2025/05/02

Newest at the top

2025-05-02 04:08:16 +0200 <geekosaur> highly unlikely. most probable to me is that it's simply ignored when parsing a cabal file, but it's in LocalBuildInformation so when the cabal.project parser sees it it can add it there
2025-05-02 04:07:37 +0200Typedfern(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net) (Ping timeout: 276 seconds)
2025-05-02 04:07:21 +0200typedfern_(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net) typedfern
2025-05-02 04:06:55 +0200 <sim590> Or could it be that c2hs actually reads the `cabal.project` file and parses the options himself? That would not have been cool, but it's just a thought.
2025-05-02 04:06:13 +0200 <sim590> So, in the whole codebase, the string c2hs-optins only appears in test files.
2025-05-02 04:02:18 +0200 <geekosaur> er, LocalBuildInformation
2025-05-02 04:01:47 +0200 <geekosaur> (sorry, that might not be clear: LocalBuildInfo. the central data structures are abbreviated all over the codebase)
2025-05-02 04:00:28 +0200 <geekosaur> makes me wonder how many other fields might be missing
2025-05-02 03:59:43 +0200 <geekosaur> did someone put it in the LBI but not then add a field for parsing?
2025-05-02 03:59:41 +0200 <sim590> Yeah I agree.
2025-05-02 03:58:57 +0200 <geekosaur> it explains the parsing, but not why it works at all
2025-05-02 03:58:52 +0200 <sim590> So, actually, the only references I saw seems to be in test files. And they are related to the $HOME/.cabal/config file interpreter I think.
2025-05-02 03:58:42 +0200stef204(~stef204@user/stef204) stef204
2025-05-02 03:57:52 +0200 <sim590> hmmmm. So, there are references in the code about `c2hs-options`, but they're not in the FieldGrammar module under the PackageDescription dir. I think that explains the difference.
2025-05-02 03:57:12 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 252 seconds)
2025-05-02 03:52:17 +0200 <geekosaur> what's puzzling me is, if it's not supported there, it shouldn't work in `cabal.project` either
2025-05-02 03:51:49 +0200 <sim590> Thanks again for the support
2025-05-02 03:50:38 +0200volsand(~volsand@2804:1b1:1080:da6:192e:6849:5950:90dd) (Quit: volsand)
2025-05-02 03:50:29 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn
2025-05-02 03:49:04 +0200 <geekosaur> 👍
2025-05-02 03:47:47 +0200sajenim(~sajenim@user/sajenim) sajenim
2025-05-02 03:46:45 +0200j1n37(~j1n37@user/j1n37) j1n37
2025-05-02 03:46:12 +0200j1n37-(~j1n37@user/j1n37) (Ping timeout: 252 seconds)
2025-05-02 03:43:34 +0200 <sim590> It might be an easy PR for me to make. I could basically mimic pretty much everything that's done for hsc2hs-options, but do it for c2hs-options.
2025-05-02 03:40:40 +0200 <sim590> But c2hs is not there.
2025-05-02 03:40:33 +0200 <sim590> I don't know if I'm at the right spot, but I see hsc2hs-options seems to be be part of the list of recognized fields here: https://github.com/haskell/cabal/blob/7e0f040d1147770b57e76b2e51123f01e83ce263/Cabal-syntax/src/Di…
2025-05-02 03:40:19 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 276 seconds)
2025-05-02 03:37:47 +0200cyphase(~cyphase@user/cyphase) cyphase
2025-05-02 03:35:17 +0200bitdex(~bitdex@gateway/tor-sasl/bitdex) bitdex
2025-05-02 03:34:42 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn
2025-05-02 03:26:28 +0200Typedfern(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net) typedfern
2025-05-02 03:25:47 +0200 <geekosaur> I told you, it's a development thing. You're trying to use it for something else
2025-05-02 03:25:20 +0200 <sim590> Hmmmm. OK. I don't really know why, but I guess I need prerequisites to undestand it. I don't really know how the build system does its thing, so I don't see the logic here, but OK. So I guess, my only chance now, is to have a fix from the cabal team so that c2hs-options be added to a cabal release.
2025-05-02 03:24:09 +0200Typedfern(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net) (Remote host closed the connection)
2025-05-02 03:23:44 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds)
2025-05-02 03:21:28 +0200typedfern_(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net) (Ping timeout: 276 seconds)
2025-05-02 03:18:56 +0200merijn(~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn
2025-05-02 03:18:26 +0200Typedfern(~Typedfern@135.red-83-37-43.dynamicip.rima-tde.net)
2025-05-02 03:17:30 +0200 <geekosaur> but you need to include a `packages:` entry, which would override the main one and result it the package being installed in its own project's `dist-newstyle` where it won't be found by the main project
2025-05-02 03:16:12 +0200 <sim590> So for your question, I guess it both files could be nested. The user's file and the dependency's file?
2025-05-02 03:15:47 +0200 <geekosaur> (unfortunately most of the dev team is asleep at this time of their night)
2025-05-02 03:15:39 +0200 <sim590> OK, thanks for the initiative.
2025-05-02 03:15:09 +0200 <geekosaur> fwiw I just asked about `c2hs-options` not working in the cabal dev channel; it seems very wrong that it works in a project file but not the cabal file
2025-05-02 03:14:27 +0200 <geekosaur> someone's using your package as a dependency in a development tree with a `cabal.project`. what happens?
2025-05-02 03:13:53 +0200 <sim590> I don't understand the question.
2025-05-02 03:13:33 +0200 <geekosaur> what's supposed to happen if it's already using `cabal.project`?
2025-05-02 03:13:12 +0200 <geekosaur> 100% sure
2025-05-02 03:12:02 +0200xff0x(~xff0x@2409:251:9040:2c00:1c47:7f78:37e8:e2a1) (Ping timeout: 265 seconds)
2025-05-02 03:11:55 +0200 <sim590> hmmmm. I tried many things with the cabal file but it didn't work. :/ Are you 100% sure cabal.project won't be used when compiling as transitive dependency? If the file is there, why would it work differently when it's build as a transitive dependency?
2025-05-02 03:09:32 +0200cyphase(~cyphase@user/cyphase) (Ping timeout: 252 seconds)