Logs: liberachat/#haskell
| 2021-06-30 13:27:40 | × | shapr quits (~user@pool-108-28-144-11.washdc.fios.verizon.net) (Ping timeout: 256 seconds) |
| 2021-06-30 13:29:11 | <dminuoso> | Learning Haskell with GHC plugins right from the beginning. |
| 2021-06-30 13:29:35 | <Cale> | Not GHC plugins, but source-to-source preprocessors even :) |
| 2021-06-30 13:30:23 | <Cale> | Everything that happened with the RecordDotSyntax proposal makes me sad |
| 2021-06-30 13:30:49 | <Cale> | If a few people wanted to use this preprocessor that's one thing |
| 2021-06-30 13:31:56 | <dminuoso> | What will the "second week of Haskell" introduce? An SMT solver plugged in via TC plugin for better type level naturals. |
| 2021-06-30 13:32:06 | × | wallymathieu quits (~wallymath@81-234-151-21-no94.tbcn.telia.com) (Quit: My MacBook has gone to sleep. ZZZzzz…) |
| 2021-06-30 13:32:19 | × | maerwald quits (~maerwald@user/maerwald) (Ping timeout: 244 seconds) |
| 2021-06-30 13:32:30 | <yushyin> | I like NoFieldSelectors and maybe I will also like RecordDotSyntax, i'll give it a chance |
| 2021-06-30 13:33:46 | → | AgentM joins (~agentm@pool-162-83-130-212.nycmny.fios.verizon.net) |
| 2021-06-30 13:35:38 | → | shapr joins (~user@pool-108-28-144-11.washdc.fios.verizon.net) |
| 2021-06-30 13:35:43 | <Cale> | There were like 4 different dimensions of how to disambiguate expressions involving RecordDotSyntax, and 8 different options proposed (which didn't even fully cover the space of what would be reasonable), and then the GHC committee used Condorcet voting to pick one. |
| 2021-06-30 13:40:57 | <Cale> | (Which I think is a really poor way to decide on things like that, where it's clear that people's first impression about what expressions will mean are so different) |
| 2021-06-30 13:42:06 | × | amcleodca quits (~amcleodca@user/amcleodca) (Killed (cadmium.libera.chat (Nickname regained by services))) |
| 2021-06-30 13:42:18 | → | amcleodca joins (~andy@user/amcleodca) |
| 2021-06-30 13:43:41 | × | cfricke quits (~cfricke@user/cfricke) (Quit: WeeChat 3.2) |
| 2021-06-30 13:45:02 | <Cale> | (a copy of the document where the choices were being discussed... https://docs.google.com/document/d/1Hmviy8xJVw3RBnpXyQxb9izchXW-ZtaXMzzAX1iNuqU/edit?usp=sharing ) |
| 2021-06-30 13:46:08 | <tomsmeding> | was there a sensible subset of expressions where everyone agreed on the semantics, so that you can just reject all programs outside that intersection? |
| 2021-06-30 13:48:21 | × | Gurkenglas quits (~Gurkengla@dslb-002-203-144-156.002.203.pools.vodafone-ip.de) (Ping timeout: 258 seconds) |
| 2021-06-30 13:51:17 | <yushyin> | i think this was a very heated and difficult to follow discussion on github. I found the online document in which they tried to neatly write down the most common variants quite helpful. the discussion was already pretty deadlocked anyway. in the end, they didn't decide on the variant that I personally would have found better, but I can live with the now somewhat more conservative variant for now. |
| 2021-06-30 13:52:25 | → | Gurkenglas joins (~Gurkengla@dslb-002-203-144-156.002.203.pools.vodafone-ip.de) |
| 2021-06-30 13:52:40 | <merijn> | yushyin: NoFieldSelectors is amazing, but I don't think anyone really objects to that one :p |
| 2021-06-30 13:53:12 | <yushyin> | you could very much see wadler's law in action there :D |
| 2021-06-30 13:53:23 | <tomsmeding> | wait people disagreed on what 'f x.r' should mean? |
| 2021-06-30 13:53:31 | <yushyin> | tomsmeding: shure! |
| 2021-06-30 13:53:33 | <tomsmeding> | okay at that point there is no way you're going to satisfy everyone |
| 2021-06-30 13:53:35 | <merijn> | tomsmeding: Yes |
