Logs: liberachat/#haskell
| 2021-08-23 12:52:16 | <[exa]> | Jack: not sure about parsec but megaparsec has `getSourcePos` |
| 2021-08-23 12:52:33 | → | amahl joins (~amahl@dsl-jklbng12-54fbca-64.dhcp.inet.fi) |
| 2021-08-23 12:53:54 | <[exa]> | ha, parsec has equivalent `getPosition` |
| 2021-08-23 12:55:56 | → | keutoi joins (~keutoi@157.48.226.59) |
| 2021-08-23 12:59:12 | → | shriekingnoise joins (~shrieking@186.137.144.80) |
| 2021-08-23 13:00:37 | → | machinedgod joins (~machinedg@135-23-192-217.cpe.pppoe.ca) |
| 2021-08-23 13:01:53 | × | burnsidesLlama quits (~burnsides@dhcp168-023.wadham.ox.ac.uk) (Remote host closed the connection) |
| 2021-08-23 13:07:46 | × | tremon_ quits (~tremon@217-63-61-89.cable.dynamic.v4.ziggo.nl) (Ping timeout: 240 seconds) |
| 2021-08-23 13:09:17 | → | drd joins (~drd@2001:b07:a70:9f1f:1562:34de:f50f:77d4) |
| 2021-08-23 13:10:58 | → | markpythonicbtc joins (~textual@50.228.44.6) |
| 2021-08-23 13:11:10 | <ggVGc> | isn't it about time Haskell gets one officially standard parsing library |
| 2021-08-23 13:12:14 | <tdammers> | "official" by whose authority? |
| 2021-08-23 13:12:57 | <merijn> | We already have an official standard parsing library in base :p |
| 2021-08-23 13:13:03 | × | pavonia quits (~user@user/siracusa) (Quit: Bye!) |
| 2021-08-23 13:13:11 | <merijn> | Actually, two, I think |
| 2021-08-23 13:13:15 | <merijn> | And they suck :p |
| 2021-08-23 13:13:25 | <tdammers> | readP? |
| 2021-08-23 13:13:34 | <merijn> | and readS |
| 2021-08-23 13:13:40 | <tdammers> | ah yes |
| 2021-08-23 13:15:10 | × | Ariakenom quits (~patrik@c83-255-154-140.bredband.tele2.se) (Ping timeout: 240 seconds) |
| 2021-08-23 13:22:02 | × | acidjnk_new quits (~acidjnk@p200300d0c72b9556a59b812708dd2de4.dip0.t-ipconnect.de) (Ping timeout: 250 seconds) |
| 2021-08-23 13:22:35 | <[exa]> | ggVGc: y tho |
| 2021-08-23 13:22:38 | → | Ariakenom joins (~patrik@c83-255-154-140.bredband.tele2.se) |
| 2021-08-23 13:25:55 | → | sim590 joins (~simon@modemcable090.207-203-24.mc.videotron.ca) |
| 2021-08-23 13:27:00 | × | gehmehgeh quits (~user@user/gehmehgeh) (Quit: Leaving) |
| 2021-08-23 13:27:09 | <merijn> | If I've learned anything from my time in python, it's that "officially blessed and in the standard library" has a 100% fatality rate on libraries :p |
| 2021-08-23 13:28:06 | × | Ariakenom quits (~patrik@c83-255-154-140.bredband.tele2.se) (Ping timeout: 250 seconds) |
| 2021-08-23 13:28:07 | → | byorgey joins (~byorgey@155.138.238.211) |
| 2021-08-23 13:31:23 | <ggVGc> | I just wish that wasn't the case |
| 2021-08-23 13:31:45 | <ggVGc> | imagine if we could collectively stop writing and rewriting 20 versions of the same things all the time |
| 2021-08-23 13:35:08 | <[exa]> | producing quality packages requires training on other ones |
| 2021-08-23 13:35:49 | <maerwald> | yeah, I will support streamly based parsers in base |
| 2021-08-23 13:35:52 | → | burnsidesLlama joins (~burnsides@client-8-88.eduroam.oxuni.org.uk) |
| 2021-08-23 13:35:59 | → | keutoi_ joins (~keutoi@223.238.95.231) |
| 2021-08-23 13:36:29 | <merijn> | ggVGc: That seems an orthogonal issue |
| 2021-08-23 13:36:33 | × | keutoi quits (~keutoi@157.48.226.59) (Killed (lithium.libera.chat (Nickname regained by services))) |
| 2021-08-23 13:36:33 | keutoi_ | is now known as keutoi |
