01:51 <FromGitter> <mjfiano:matrix.org> Might be a dumb question, but specs gives an internal error when i switch my code to use struct instead of class, saying `same?` method is undefined, probably from using the spectation `should be`.

01:52 <FromGitter> <Blacksmoke16> yes

01:53 <FromGitter> <Blacksmoke16> makes sense

01:56 <FromGitter> <mjfiano:matrix.org> Yeah thought so, since they are pbv

01:56 <FromGitter> <mjfiano:matrix.org> ok

01:56 <FromGitter> <Blacksmoke16> you could use `should be` if your struct implemented that method

01:57 <FromGitter> <Blacksmoke16> but at that point its basically the same as `should eq`

01:57 <FromGitter> <mjfiano:matrix.org> Right

01:58 <FromGitter> <Blacksmoke16> 👍

01:59 <FromGitter> <mjfiano:matrix.org> I'm just experimenting by removing all those "into" methods and making the types structs

01:59 <FromGitter> <mjfiano:matrix.org> It would surely be a lot less code to maintain and probably less room for user error

02:02 <FromGitter> <Blacksmoke16> indeed

11:01 <FromGitter> <manveru:matrix.org> is there a better way to write this? ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ I think there must be a better way to convert the string back... [https://gitter.im/crystal-lang/crystal?at=61052d8623956a5aa466914c]

11:04 <FromGitter> <HertzDevil> ```code paste, see link``` [https://gitter.im/crystal-lang/crystal?at=61052e371a1db149e9ef7e65]

11:06 <FromGitter> <manveru:matrix.org> oh, nice :)

11:07 <FromGitter> <manveru:matrix.org> i really wish there was a finished libgit2 crystal binding... but for now this'll have to do

11:42 <FromGitter> <manveru:matrix.org> finally got cloning and checking out a commit working :D

15:42 <FromGitter> <Blacksmoke16> https://github.com/rust-lang/rust/pull/43170 is this something we should also do?

16:04 <FromGitter> <mjfiano:matrix.org> heh this library is only 1500 lines, but test boilerplate is about 5000

18:07 <FromGitter> <Blacksmoke16> :P sounds about right

22:51 <FromGitter> <mattrberry> `out` just creates an uninitialized variable for whatever type is being pointed to, right?

22:53 <FromGitter> <mattrberry> So `out` for an argument that's a `Pointer(Char)` effectively just creates an `uninitialized Char` then passes its pointer as an argument, right?

22:54 <FromGitter> <Blacksmoke16> based on https://crystal-lang.org/reference/syntax_and_semantics/c_bindings/out.html, yes

22:57 <FromGitter> <mattrberry> If I want to allocate space for a string, I'd need to allocate that myself using something like Pointer.malloc or just `buffer = " " * 512` then pass the pointer

22:58 <FromGitter> <mattrberry> When I just use `out` in the case where I *know* it's filling many characters it's still working, but I'm guessing that it's just writing to memory that's still addressable so it's getting lucky and not segfaulting. I just wanted to make sure the Crystal compiler wasn't doing anything tricky with the `out` syntax that I wasn't expecting

22:59 <FromGitter> <mattrberry> So I'll stick to just allocating space myself first

22:59 <FromGitter> <Blacksmoke16> could you do something with a slice?

23:00 <FromGitter> <Blacksmoke16> like `Bytes.new size`, then pass a pointer to that buffer to fill it?

23:00 <FromGitter> <Blacksmoke16> then you'd be able to do `String.new bytes`

23:00 <FromGitter> <mattrberry> The benefit of just setting `buffer = " " * n` is that to_unsafe already gives me a Pointer(UInt8), so I can just pass it directly to the c binding, and in Crystal it's already a string so I don't need to do anything funky with it

23:01 <FromGitter> <Blacksmoke16> thats fair, the other option would be to `maloc` a pointer and use the `String.new ptr, size` constructor

23:01 <FromGitter> <Blacksmoke16> `malloc`*

23:02 <FromGitter> <mattrberry> Yeah that was my other thought too. Any thought if there's any perf concern with `" " * n`? It's not in a hotpath so I'm not overly concerned, but worth considering at least

23:02 <FromGitter> <mattrberry> Currently I'm just sticking with this since it's pretty simple ⏎ ⏎ ```code paste, see link``` [https://gitter.im/crystal-lang/crystal?at=6105d69e8587b302bfd5b82a]

23:05 <FromGitter> <Blacksmoke16> prob be fine

23:05 <FromGitter> <Blacksmoke16> :shrug:

23:05 <FromGitter> <mattrberry> Yeah same haha

23:15 <Guest13> hi all. i am wondering if constants can be defined with leading underscores. For example are both Hmm, and __Hmm both considered constants

23:16 <Guest13> i saw there are psudo constants such as __LINE__ defined so i would think such a variable definition would be considered a constant?