| 2021-06-30 13:53:43 | → | cloudy joins (~user@2600:8807:c207:f00:d022:dd81:f0f8:bb22) |
| 2021-06-30 13:54:00 | <merijn> | tbh, the real mistake in Haskell was not requiring whitespace around operators |
| 2021-06-30 13:54:03 | <tomsmeding> | well, I guess the "sensible subset" that I mentioned would then the programs that do not use record dot syntax :p |
| 2021-06-30 13:54:05 | <tomsmeding> | i.e. C1 |
| 2021-06-30 13:54:12 | <merijn> | If had required whitespace around operators none of this would be a problem |
| 2021-06-30 13:54:21 | <tomsmeding> | dminuoso: you can stop waiting |
| 2021-06-30 13:54:57 | <merijn> | I don't really see any value in operators without whitespace anyway |
| 2021-06-30 13:55:31 | <merijn> | This would've eliminated the need for unary minus too |
| 2021-06-30 13:55:39 | <tomsmeding> | it's sometimes nice to compress e.g. [0 .. i-1] |
| 2021-06-30 13:55:43 | <dminuoso> | tomsmeding: ! |
| 2021-06-30 13:55:43 | <tomsmeding> | lol |
| 2021-06-30 13:55:45 | <dminuoso> | :-) |
| 2021-06-30 13:55:56 | <merijn> | tomsmeding: you really think two spaces would ruin that? |
| 2021-06-30 13:56:25 | <merijn> | tomsmeding: Or rather, would you put up with 2 spaces there if it completely eliminated the problems of: 1) unary minus and 2) record indexing? |
| 2021-06-30 13:57:08 | <Rembane> | What about always demanding whitespace around the function composition operator, but keeping the rest as it is? |
| 2021-06-30 13:57:11 | → | trcc_ joins (~trcc@eduroam09.au.dk) |
| 2021-06-30 13:57:31 | <merijn> | Rembane: Now you're special casing a specific function name in the grammar |
| 2021-06-30 13:57:33 | <merijn> | Hard pass |
| 2021-06-30 13:57:44 | <merijn> | I want *less* special casing and weird ass rules |
| 2021-06-30 13:57:45 | <merijn> | Not more |
| 2021-06-30 13:57:53 | <merijn> | Rembane: What happens if a user defines a local (.) ? |
| 2021-06-30 13:57:59 | <Rembane> | merijn: We already have one, the dollar sign, lets give it exactly one friend. :) |
| 2021-06-30 13:58:19 | <Rembane> | merijn: Then it must behave in the exact same way as the global one. |
| 2021-06-30 13:58:37 | <merijn> | This sounds incredibly unprincipled and messy |
| 2021-06-30 13:58:44 | <tomsmeding> | merijn: yes I would put up with some spaces :p |
| 2021-06-30 13:58:50 | <merijn> | Rembane: Also, then you still have the unary minus problem |
| 2021-06-30 13:59:31 | <Rembane> | merijn: Indeed. The unary minus problem is definitely a wart. |
| 2021-06-30 14:00:53 | × | trcc quits (~trcc@users-1190.st.net.au.dk) (Ping timeout: 268 seconds) |
| 2021-06-30 14:01:18 | <merijn> | Mandatory whitespace around operator solves: 1) ambiguity of qualified import vs constructor composition with ., 2) solves the problem of finding a non-conflicting record index syntax, 3) eliminates unary minus. I, the prophet Merijn, have foretold the 2nd^H^H^H19th coming of Haskell Prime and it will redeem us all! Heed my words, oh ye unworthy! |
| 2021-06-30 14:01:58 | × | trcc_ quits (~trcc@eduroam09.au.dk) (Ping timeout: 265 seconds) |
| 2021-06-30 14:02:42 | <merijn> | While we're at it and need to accomodate our new "negative literals", let's splice in a new SignedNum class and move abs, signum and - into it >.> |
| 2021-06-30 14:03:23 | → | neo joins (~neo3@cpe-292712.ip.primehome.com) |
| 2021-06-30 14:06:08 | → | delYsid joins (~user@84-115-55-45.cable.dynamic.surfer.at) |
| 2021-06-30 14:06:10 | <yushyin> | i vote merijn for haskell prime chairman! \o/ let the bikeshedding begin \o/ |
| 2021-06-30 14:07:44 | × | mastarija quits (~mastarija@31.217.22.187) (Quit: Leaving) |
| 2021-06-30 14:08:00 | → | eggplantade joins (~Eggplanta@2600:1700:bef1:5e10:dc49:b28b:5485:3e7e) |
