Home liberachat/#haskell: Logs Calendar

Logs: liberachat/#haskell

←Prev  Next→ 1,804,084 events total
2025-09-07 17:05:37 <slondr> is there a way to tell hls-wrapper to use a specific hls version?
2025-09-07 17:05:53 Raito_Bezarius joins (~Raito@libera/contributor/wireguard.tunneler.raito-bezarius)
2025-09-07 17:06:05 tv joins (~tv@user/tv)
2025-09-07 17:09:15 × merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 248 seconds)
2025-09-07 17:11:08 divya joins (divya@140.238.251.170)
2025-09-07 17:12:02 zzz joins (~zero@user/zero)
2025-09-07 17:13:10 <slondr> actually, what version of hls should I even be using? I tried reinstalling 2.10 using ghcup but I still ended up with it compiled using ghc 9.10 even though my ghc is 9.6
2025-09-07 17:13:44 × Lord_of_Life quits (~Lord@user/lord-of-life/x-2819915) (Quit: Laa shay'a waqi'un moutlaq bale kouloun moumkine)
2025-09-07 17:15:25 <geekosaur> that's not precise enough: minor versions matter. 2.10 supports 9.6.1-6 but not 9.6.7 because the ghc-api changed (as yet, it's not stable even across major versions)
2025-09-07 17:16:39 <slondr> yeah I'm just getting "ghcide compiled against GHC 9.10.1 but currently using 9.6.7 This is unsupported, ghcide must be compiled with the same GHC version as the project." when I try to run it
2025-09-07 17:16:58 <geekosaur> also, hls doesn't pick a ghc, it tries to use whichever one your project uses by figuring out if you're using cabal or stack and asking that, otherwise using whatever it finds first in $PATH
2025-09-07 17:17:43 <slondr> So maybe my problem is that GHCup compiled hls using a ghc version that isn't the version on my path/set in GHCup?
2025-09-07 17:18:17 <geekosaur> that sounds like haskell-language-server-wrapper is picking the wrong HLS (ghc-ide is a component of HLS)
2025-09-07 17:18:40 <slondr> I agree; is there a way to change which HLS is picked by HLS-wrapper?
2025-09-07 17:20:23 merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl)
2025-09-07 17:20:38 × MelodyOwO quits (~MelodyOwO@user/MelodyOwO) (Quit: Leaving.)
2025-09-07 17:21:23 <geekosaur> the right way to do it is to switch to HLS 2.11, which supports ghc 9.6.7. but I thought HLS-=wrapper just faield to start if it didn't find a matching binary, here it sounds like it's just trying the newest HLS binary present which is just wrong
2025-09-07 17:23:41 <slondr> I was just on all the "recommended" versions of everything in GHCup, if I switch to "latest" everything should I expect that to work?
2025-09-07 17:24:36 <geekosaur> wait, I keep forgetting 2.10 does support 9.6.7
2025-09-07 17:24:46 <geekosaur> so I have no idea why the wrapper isn't selecting it
2025-09-07 17:25:45 <geekosaur> do you have ~/.ghcup/bin/haskell-language-server-9.6.7-2.10.0.0?
2025-09-07 17:26:54 <slondr> yes (it's a ~ not a - but otherwise yes)
2025-09-07 17:27:14 × merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 256 seconds)
2025-09-07 17:30:50 <geekosaur> weird
2025-09-07 17:30:53 <geekosaur> (not the ~)
2025-09-07 17:31:08 <geekosaur> I think at this point you need to ask over in #haskell-language-server
2025-09-07 17:32:04 Lord_of_Life joins (~Lord@user/lord-of-life/x-2819915)
2025-09-07 17:33:55 <tomsmeding> slondr: have you tried running 'haskell-language-server-wrapper' in a terminal and seeing what it prints?
2025-09-07 17:33:59 merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl)
2025-09-07 17:34:34 <tomsmeding> Do you perhaps have a stack.yaml in the directory too but want to use cabal, and the stack.yaml specifies a different GHC version (via the snapshot) than your cabal project?
2025-09-07 17:34:47 <slondr> I'm just in a directory with a single .hs file nothing else
2025-09-07 17:35:12 <tomsmeding> ah
2025-09-07 17:35:30 <tomsmeding> then the first thing I said
2025-09-07 17:35:30 × tv quits (~tv@user/tv) (Read error: Connection reset by peer)
2025-09-07 17:35:45 <tomsmeding> if you're not sure whether that outputs something relevant, paste the output in a pastebin (see /topic)
2025-09-07 17:35:47 <slondr> Running h-l-s-wrapper doesn't report any obvious errors, it prints a bunch of stuff but looks fine?
2025-09-07 17:36:06 <tomsmeding> what ghc version does it select? And is that the same version as you get reported in your editor?
2025-09-07 17:36:17 <geekosaur> ctly is it printing? (use paste.tomsmeding.com)
2025-09-07 17:36:35 tomsmeding . o O ( "ctly"? )
2025-09-07 17:37:14 <int-e> I'd guess "what exactly"
2025-09-07 17:37:29 <geekosaur> whoops, yes
2025-09-07 17:37:34 <int-e> no clue how it got cut off like that though
2025-09-07 17:37:36 <geekosaur> I keep fatfingering ctrl and a
2025-09-07 17:37:51 <geekosaur> (I got up unusually early this morning and it shows)
2025-09-07 17:37:59 <tomsmeding> you have ctrl on capslock?
2025-09-07 17:38:08 <geekosaur> yeh, because I'm ancient
2025-09-07 17:38:22 <slondr> https://paste.tomsmeding.com/IZI6qrQg
