Newest at the top
| 2026-02-04 16:10:50 +0100 | <mesaoptimizer> | also any package that involves networking seems to throw an error |
| 2026-02-04 16:10:35 +0100 | <mesaoptimizer> | any package that depends on ghc-prim cannot be installed via mcabal |
| 2026-02-04 16:10:23 +0100 | <mesaoptimizer> | IIRC it couldn't build aeson for example, because that relied on ghc-prim |
| 2026-02-04 16:09:57 +0100 | <mesaoptimizer> | microhs + mcabal cannot yet install / build most hackage packages |
| 2026-02-04 16:09:31 +0100 | trickard_ | (~trickard@cpe-61-98-47-163.wireline.com.au) |
| 2026-02-04 16:09:17 +0100 | trickard | (~trickard@cpe-61-98-47-163.wireline.com.au) (Read error: Connection reset by peer) |
| 2026-02-04 16:06:59 +0100 | noctux | (~noctux@user/noctux) noctux |
| 2026-02-04 16:04:28 +0100 | <gentauro> | Morj: Got it |
| 2026-02-04 16:01:55 +0100 | <haskellbridge> | <Morj> I wanted to suggest memorizing that ($) is zero and using that as a mnemonic =) |
| 2026-02-04 16:01:30 +0100 | <dutchie> | i always have to :i $ in ghci to see whether low number is tight binding or not |
| 2026-02-04 16:00:40 +0100 | <dutchie> | why can I never remember which way the numbers go in fixity declarations |
| 2026-02-04 15:59:58 +0100 | <haskellbridge> | <Morj> I could create an array of size 33 padded with zeroes, true, but it's even less convenient to use - there is no simple conversion from size 33 to size 32 |
| 2026-02-04 15:57:19 +0100 | <haskellbridge> | <Morj> If it were just Box<[u8]>, this could be achieved very simply. If it were [u8; 32] to append a byte and get [u8; 33] - also very simple. But I wanted to write one function for several N sizes |
| 2026-02-04 15:56:33 +0100 | <haskellbridge> | <Morj> I wanted to write this function because I have arrays of different sizes (for XOF seeding), and the algorithm wants to append one or two bytes to the seeds sometimes |
| 2026-02-04 15:55:57 +0100 | <haskellbridge> | <Morj> In rust you can have a choice between statically and dynamically known sizes, roughly [u8; N] and Box<[u8]> |
| 2026-02-04 15:55:35 +0100 | <gentauro> | And you end up padding/populating non-used array items with some value. |
| 2026-02-04 15:54:54 +0100 | <gentauro> | Morj: I noticed myself that Rust needs to know (statically) the size of arrays. Don't you end up by using a fixed buffer size and then just iterate? |
| 2026-02-04 15:49:05 +0100 | <akegalj> | magic_rb: found this https://mail.haskell.org/pipermail/beginners/2010-October/005571.html |
| 2026-02-04 15:49:04 +0100 | noctux | (~noctux@user/noctux) (Ping timeout: 246 seconds) |
| 2026-02-04 15:39:55 +0100 | <merijn> | I mean, that's kinda easy, just turn into ByteString and use attoparsec and/or binary :p |
| 2026-02-04 15:38:32 +0100 | <haskellbridge> | <magic_rb> i wrote a fat32 filesystem driver in haskell for my thesis, it was an experience |
| 2026-02-04 15:38:20 +0100 | <haskellbridge> | <magic_rb> my condolences |
| 2026-02-04 15:33:43 +0100 | <haskellbridge> | <Morj> *with in haskell |
| 2026-02-04 15:33:26 +0100 | <haskellbridge> | <Morj> «I have a [Word8] that I would like to convert to a Word32 in big endian. How can I do that?» - type shit that I have to deal with haskell |
| 2026-02-04 15:33:18 +0100 | <haskellbridge> | <Morj> I think it's gc. I was rather talking about not having a big runtime, the environment is rather memory constrained |
| 2026-02-04 15:32:33 +0100 | <haskellbridge> | <magic_rb> or does it do refcounting |
| 2026-02-04 15:32:26 +0100 | <haskellbridge> | <magic_rb> microhs still does gc no? |
| 2026-02-04 15:32:17 +0100 | <haskellbridge> | <Morj> Although, performance.. |
| 2026-02-04 15:31:27 +0100 | <haskellbridge> | <Morj> Hm, actually microhs might fit there, need to evaluate |
| 2026-02-04 15:31:12 +0100 | <haskellbridge> | <magic_rb> :( |
| 2026-02-04 15:31:06 +0100 | <haskellbridge> | <Morj> «just, give me haskell thank you very much» yeh. Except I currently have to write a thing without gc or a large runtime at all |
| 2026-02-04 15:30:41 +0100 | <haskellbridge> | <magic_rb> honestly the only reason i may personally choose rust over haskell is due to the library ecosystem. for embedded i would try microhs, though idk how well that would work |
| 2026-02-04 15:30:29 +0100 | <haskellbridge> | <Morj> If this were in rust, some problems of async runtimes would be removed |
| 2026-02-04 15:30:22 +0100 | Enrico63 | (~Enrico63@148.252.128.12) Enrico63 |
| 2026-02-04 15:30:11 +0100 | <haskellbridge> | <Morj> I'm partial to the RIO approach of having only one monad, but you parameterize the context you read from it |
| 2026-02-04 15:29:57 +0100 | <haskellbridge> | <magic_rb> just, give me haskell thank you very much |
| 2026-02-04 15:29:50 +0100 | <haskellbridge> | <magic_rb> and a GC |
| 2026-02-04 15:29:39 +0100 | <haskellbridge> | <Morj> And monad readers |
| 2026-02-04 15:29:12 +0100 | <haskellbridge> | <Morj> «what features would you like to see in rust» I want `pub fn append_byte<const N: usize>(array: [u8; N], b: u8) -> [u8; N + 1]` to compile |
| 2026-02-04 15:25:58 +0100 | <lucabtz> | yeah they missed a chance with do notation maybe |
| 2026-02-04 15:25:43 +0100 | <haskellbridge> | <magic_rb> things get so much more awkward with those imo |
| 2026-02-04 15:25:23 +0100 | <haskellbridge> | <magic_rb> not having actual do notation with Monads is to me a deal breaker |
| 2026-02-04 15:25:07 +0100 | <haskellbridge> | <magic_rb> in 99% of cases i dont give a rats ass about not having a GC and having to fight the borrow checker or sprinkle Arc<> absolutely everywhere is not what i want to be doing |
| 2026-02-04 15:25:03 +0100 | <lucabtz> | and for future |
| 2026-02-04 15:25:03 +0100 | <merijn> | If you can afford GC then something like Haskell, Scala, any ML will give you essentially all the stuff you want |
| 2026-02-04 15:24:45 +0100 | <lucabtz> | it kinda has do notation, but only for Option and Result |
| 2026-02-04 15:24:38 +0100 | <merijn> | Now, personally I think the vast majority of code is perfectly fine with a GC. In which case you pay the price for nothing |
| 2026-02-04 15:24:18 +0100 | <haskellbridge> | <magic_rb> lucabtz: monads, more type magic, GC |
| 2026-02-04 15:24:03 +0100 | <merijn> | So Rust is supercool. IFF you cannot afford/accept a GC |
| 2026-02-04 15:23:47 +0100 | <merijn> | It's more like "Rust pays a (fairly heavy) price, to attain something hard to achieve" (i.e. memory safety without GC) |