/random-jellyfish) (Quit: Client closed) |
| 2021-08-09 18:01:51 | <lechner> | please let me say that i adore Microsoft's support for Haskell. sorry if i offended anyone. bad old habit |
| 2021-08-09 18:02:39 | <lechner> | i really meant that. my apologies |
| 2021-08-09 18:03:16 | → | cheater joins (~Username@user/cheater) |
| 2021-08-09 18:03:22 | <maerwald[m]> | Microsoft Research |
| 2021-08-09 18:03:23 | <maerwald[m]> | which is more like a completely independent thing |
| 2021-08-09 18:03:42 | <lechner> | dminuoso: isn't stripping on by default? https://cabal.readthedocs.io/en/3.4/cabal-project.html#cfg-field-executable-stripping |
| 2021-08-09 18:03:56 | <lechner> | maerwald[m]: crdeit has to be given when due |
| 2021-08-09 18:04:09 | <dminuoso> | lechner: I dont know about your cabal installation, configuration, cabal.project/cabal.project.local |
| 2021-08-09 18:04:10 | <lechner> | credit, sorry poor speller |
| 2021-08-09 18:04:21 | <monochrom> | Yes, this is why the correction "MS Research" is important. |
| 2021-08-09 18:04:54 | <dminuoso> | lechner: Which is why I said "ensure you have..." |
| 2021-08-09 18:05:22 | <dminuoso> | And maybe this changed in cabal versions, i have no clue. |
| 2021-08-09 18:05:26 | <maerwald[m]> | MS Research... oddly... is pretty insanely great. The projects they do, the papers, the reserachers |
| 2021-08-09 18:05:36 | <lechner> | yeah |
| 2021-08-09 18:05:38 | <monochrom> | Approximately the only credit for MS the parent company is the money. |
| 2021-08-09 18:06:03 | <monochrom> | Oh the history backdrop is even more amazing. |
| 2021-08-09 18:06:10 | <maerwald[m]> | excess money needs to be spent somehow |
| 2021-08-09 18:06:20 | <monochrom> | In the beginning, Bell Labs was insanely great. |
| 2021-08-09 18:06:40 | <lechner> | aside from the money, they also had the wits not to pour it into Pascal |
| 2021-08-09 18:07:04 | → | Obo joins (~roberto@70.pool90-171-81.dynamic.orange.es) |
| 2021-08-09 18:07:18 | → | drewr joins (~drew@user/drewr) |
| 2021-08-09 18:07:21 | <geekosaur> | Pascal-izing C (see ther original Windows API) isn't necessarily better thiugh |
| 2021-08-09 18:07:45 | <monochrom> | On day, Bell the parent company became not so friendly with the research branch, so the great researchers moved to DEC Lab. |
| 2021-08-09 18:08:10 | <monochrom> | So DEC Lab was insanely great. |
| 2021-08-09 18:09:15 | <monochrom> | Then one day, DEC sold the lab to Compaq, and later Compaq got bought up by Dell. So the parent company, whichever it was at the time, became not so friendly to the lab, so the great researches moved to MS Research. |
| 2021-08-09 18:09:26 | <monochrom> | So MS Research became insanely great. |
| 2021-08-09 18:10:00 | <monochrom> | History can repeat again. If one day MS becomes unfriendly to MS Research, the great researchers can still move to another company and make it great. |
| 2021-08-09 18:10:09 | <lechner> | i thought Apple had trademarked "insanely" |
| 2021-08-09 18:10:28 | <lechner> | or rather "insanely great" |
| 2021-08-09 18:11:21 | <monochrom> | Oh, Windows API being pascalish also has its historical backdrop, probably. |
| 2021-08-09 18:11:38 | <Drew[m]1> | Compaq merged with HP, not Dell |
| 2021-08-09 18:11:55 | <davean> | lechner: so what if they did? |
| 2021-08-09 18:11:55 | <monochrom> | So it began with Classical MacOS choosing Pascal and giving you only a Pascal binding of its OS and GUI. |
All times are in UTC.