| 2021-08-23 13:36:43 | <merijn> | ggVGc: If there is a good and adequate parser library, what's stopping you from using it |
| 2021-08-23 13:37:11 | <merijn> | Like, how would an "officially blessed and in base" version of megaparsec be better than, just megaparsec? |
| 2021-08-23 13:37:42 | <[exa]> | also I guess we can kinda agree that almost all SI-prefixed variants of parsec have their own place |
| 2021-08-23 13:37:46 | <maerwald> | well... haskell is really bad at composition |
| 2021-08-23 13:37:47 | × | burnsidesLlama quits (~burnsides@client-8-88.eduroam.oxuni.org.uk) (Remote host closed the connection) |
| 2021-08-23 13:37:53 | <maerwald> | more stuff in base can improve that |
| 2021-08-23 13:37:54 | <ggVGc> | I believe the possible positive outcomes are always more focused efforts, and better documentation |
| 2021-08-23 13:37:55 | → | slowButPresent joins (~slowButPr@user/slowbutpresent) |
| 2021-08-23 13:38:02 | <merijn> | maerwald: So is pretty much every mainstream language :p |
| 2021-08-23 13:38:06 | <maerwald> | merijn: no |
| 2021-08-23 13:38:20 | <maerwald> | the weaker the types, the easier the composition, sometimes |
| 2021-08-23 13:38:32 | <maerwald> | since haskells API is all very type heavy, you can't easily exchange implementations |
| 2021-08-23 13:38:53 | <maerwald> | that's often less of an issue in, say, Go |
| 2021-08-23 13:39:10 | <maerwald> | because the few types they use are everywhere |
| 2021-08-23 13:39:13 | <[exa]> | can we count in the debugging time? :D |
| 2021-08-23 13:39:26 | <merijn> | maerwald: Depends how you define "easier", sure you can combine more things, but I assumed we only cared about *bugfree* combinations :p |
| 2021-08-23 13:40:05 | <maerwald> | merijn: you don't have bugfree anywhere |
| 2021-08-23 13:40:19 | <merijn> | maerwald: I also disagree that's an inherent flaw in Haskell, it's just a flaw in Haskell *programmers* to make their API's types inconvenient |
| 2021-08-23 13:40:45 | <maerwald> | sure, it's an ecosystem effect, but the language is more convenient for it |
| 2021-08-23 13:40:48 | <maerwald> | so it's both sides |
| 2021-08-23 13:40:49 | × | jgeerds quits (~jgeerds@55d4b311.access.ecotel.net) (Ping timeout: 248 seconds) |
| 2021-08-23 13:41:04 | × | azeem quits (~azeem@dynamic-adsl-94-34-33-6.clienti.tiscali.it) (Ping timeout: 252 seconds) |
| 2021-08-23 13:42:02 | <maerwald> | people thought backpack might solve this, but lol |
| 2021-08-23 13:42:14 | → | azeem joins (~azeem@176.201.15.223) |
| 2021-08-23 13:42:22 | <merijn> | I blame pytorch hijacking ezyang |
| 2021-08-23 13:42:24 | <maerwald> | you still have to agree on an API, like... what's parsing, what's streaming etc |
| 2021-08-23 13:44:06 | <maerwald> | and looking at conduit and streamly, the APIs are already incompatible |
| 2021-08-23 13:44:07 | <maerwald> | inherently |
| 2021-08-23 13:44:25 | <maerwald> | so backpack is not the solution |
| 2021-08-23 13:45:22 | <maerwald> | if we had an actual functioning CLC that is working on base actively, I don't think it would be an issue to broaden the scope |
| 2021-08-23 13:45:29 | → | acidjnk_new joins (~acidjnk@p200300d0c72b9556a59b812708dd2de4.dip0.t-ipconnect.de) |
| 2021-08-23 13:45:50 | <[exa]> | ...write a grant? |