2025-09-07 17:38:26 <tomsmeding> no need to use that excuse, I know very young people who do that too :p (I don't though)
2025-09-07 17:38:44 <slondr> I have Compose on caps lock :)
2025-09-07 17:39:07 × merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 265 seconds)
2025-09-07 17:39:11 <tomsmeding> slondr: that... looks very innocuous
2025-09-07 17:39:20 <tomsmeding> it just selects HLS for GHC 9.6.7 and it works
2025-09-07 17:39:40 <geekosaur> right, that works right apparently
2025-09-07 17:39:57 <geekosaur> hm, do you have a stray ghc-ide somewhere on your PATH?
2025-09-07 17:40:18 <tomsmeding> slondr: how have you set up HLS in your editor? Have you, perchance, told it to start h-l-s-9.10.1 specifically? Or that ^ I guess
2025-09-07 17:40:26 <geekosaur> er, ghcide
2025-09-07 17:42:05 <tomsmeding> also, what editor are you using, and does that editor have the same $PATH as in a terminal?
2025-09-07 17:42:11 <slondr> All I did was install lsp-haskell and run lsp in a .hs file. I haven't configured anything else in particular
2025-09-07 17:42:35 <tomsmeding> your editor does not magically know about "haskell-language-server"
2025-09-07 17:42:53 <tomsmeding> you may have a plugin in your editor that knows about it and has lain dormant until now
2025-09-07 17:43:09 <slondr> oh to be clear lsp-haskell is a package for my editor
2025-09-07 17:43:16 <slondr> https://github.com/emacs-lsp/lsp-haskell
2025-09-07 17:43:35 <tomsmeding> looks like that starts h-l-s-wrapper
2025-09-07 17:43:50 <tomsmeding> is the lsp-haskell-server-path still on the default value?
2025-09-07 17:44:18 <slondr> yep it's "haskell-language-server-wrapper"
2025-09-07 17:44:34 <geekosaur> is there some reasom you're not using the built-in eglot?
2025-09-07 17:44:55 tomsmeding will defer to people who actually know something about emacs
2025-09-07 17:45:21 × sajenim quits (~sajenim@user/sajenim) (Ping timeout: 256 seconds)
2025-09-07 17:45:35 <geekosaur> I only know a little bit more, but eglot's supposed to be pretty good these days (originally it wasn't and people used other solutions)
2025-09-07 17:45:44 <slondr> lsp::log shows that it's starting haskell-language-server-wrapper with log output to ../hls.log
2025-09-07 17:45:56 <geekosaur> (I switched to codium)
2025-09-07 17:46:25 <slondr> geekosaur: I prefer lsp's output and interactions to eglot generally
2025-09-07 17:47:09 <tomsmeding> slondr: cool, is there anything relevant in that hls.log?
2025-09-07 17:50:27 × peterbecich quits (~Thunderbi@syn-172-222-149-049.res.spectrum.com) (Ping timeout: 252 seconds)
2025-09-07 17:50:39 <slondr> https://paste.tomsmeding.com/S5NDK1ZO
2025-09-07 17:50:56 <slondr> I don't think so? nothing jumps out as an obvious error to mee at least. I'm not even seeing the error reported in my haskell buffer here
2025-09-07 17:53:28 tv joins (~tv@user/tv)
2025-09-07 17:54:05 <slondr> ok this is really weird actually. some lsp features are totally working! when I mouse over functions I see their docs in a pop-up, when I mouse over definitions it shows me their inferred types. all that is working. it's just also randomly giving me this critical error on line 1 character 1 with a nonsense error message
2025-09-07 17:54:31 × humasect quits (~humasect@dyn-192-249-132-90.nexicom.net) (Quit: Leaving...)
2025-09-07 17:55:56 <slondr> I do *not* get this error with eglot, which does imply that this is a problem with lsp-haskell and not with HLS itself
2025-09-07 17:56:04 <slondr> Hmmmmmmmmmmm
2025-09-07 17:58:57 × img quits (~img@user/img) (Quit: ZNC 1.8.2 - https://znc.in)
2025-09-07 17:59:15 ljdarj joins (~Thunderbi@user/ljdarj)
2025-09-07 18:00:15 img joins (~img@user/img)
2025-09-07 18:02:41 <slondr> well I filed a bug report with lsp-haskell: https://github.com/emacs-lsp/lsp-haskell/issues/201 we'll see what they say
2025-09-07 18:04:25 trickard_ is now known as trickard
2025-09-07 18:10:28 vanishingideal joins (~vanishing@user/vanishingideal)
2025-09-07 18:11:32 <haskellbridge> <geekosaur> Fwiw that looks like lsp-haskell is mis-parsing HLS output
2025-09-07 18:19:35 <[exa]> evenin' guys
2025-09-07 18:19:50 <[exa]> hows your thunks
2025-09-07 18:22:55 <Rembane> Evening! They're thunking!
2025-09-07 18:24:19 × divlamir quits (~divlamir@user/divlamir) (Read error: Connection reset by peer)
2025-09-07 18:24:44 divlamir joins (~divlamir@user/divlamir)
2025-09-07 18:29:00 × vanishingideal quits (~vanishing@user/vanishingideal) (Ping timeout: 256 seconds)
2025-09-07 18:30:17 Lycurgus joins (~juan@user/Lycurgus)
2025-09-07 18:30:31 vanishingideal joins (~vanishing@user/vanishingideal)
2025-09-07 18:33:32 × tomboy64 quits (~tomboy64@user/tomboy64) (Ping timeout: 256 seconds)
2025-09-07 18:34:59 merijn joins (~merijn@host-vr.cgnat-g.v4.dfn.nl)
2025-09-07 18:35:53 tomboy64 joins (~tomboy64@user/tomboy64)
2025-09-07 18:39:46 × merijn quits (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 256 seconds)

All times are in UTC.