Newest at the top
| 2024-11-06 02:43:07 +0100 | <haskellbridge> | <sm> a language server is a small internal component that lets IDEs communicate with the language tools |
| 2024-11-06 02:43:04 +0100 | <monochrom> | I talked about lacking a soul, but when you do have a soul, it needs a body too... |
| 2024-11-06 02:42:58 +0100 | <haskellbridge> | <zwro> I don't think people like changing editors just for the language |
| 2024-11-06 02:42:31 +0100 | <monochrom> | A face? :) |
| 2024-11-06 02:42:18 +0100 | <haskellbridge> | <zwro> what can an IDE provide that a language server can't? |
| 2024-11-06 02:41:54 +0100 | <haskellbridge> | <sm> I did see a more recent brick based debugger ui. I'd find it, if hackage wasn't down just now |
| 2024-11-06 02:41:26 +0100 | <monochrom> | The cartoon? Yes I always link to it to rub it into people's faces. |
| 2024-11-06 02:41:14 +0100 | <haskellbridge> | <sm> monochrom: I think those things prove that packaging and integration are also necessary.. so all of those are dead.. except the debugger built in to GHCI, which works perfectly (or anyway, works) |
| 2024-11-06 02:39:52 +0100 | <haskellbridge> | <sm> wait, didn't I see that already... |
| 2024-11-06 02:39:41 +0100 | <monochrom> | Oh it's an extreme case of the 80-20 rule. Or at this extreme it's like a 99.99-0.01 rule. Hat, Buddha, GHCi, and that plug in prove that prototyping a Haskell debugger is easy, sure, and at the same time by their incompleteness and even bitrot, also prove that actually covering all of Haskell is way too annoying for anyone to finish it. |
| 2024-11-06 02:38:45 +0100 | <haskellbridge> | <sm> someone could whip that up in brick |
| 2024-11-06 02:36:19 +0100 | <haskellbridge> | <sm> It doesn't seem _that_ hard. GHCI already has the core of it, and https://marketplace.visualstudio.com/items?itemName=phoityne.phoityne-vscode is an example of a UI on top |
| 2024-11-06 02:36:00 +0100 | <monochrom> | OTOH, a nice IDE that lacks debugging feels like lacking a soul. |
| 2024-11-06 02:35:30 +0100 | <monochrom> | Debugger is extremely hard. I don't really complain about lacking it. |
| 2024-11-06 02:34:51 +0100 | <haskellbridge> | <sm> * UI though, that's what I'm thinking of particularly |
| 2024-11-06 02:34:26 +0100 | <haskellbridge> | <sm> Not a debugger thought |
| 2024-11-06 02:34:23 +0100 | <haskellbridge> | <sm> Ok. That IDE was good as I recall ! Lots of people had a beef with FP Complete so it had an uphill path. |
| 2024-11-06 02:33:11 +0100 | <monochrom> | IIRC the company is precisely FPComplete. |
| 2024-11-06 02:33:05 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
| 2024-11-06 02:32:19 +0100 | <haskellbridge> | <sm> I think in this cartoon, something is missing from the company's offering. We just don't know what it is :) |
| 2024-11-06 02:30:36 +0100 | <monochrom> | https://ro-che.info/ccc/26 |
| 2024-11-06 02:29:18 +0100 | <haskellbridge> | <sm> a built in cheaper just works ui could help build the foundation |
| 2024-11-06 02:28:12 +0100 | <haskellbridge> | <sm> I have got the VS Code debugger extension working once, and it was so fragile and little known that pretty much no-one else will |
| 2024-11-06 02:26:47 +0100 | xff0x | (~xff0x@fsb6a9491c.tkyc517.ap.nuro.jp) |
| 2024-11-06 02:26:43 +0100 | <haskellbridge> | <sm> I promise you this one would |
| 2024-11-06 02:26:29 +0100 | CoolMa7 | (~CoolMa7@ip5f5b8957.dynamic.kabel-deutschland.de) (Ping timeout: 252 seconds) |
| 2024-11-06 02:26:26 +0100 | <haskellbridge> | <sm> and sometimes a monolithic just works out of box tool is the only thing that'll get used |
| 2024-11-06 02:26:12 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
| 2024-11-06 02:25:51 +0100 | <monochrom> | In practice, that doesn't happen often enough in the Haskell community. Look at how GHCi debugger has not received any further work, and how HLS still doesn't have any debugger. Even the GHCup effort is wearing thin and the guy is complaining: https://discourse.haskell.org/t/priorities-for-upcoming-ghc-releases/9605/60 |
| 2024-11-06 02:25:50 +0100 | postcerebral | (~postcereb@user/postcerebral) (Quit: WeeChat 4.4.2) |
| 2024-11-06 02:25:15 +0100 | <haskellbridge> | <sm> or anyway, _a_ way |
| 2024-11-06 02:24:54 +0100 | <haskellbridge> | <sm> I could imagine integrating an already-maintained TUI editor (yi ?) could also be the way to improve GHCI's debugger, because a more visual UI is what it needs |
| 2024-11-06 02:24:50 +0100 | <zero> | which wraps it "nicely" |
| 2024-11-06 02:23:51 +0100 | <zero> | foreign-store is hacky. i would be more interested in something like https://hackage.haskell.org/package/rapid-0.1.4/docs/Rapid.html |
| 2024-11-06 02:22:07 +0100 | CoolMa7 | (~CoolMa7@ip5f5b8957.dynamic.kabel-deutschland.de) CoolMa7 |
| 2024-11-06 02:21:59 +0100 | Tuplanolla | (~Tuplanoll@91-159-69-59.elisa-laajakaista.fi) (Quit: Leaving.) |
| 2024-11-06 02:19:53 +0100 | <haskellbridge> | <sm> just saying |
| 2024-11-06 02:19:35 +0100 | <haskellbridge> | <sm> dev time doesn't always have to be zero sum in FOSS.. motivated new devs / funders sometimes show up |
| 2024-11-06 02:18:36 +0100 | <monochrom> | In the particular case of GHCi, given that dev time is a zero-sum game, when (not if, it's a certainty) the devs have to choose between improving GHCi's debugger and adding an editor to GHCi, why would anyone want the latter? Isn't debugger a much higher priority? Editors already exist, plentiifully. |
| 2024-11-06 02:18:35 +0100 | <constxd> | but yeah i wouldn't want to implement that and maintain it as a part of a REPL |
| 2024-11-06 02:18:15 +0100 | <constxd> | wouldn't be the end of the world imo if it's "for free" by virtue of being built into libreadline or something |
| 2024-11-06 02:16:53 +0100 | <haskellbridge> | <sm> +how it goes |
| 2024-11-06 02:16:33 +0100 | <haskellbridge> | <sm> try it out, let us know ? |
| 2024-11-06 02:15:56 +0100 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 272 seconds) |
| 2024-11-06 02:15:41 +0100 | <zero> | ooh is this hot reloading? |
| 2024-11-06 02:15:26 +0100 | <monochrom> | Instead, since iPython and iHaskell exist, there is no point duplicating the effort in the basic REPL. |
| 2024-11-06 02:14:37 +0100 | <constxd> | ah |
| 2024-11-06 02:14:29 +0100 | <monochrom> | This is different from iPython and iHaskell. I have no quarrel with them. |
| 2024-11-06 02:13:29 +0100 | <monochrom> | tl;dr the python REPL is now its own editor. |
| 2024-11-06 02:12:57 +0100 | <monochrom> | constxd: This: https://discourse.haskell.org/t/multiline-interpreter-with-editing-of-previous-lines/10551 |