| 2021-08-23 13:46:10 | × | drd quits (~drd@2001:b07:a70:9f1f:1562:34de:f50f:77d4) (Ping timeout: 240 seconds) |
| 2021-08-23 13:46:23 | <maerwald> | money solves it? |
| 2021-08-23 13:46:27 | <[exa]> | honestly I think e.g. MS research would not think twice about this precise ecosystem improvement |
| 2021-08-23 13:46:32 | <[exa]> | well, you need people to do that |
| 2021-08-23 13:46:37 | <maerwald> | HF has money |
| 2021-08-23 13:46:47 | <merijn> | I mean, base's problem is that it can't move fast without breaking *everything* |
| 2021-08-23 13:46:58 | <merijn> | Not to mention, significant parts are in the report |
| 2021-08-23 13:47:31 | → | waleee joins (~waleee@2001:9b0:216:8200:d457:9189:7843:1dbd) |
| 2021-08-23 13:47:58 | <[exa]> | maerwald: still it somehow chooses to work on other stuff |
| 2021-08-23 13:48:02 | <maerwald> | yes |
| 2021-08-23 13:48:15 | → | burnsidesLlama joins (~burnsides@client-8-88.eduroam.oxuni.org.uk) |
| 2021-08-23 13:48:27 | <[exa]> | maerwald: anyway the issue might just be more problematic, driven by actual programming deficiency than just people being lazy |
| 2021-08-23 13:48:35 | <maerwald> | they're more concerned with broadening Haskell adoption in industry I think... at least that's what I gather from their mission statement discussion |
| 2021-08-23 13:49:11 | <[exa]> | writing super-compatible stuff that still has types and not just "FromResult a => IO a" takes insane amounts of abstraction |
| 2021-08-23 13:49:18 | × | burnsidesLlama quits (~burnsides@client-8-88.eduroam.oxuni.org.uk) (Remote host closed the connection) |
| 2021-08-23 13:49:27 | <maerwald> | broadening adoption is more about building a brand/an image, providing consulting services, etc etc |
| 2021-08-23 13:49:31 | <maerwald> | less about improving ecosystem |
| 2021-08-23 13:49:36 | <[exa]> | monkeypatching stuff together (python/ruby way yay) certainly works but that's one thing we wanted to prevent right? |
| 2021-08-23 13:50:07 | <[exa]> | adoption broads itself with killer apps |
| 2021-08-23 13:51:22 | <[exa]> | anyway I'd certainly vote for having the internals of the packages exposed a bit more, having some way of recombining them in a little different way/with different glue types in case it's needed |
| 2021-08-23 13:51:51 | <[exa]> | the problem is that there's no such now, except for 1] having 100x more type variables 2] backpack, kinda |
| 2021-08-23 13:52:07 | <[exa]> | *no such standardized way to do that |
| 2021-08-23 13:52:36 | <[exa]> | not standardized -- not even available |
| 2021-08-23 13:55:28 | <[exa]> | hm now that's a research topic, assurance of some api compatibility without standardization |
| 2021-08-23 13:55:33 | <[exa]> | ...let's write a grant? |
| 2021-08-23 13:56:49 | × | aguapesada quits (~aguapesad@191.177.175.57) (Ping timeout: 248 seconds) |
| 2021-08-23 13:57:15 | → | burnsidesLlama joins (~burnsides@client-8-88.eduroam.oxuni.org.uk) |
| 2021-08-23 13:59:46 | × | enoq quits (~enoq@2a05:1141:1f5:5600:b9c9:721a:599:bfe7) (Quit: enoq) |
| 2021-08-23 14:00:37 | × | burnsidesLlama quits (~burnsides@client-8-88.eduroam.oxuni.org.uk) (Remote host closed the connection) |
| 2021-08-23 14:01:55 | <jacks2> | [exa], nice, thanks! |
| 2021-08-23 14:02:37 | → | eggplantade joins (~Eggplanta@2600:1700:bef1:5e10:71e9:72b:1860:cb0f) |
All times are in UTC.