| 2021-06-30 14:08:24 | <turlando> | Do we really need negative numbers after all? |
| 2021-06-30 14:08:42 | <merijn> | yushyin: https://files.inconsistent.nl/merijn.jpg |
| 2021-06-30 14:09:11 | <yushyin> | ;D |
| 2021-06-30 14:10:58 | × | MQ-17J quits (~MQ-17J@d14-69-206-129.try.wideopenwest.com) (Read error: Connection reset by peer) |
| 2021-06-30 14:12:02 | <yushyin> | merijn: btw. I get a bad cert domain ssl error with your link. |
| 2021-06-30 14:12:18 | × | eggplantade quits (~Eggplanta@2600:1700:bef1:5e10:dc49:b28b:5485:3e7e) (Ping timeout: 240 seconds) |
| 2021-06-30 14:12:24 | → | fendor__ joins (~fendor@77.119.196.177.wireless.dyn.drei.com) |
| 2021-06-30 14:13:28 | → | chris_ joins (~chris@81.96.113.213) |
| 2021-06-30 14:13:35 | × | sander quits (~sander@user/sander) (Quit: So long! :)) |
| 2021-06-30 14:13:49 | → | maerwald joins (~maerwald@mail.hasufell.de) |
| 2021-06-30 14:13:51 | → | Sgeo joins (~Sgeo@user/sgeo) |
| 2021-06-30 14:14:35 | → | sander joins (~sander@user/sander) |
| 2021-06-30 14:14:42 | <merijn> | oh, heh |
| 2021-06-30 14:14:47 | <merijn> | it shouldn't be https |
| 2021-06-30 14:14:54 | × | fendor quits (~fendor@77.119.195.142.wireless.dyn.drei.com) (Ping timeout: 252 seconds) |
| 2021-06-30 14:14:56 | <merijn> | I blame my browser |
| 2021-06-30 14:16:16 | <merijn> | Because that server indeed doesn't have any certificate for my domain, on account of it not being my server :p |
| 2021-06-30 14:16:42 | → | MQ-17J joins (~MQ-17J@8.21.10.15) |
| 2021-06-30 14:17:23 | → | Pickchea joins (~private@user/pickchea) |
| 2021-06-30 14:18:52 | × | neo quits (~neo3@cpe-292712.ip.primehome.com) (Quit: Leaving) |
| 2021-06-30 14:19:11 | → | neo joins (~neo3@cpe-292712.ip.primehome.com) |
| 2021-06-30 14:20:13 | → | robbert-vdh joins (~robbert@robbertvanderhelm.nl) |
| 2021-06-30 14:23:43 | <guest61> | . |
| 2021-06-30 14:24:04 | × | jao quits (jao@gateway/vpn/protonvpn/jao) (Ping timeout: 272 seconds) |
| 2021-06-30 14:24:04 | Fendor_ | is now known as fendor |
| 2021-06-30 14:27:25 | → | pagnol joins (~user@014198154145.ctinets.com) |
| 2021-06-30 14:33:12 | <robbert-vdh> | Has anyone ever tried using RTS Execution Stacks? (after compiling GHC with --enable-dwarf-unwind) Is that supposed to work? |
| 2021-06-30 14:34:20 | <merijn> | robbert-vdh: *someone*? probably, whether that person is here, who knows :p |
| 2021-06-30 14:34:44 | <robbert-vdh> | merijn: At this point I really wouldn't be surprised if literally noone has tried using this feature lol |
| 2021-06-30 14:34:45 | <merijn> | I'd be curious to see how well it works if you succeed, though |
| 2021-06-30 14:36:01 | <robbert-vdh> | I tried compiling both GHC 8.10.5 and GHC 9.2.0.whatevertheprereleaseis with that option and producing a stack trace gives me "execution-stacks: Failed to get stack frames of current process: no matching address range: Invalid argument", so I was hoping there's some obvious flag I'm missing or something :) |
| 2021-06-30 14:36:13 | <robbert-vdh> | (and the stack trace ends in rts/Libdw.c which uh isn't very helpful) |
| 2021-06-30 14:36:37 | <robbert-vdh> | Trying to use this to get some additional source mapping information in Accelerate |
| 2021-06-30 14:38:35 | <merijn> | robbert-vdh: Probably wanna ask in #ghc |
| 2021-06-30 14:38:53 | <merijn> | robbert-vdh: Since this dealing with rather deep internsl of ghc |
| 2021-06-30 14:39:18 | × | jumper149 quits (~jumper149@80.240.31.34) (Quit: WeeChat 3.1) |
| 2021-06-30 14:39:29 | <robbert-vdh> | merijn: Yeah I guess I'll have to do that, was hoping someone else miraculously also tried using this feature and knew the magic incantation to make it work :) |
All times are in UTC.