Newest at the top
2024-10-13 21:14:48 +0200 | machinedgod | (~machinedg@d50-99-47-73.abhsia.telus.net) (Ping timeout: 252 seconds) |
2024-10-13 21:14:09 +0200 | rvalue | (~rvalue@user/rvalue) rvalue |
2024-10-13 21:13:36 +0200 | rvalue | (~rvalue@user/rvalue) (Read error: Connection reset by peer) |
2024-10-13 21:08:33 +0200 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 252 seconds) |
2024-10-13 21:08:03 +0200 | <monochrom> | Is it like you have your own shell script and you name it "gcc" to intercept usages of gcc and change some arguments and you would really like to avoid parsing a response file? :) |
2024-10-13 21:06:42 +0200 | <monochrom> | _d0t: Now I'm curious. (Just curious.) What needs to be hacked around? :) |
2024-10-13 21:06:18 +0200 | acidjnk_new | (~acidjnk@p200300d6e72cfb115c8ca2f660dd1d29.dip0.t-ipconnect.de) |
2024-10-13 21:05:56 +0200 | Pixi | (~Pixi@user/pixi) Pixi |
2024-10-13 21:03:56 +0200 | <Square> | dminuoso, "Q: how many types". Unclear. But possibly 50ish. |
2024-10-13 21:03:32 +0200 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) merijn |
2024-10-13 21:03:00 +0200 | <Square> | The generation of schema + java types would be part of the build, so any changes in types would lead to compile errors - if not adapted in java. |
2024-10-13 21:02:29 +0200 | vanishingideal | (~vanishing@user/vanishingideal) vanishingideal |
2024-10-13 21:01:53 +0200 | <Square> | parts of a java application* |
2024-10-13 21:01:38 +0200 | <Square> | My use case is migrating parts of an application to Haskell and having an RPC mechanism calling Haskell from Java. Ideally i would like to replicate a bunch of haskell types in java (limiting myself to constructs it supports). |
2024-10-13 21:00:42 +0200 | caconym | (~caconym@user/caconym) caconym |
2024-10-13 21:00:17 +0200 | <dminuoso> | Also think about code generators (the other way around), they tend to be esaier. |
2024-10-13 21:00:06 +0200 | caconym | (~caconym@user/caconym) (Quit: bye) |
2024-10-13 21:00:02 +0200 | <dminuoso> | How many data types are we talking about here? |
2024-10-13 20:59:31 +0200 | <dminuoso> | It increases coupling, introduces bugs quickly when you make modifications, and delays separating the network representation properly, possibly with versioning even. |
2024-10-13 20:58:25 +0200 | <Square> | s/in don't/in I don't/ |
2024-10-13 20:58:07 +0200 | <Square> | dminuoso, I know an argument, it could make my work so much easier in don't having to replicating the model by hand in another language. |
2024-10-13 20:57:53 +0200 | lxsameer | (~lxsameer@Serene/lxsameer) (Ping timeout: 245 seconds) |
2024-10-13 20:57:13 +0200 | <_d0t> | Otherwise I'll just have to hack around this. |
2024-10-13 20:57:09 +0200 | <dminuoso> | Certainly promoted by the ease of `Generic` and` ToJSON/FromJSON. |
2024-10-13 20:57:07 +0200 | <_d0t> | geekosaur: i mean, i was just wondering if there was a way already. |
2024-10-13 20:56:58 +0200 | <dminuoso> | Tying your wire format to your semantic data model is a classic mistake. |
2024-10-13 20:56:43 +0200 | <dminuoso> | In fact, you should probably have an intermediate data type representing the format instead. |
2024-10-13 20:56:36 +0200 | <geekosaur> | you''d have to ask in #ghc but I strongly suspect they're not interested in having to maintain multiple ways to invoke gcc |
2024-10-13 20:56:29 +0200 | <dminuoso> | Square: As they shouldn't. |
2024-10-13 20:56:13 +0200 | <Square> | dminuoso, Maybe I should. I just feel most of these formats are crude and doesn't lend themselves in replicating a data model in another language. |
2024-10-13 20:56:09 +0200 | <_d0t> | yeah, i meant response file https://gcc.gnu.org/wiki/Response_Files |
2024-10-13 20:55:47 +0200 | <geekosaur> | the problem is that Windows command lines are way too short |
2024-10-13 20:55:41 +0200 | <_d0t> | geekosaur: ok but what if I don't care about windows? |
2024-10-13 20:55:36 +0200 | <geekosaur> | I think they mean response file |
2024-10-13 20:55:28 +0200 | <geekosaur> | _d0t, generally not because of Windows compatibility |
2024-10-13 20:55:28 +0200 | <monochrom> | resource file? Is that Windows? |
2024-10-13 20:54:39 +0200 | Pixi | (~Pixi@user/pixi) (Quit: Leaving) |
2024-10-13 20:54:36 +0200 | Sgeo | (~Sgeo@user/sgeo) Sgeo |
2024-10-13 20:54:34 +0200 | <dminuoso> | Dont think of a 1:1 mapping of haskell, and just think about how you encode the idea in a format, and then keep a schema for that. |
2024-10-13 20:54:29 +0200 | CrunchyFlakes | (~CrunchyFl@ip1f13e94e.dynamic.kabel-deutschland.de) (Quit: ZNC 1.8.2 - https://znc.in) |
2024-10-13 20:54:19 +0200 | <_d0t> | ohai! Is there a way to force GHC to pass command-line args to GCC directly instead of through a resource file (ghc_X.rsp)? |
2024-10-13 20:53:40 +0200 | <Square> | dminuoso, I feel it does have something todo with schema generation though. |
2024-10-13 20:53:38 +0200 | <dminuoso> | Depending on the surrounding code, you could also work with a GADT |
2024-10-13 20:52:29 +0200 | <dminuoso> | What you are asking about has nothing to do with the serialization format |
2024-10-13 20:52:10 +0200 | <dminuoso> | Add a (Typeable a, Typeable b) constraint to your serialization function, and work with that? |
2024-10-13 20:51:57 +0200 | merijn | (~merijn@128-137-045-062.dynamic.caiway.nl) (Ping timeout: 246 seconds) |
2024-10-13 20:51:24 +0200 | <Square> | data MyType a b = MyType a b |
2024-10-13 20:50:44 +0200 | <dminuoso> | What do you mean by type parameters? |
2024-10-13 20:50:24 +0200 | <Square> | dminuoso, Maybe I missed it, but I thought it didn't have support for type parameters. |
2024-10-13 20:47:44 +0200 | <monochrom> | Yeah, binary and cereal should also do those things and pretty much automatable. |